Development of a Virtual Drone System for Exploring Natural Landscapes and Enhancing Junior High School Students’ Learning of Indigenous Settlement Site Selection

Highlights

What are the main findings?

• Integrating virtual drones with 3D terrain modeling enables safe and efficient aerial observation, overcoming the limitations of traditional instruction. Task-oriented simulated flights enhance students’ spatial understanding, critical judgment, and decision-making in authentic contexts such as settlement site selection.
• Experimental results show that students improved in learning outcomes, intrinsic motivation, and system satisfaction, while maintaining a stable cognitive load. This suggests that the virtual drone system effectively supports exploration of geographic features, hydrological patterns, and settlement relationships.

What is the implication of the main finding?

• The virtual drone system developed in this study functions as a portable and interactive platform, allowing students to investigate mountains, river terraces, and meanders through exploratory tasks. Viewing landscapes from multiple perspectives promotes spatial thinking and cognitive flexibility.
• Drone simulations provide a cost-effective, safe, and flexible approach to authentic environmental observation and spatial literacy. They integrate fieldwork-like experiences with technological skill development, offering high interactivity and low entry barriers for geography education.

Abstract

This study combined virtual reality technology with drone aerial imagery of Smangus, a remote Atayal tribe situated 1500 m above sea level in Hsinchu County, Taiwan, to develop a virtual drone system. This study aims to investigate the learning effectiveness and operational experience associated with the application of the virtual drone system for exploring tribal natural landscapes and enhancing junior high school students’ learning of Indigenous settlement site selection. A quasi-experimental design was conducted with two seventh-grade classes from a junior high school in Hsinchu County, Taiwan. The experimental group (n = 43) engaged with the virtual drone system to perform settlement site selection tasks, while the control group (n = 42) learned using traditional materials such as PowerPoint slides and maps. The intervention consisted of two instructional sessions, with data collected via achievement tests, questionnaires, and open-ended feedback. The results indicated that students in the experimental group significantly outperformed the control group in learning outcomes. Positive responses were also observed in learning motivation, cognitive load, and system satisfaction. Students reported that the virtual drone system improved students’ understanding of terrain and enhanced their skills in selecting appropriate sites while increasing their interest and motivation in learning. Moreover, the course incorporated the Atayal people’s migration history and field interview data, enriching its cultural authenticity and contextual relevance.

1. Introduction

The site selection and formation of settlements are essential concerns in geography education, reflecting how humans interact with natural environments when choosing residential locations. Students must integrate multiple geographic factors—topography, hydrology, wind patterns, and transportation—to develop spatial reasoning that is both logical and contextually grounded. Traditional geography instruction, relying on paper maps, images, or teacher-centered explanations, often limits understanding due to the absence of three-dimensional observation and contextual imagination [1,2].

Geographic Information Systems (GIS) and map-based visualization enhance understanding of spatial relationships, geographic conditions, and human–environment interactions [3]. However, conventional GIS applications often require desktop computers and teacher guidance, limiting hands-on engagement with topography and real-world contexts. Without immersive, interactive experiences, students may struggle to construct meaningful geographic concepts, suggesting the need for designs that incorporate visual, contextualized, and experiential learning.

Advances in information and communication technology have enabled geography education to adopt immersive tools such as Virtual Reality (VR), Augmented Reality (AR), and 3D simulations. These tools allow students to simulate field observations, strengthening spatial reasoning and decision-making skills [4]. VR provides a first-person, interactive environment with visual feedback and navigation closely approximating the real world [5,6]. Research shows that 360-degree VR (or VR360) improves motivation, understanding of geographic concepts, and spatial exploration, particularly among low- and mid-achieving students [7]. Immersive VR has become an important tool in STEM and geography education, enabling learners to explore complex or inaccessible environments through experiential interaction. Research indicates that VR can enhance engagement, observational learning, and spatial reasoning when supported by effective scaffolding and cognitive-load management strategies [8].

Virtual drones combined with 3D terrain modeling and interactive interfaces enable aerial observation and address traditional instruction’s limitations in time, space, and safety [9]. Students can explore mountains, river terraces, and meanders while engaging in investigative tasks similar to field study. Viewing the landscape from multiple perspectives fosters spatial thinking and cognitive flexibility. Drone simulations are cost-effective, flexible, and safe, supporting authentic environmental observation and spatial literacy while integrating fieldwork simulation with technological skill development.

Recent studies have expanded the educational use of VR through integration with GIS and 3D terrain modeling. Bursztyn et al. [10] developed an immersive VR system to teach geological field measurement, demonstrating how realistic terrain visualization improves learners’ observational accuracy. Havenith et al. [11] combined drone surveys, 3D geomodels, and VR for structural geology analysis of large rockslides, highlighting the value of authentic, data-driven virtual environments for spatial understanding. These studies emphasize the potential of VR for place-based learning, especially when paired with real-world data and fieldwork tasks.

Emerging work has also focused on improving user adaptability and interface fidelity. Mota et al. [12] proposed uncertainty-aware VR interfaces that dynamically adjust visual guidance based on user behavior, illustrating how adaptive VR systems can support decision-making in environmental analysis. Makransky and Mayer [13] showed that immersive virtual field trips enhance students’ spatial awareness and engagement when tasks are scaffolded and directly linked to curriculum objectives.

Despite their potential, immersive technologies face practical challenges. VR often requires expensive equipment, high-performance hardware, and complex operations, creating barriers for widespread adoption. To overcome these issues, this study developed a virtual drone system that is low-cost, highly interactive, and portable. Touchscreen controls allow students to navigate 3D terrain, observe landforms, and simulate settlement site selection by evaluating water availability, slope, and accessibility, providing immersive learning while remaining practical for classroom application.

Immersion in virtual environments can vary across different levels, typically categorized as non-immersive, semi-immersive, and fully immersive experiences. These levels differ in the degree to which users feel presence and interaction within a simulated environment. In the proposed virtual drone system, immersion is realized at a semi-immersive level, achieved through a 3D interactive interface that allows learners to manipulate drone perspectives, explore terrain features, and observe spatial relationships in real time. This design provides sufficient sensory engagement and interactivity to support spatial learning without requiring specialized VR hardware like a head-mounted display.

Drones have become versatile tools across agriculture, surveying, disaster relief, and transportation, with applications expanding rapidly [14]. Integrating drones into education supports spatial knowledge and environmental exploration while fostering technological literacy and problem-solving skills. Ahmad et al. [15] indicated that drone-based instructional designs, particularly when combined with simulation, scaffolded learning, and immersive virtual environments, can enhance students’ confidence, competence, and spatial reasoning prior to real-world applications. This approach not only improves operational skills in safe, controlled settings but also fosters a deeper understanding of geographic concepts and decision-making processes. STEM curricula using simulated flight operations also improve action skills, engagement, and learning interest [16,17].

A review of drone-assisted STEM education (2005–2023) reveals teamwork, hands-on practice, and contextual simulations as dominant strategies, producing outcomes in technical skills, problem-solving, and career awareness [18]. Responding to these trends, the virtual drone system developed in this study emphasizes portability, usability, and interactivity. Operated through mobile devices, it allows highly accessible immersive learning without requiring any additional specialized hardware. In this study, task-oriented and simulated flights are designed to enhance spatial understanding, critical judgment, and decision-making skills in authentic geographic contexts, reflecting the pedagogical principle of learning by doing.

This study focuses on the Smangus tribe, a remote Indigenous community in Hsinchu County, Taiwan (Figure 1). On-site topographic photography and modeling produced a realistic 3D terrain and natural landscapes integrated into the virtual drone system for interactive and immersive exploration. Historically, Smangus relocated its settlement based on water sources, topography, land suitability, and community requirements, reflecting collective Indigenous decision-making. Embedding this context aligns with place-based concepts, which emphasize local knowledge and cultural narratives as foundations for meaningful learning in geography education [19,20].

Within the virtual environment, students can select flight paths, observe settlement–landform relationships, and perform tasks simulating relocation and expansion. Combining virtual operation, task orientation, authentic terrain, and local culture fosters geographic perspective-taking, spatial reasoning, and cultural understanding. Students gain experiential insight into human–environment interactions and Indigenous decision-making, reinforcing critical thinking about settlement planning.

Traditional geography instruction often relies on static teaching materials, limiting interactive and 3D experiences essential for spatial understanding and reasoning. Immersive tools like head-mounted displays offer authentic learning experiences but often encounter cost, operational, and feasibility challenges, thereby relying on simplified or artificial terrain models. This study addresses these challenges in line with Taiwan’s 12-Year Basic Education Curriculum Guidelines, suggesting competence-based, interdisciplinary learning that integrates technology, local knowledge, and context to develop geographic literacy, technological competence, and cultural identity.

Unlike many VR-learning studies that use generic or synthetic environments, our system integrates drone-captured aerial imagery and GIS-derived 3D terrain reconstructions of Smangus (Atayal territory), preserving high-fidelity landscape and hydrological patterns that are directly relevant to Indigenous settlement issues. The learning environment fuses ecological observation with Atayal migration history and field-interview data, embedding culturally authentic narratives into task-based exploration, a combination seldom reported in prior drone/VR educational work. This study combines simple, scaffolded drone flight tasks with real-world decision-making scenarios (site selection), rather than focusing only on instrument training or technical piloting—bridging immersive experience with curriculum-relevant spatial reasoning. The quasi-experimental design measures learning achievement, learning motivation, cognitive load, and system satisfaction, providing empirical evidence that the virtual drone system can be pedagogically effective and manageable for younger learners.

The virtual drone system developed in this study serves as a portable, low-barrier, interactive platform. Authentic Smangus terrain, reconstructed through drone imagery and on-site modeling, is combined with task-based instructional design. Students can conduct virtual flights and settlement evaluations to simulate Indigenous decision-making. The 3D terrain model preserves scale and texture, enabling exploration of geographic features, hydrological patterns, and settlement relationships. This approach enhances spatial understanding, landform observation, and reasoning while balancing technological innovation with cultural preservation. By integrating immersive technology, authentic terrain, and culturally meaningful tasks, the study demonstrates a model of geography education that increases spatial literacy, decision-making, and cultural awareness.

The virtual drone system overcomes traditional limitations in geography education, providing safe, interactive, and contextually grounded learning experiences that connect geographic knowledge with the local place, natural environments, and human activities, while fostering a sense of environmental and cultural responsibility. The system seeks to promote students’ cross-disciplinary integration and technological literacy, aligning with the STEM principle of context- and task-driven knowledge application. Through empirical validation, the current study seeks to provide geography education with an innovative instructional solution characterized by high fidelity, interactivity, and portability. Specifically, the research addresses the following questions:

(1)

Can the virtual drone system effectively improve students’ ability to select suitable settlement sites?

(2)

Does the virtual drone system increase students’ learning motivation while reducing cognitive load?

(3)

How do students perceive the usability and satisfaction of the virtual drone system?

These questions are designed to assess the value and effectiveness of the virtual drone system in geography education, while also advancing our understanding of its potential impact on students’ learning processes, spatial cognition, and cultural engagement. These findings are expected to provide empirical evidence, pedagogical insights, and future directions for integrating VR technology into geography education.

The remainder of this article is organized as follows: Section 2 reviews the relevant literature; Section 3 describes the materials and methods employed in this study; Section 4 presents the results and discussion; and Section 5 concludes the study and outlines directions for future work.

2. Literature Review

This section divides the literature review into four subsections to establish the study’s theoretical foundation and provide conceptual guidance for subsequent analysis. Section 2.1 explores the role of settlement site selection within geography education and discusses the challenges involved in fostering students’ spatial literacy. Section 2.2 discusses spatial thinking skills and relevant applications to support their development. Section 2.3 reviews the educational potential and design elements of a virtual drone system in geography education. Finally, Section 2.4 provides theoretical perspectives on learning outcomes, learning motivation, cognitive load, and learner experience, which collectively inform the instructional design and evaluation of its effectiveness in this study.

2.1. Importance of Settlement Site Selection

Settlement site selection is an essential theme in geography education, integrating natural environments, human activities, and spatial reasoning. It involves multiple factors such as topography, hydrology, climate, resources, and sociocultural contexts, offering interdisciplinary educational potential [21]. According to Taiwan’s 12-Year Basic Education Curriculum Guidelines, geography education emphasizes human–environment interactions and the development of spatial thinking skills for problem-solving abilities, fostering geographic literacy and civic awareness [22,23].

Traditional instruction often depends on paper-based materials, static images, maps, and teacher-centered lectures, limiting interactive participation and three-dimensional perspectives. These approaches can hinder students’ comprehension of geographic concepts and constrain the development of spatial reasoning skills, thereby restricting both the depth of knowledge and its transfer [1,2,4]. This section proposes three key challenges: the curricular positioning of settlement site selection, difficulties in fostering spatial thinking, and limitations of traditional teaching methods. These challenges form the basis for the research design proposed in this study.

2.1.1. Criteria for Settlement Site Selection

Settlement site selection embodies human–environment interactions and is a central learning objective in junior high school geography. Students are expected to understand how natural and human factors—such as transportation, water availability, climate, and topography—influence settlement distribution, while developing geographic literacy through inquiry and problem-solving [23]. Curriculum guidance emphasizes contextualized, life-oriented instruction aligned with place-based education principles, fostering emotional connections to place and critical geographic thinking [19]. In Indigenous or remote contexts, settlement activities can integrate local history, land knowledge, and cultural memory, strengthening environmental identity [20,24].

The curriculum also promotes integration of technology and interdisciplinary learning to enhance spatial thinking. For example, GIS and virtual simulations help externalize spatial relationships, improving reasoning and decision-making. Therefore, settlement site selection supports spatial literacy, human–environment understanding, and the integration of technology, culture, and interdisciplinary learning [1,4,25].

2.1.2. Developing Students’ Spatial Thinking

Spatial thinking enables students to comprehend geographic phenomena and make informed decisions. It involves the following: (1) understanding spatial concepts such as direction, distance, scale, and distribution; (2) using representational tools such as maps and virtual models; and (3) applying geographic reasoning to decisions concerning migration, settlement site selection, and policy planning. Despite its importance, spatial thinking is rarely taught in secondary geography. As a result, students engage passively with spatial concepts, lacking sufficient opportunities for visualization, modeling, and problem-solving, which hinders the development of reasoning and decision-making abilities [22,26]. Studies using Google Earth 7.3.6.10201 or GIS indicate that, without structured guidance, students tend to passively receive visual information and fail to form meaningful connections [4,27].

Instructional design plays a crucial role in fostering spatial thinking in geography education. Integrative approaches combining spatial concepts, symbolic tools (e.g., contour maps, GIS layers), and reasoning tasks enhance students’ abilities to compare, explain, and evaluate locations [28,29,30]. Research in cognitive psychology confirms that spatial thinking is teachable, and targeted assessment tools can enhance skills in graphic transformation, positional comparison, and spatial reasoning [31]. Effective spatial thinking instruction requires innovations in teaching materials, activities, and assessments. Simulation-based operations, task-oriented learning, and digital tools (e.g., virtual terrain models and Google Earth 7.3.6.10201) provide interactive contexts for observation, classification, comparison, and judgment, enhancing spatial reasoning, representation use, and decision-making, while fostering geographic literacy and problem-solving skills.

2.1.3. Limitations of Traditional Teaching Methods

Traditional geography instruction in Taiwan often relies on maps, static images, and teacher-centered lectures. Although stable and easy to implement, these methods provide limited support for higher-order spatial reasoning and problem-solving [32]. Static materials may fail to engage students in constructing geographic concepts or linking spatial information to decision-making. Interactive 3D maps outperform static visuals in promoting spatial understanding and task accuracy, revealing the constraints of traditional learning media [33]. Lecture-based instruction also provides limited support for motivation, engagement, and critical thinking. Cooperative or game-based approaches improve motivation, participation, and cognitive engagement, suggesting a need for interactive and socially constructed learning [34,35]. In addition, digital tools and blended learning approaches, including field-based learning and virtual exploration, enhance engagement, spatial reasoning, and internalization of geographic knowledge [36,37].

While traditional methods provide underlying instructional value, they fail to meet modern educational demands for interactivity, contextual relevance, and critical engagement. To enhance spatial thinking and geographic literacy, instruction should shift toward active knowledge construction using interactive technologies, task-based learning, and collaborative strategies that foster engagement, reasoning, and problem-solving.

2.2. Spatial Thinking for Learning Geography

Given the challenges of fostering spatial thinking, geography educators must consider how to cultivate spatial literacy. Instruction should go beyond conveying concepts by integrating tools and activities that connect observation, representation, reasoning, and decision-making. Spatial literacy, a core geographic competency, is also a critical requirement in contemporary geography education, encompassing image interpretation, spatial reasoning, and the translation of information across various forms.

2.2.1. Importance of Spatial Thinking Ability

Spatial thinking refers to multilayered cognitive processes for handling spatial information, including positional understanding, visual transformation, image manipulation, and reasoning-based modeling. It is a core competency in geography education and a key bridge across STEM disciplines [38]. Spatial thinking comprises three interrelated dimensions: spatial concepts, spatial tools, and spatial reasoning—forming an integrated framework that combines perception, symbolic representation (e.g., maps, graphs, coordinate systems), and inferential reasoning.

Research shows that students often struggle with map legends, scale interpretation, and GIS use, indicating the importance of multi-perspective, interactive tasks to connect abstract geospatial data with real-world landscapes. Hence, spatial tools act as cognitive bridges, enabling knowledge transformation through visualization and manipulation, while spatial reasoning—encompassing mental reconstruction, transformation, and decision-making—serves as the core driver of spatial cognition. Training frameworks, such as Think3D!, demonstrate that targeted tasks in mental rotation, perspective-taking, and mirror-image recognition can effectively enhance spatial skills, with transfer effects to mathematics and engineering problem-solving [39]

Spatial thinking develops progressively. Piaget [40], along with subsequent studies [38], identified a transition from concrete perceptions—such as route navigation and directional awareness—to more abstract operations, including scale comprehension and three-dimensional mental rotation. This development relies on repeated engagement with hands-on tasks, interactive visualization, and reflective manipulation exercises, gradually forming stable mental models. Insufficient spatial strategies during cross-perspective tasks can impede disciplinary learning, emphasizing the importance of guided interaction to cultivate self-directed spatial strategies [41].

Empirical evidence supports the trainability of spatial thinking. Bednarz and Lee [31], using the Spatial Thinking Abilities Test (STAT), found that task-oriented, image-based activities significantly improve spatial reasoning, whereas traditional lectures and static materials yield limited effects. Mathewson [42] emphasized that visual–spatial processing—image transformation, visual memory, and spatial composition—is essential for scientific learning, necessitating diverse opportunities for model- and image-based manipulation. Spatial thinking is a trainable, context-sensitive cognitive process rooted in conceptual understanding, facilitated by spatial tools, and propelled by reasoning. Through guided practice and interactive engagement, these elements become cognitive tools that help students interpret landscapes, construct spatial models, and solve real-world problems. Understanding students’ developmental stages and learning pathways enables geography education to more effectively cultivate interdisciplinary spatial literacy.

2.2.2. Enhancing Spatial Thinking Through GIS

GIS integrates spatial data, visualization, and analytical tools, making it a powerful tool in geography education. Beyond traditional cartography, it supports interactive simulations and real-time spatial modeling, serving as an effective medium for developing students’ spatial thinking, including understanding spatial concepts, operations, and reasoning across scales. Evidence indicates that GIS not only facilitates information processing but also promotes higher-order spatial cognition. Lee and Bednarz [43] showed that GIS interventions, combined with problem-based learning and hands-on tasks, significantly improved students’ map interpretation, visual transformation, and spatial reasoning skills. Duarte et al. [44] discovered that GIS-based instruction enhanced spatial visualization and mental map construction, indicating the importance of contextualized, dynamic, and task-driven designs.

From an educational innovation perspective, GIS should evolve beyond data querying to serve as a strategic platform for students to construct geographic meaning. Donert and González [45] emphasized that learners evolve from data encoders to problem solvers, developing geographic thinking through observation, analysis, and modeling. Moorman [46] further argued that geospatial literacy requires integrating spatial understanding, tool application, and contextual reflection rather than mere software proficiency.

Competency-based GIS frameworks emphasize spatial thinking as the primary learning objective. Hickman [47] asserted that tasks such as settlement site selection, resource allocation, and disaster simulation develop robust spatial judgment under cognitively demanding conditions, encouraging cross-scale reasoning and perspective transformation. Wakabayashi and Ishikawa [48] identified three core dimensions—concepts, operations, and reasoning—showing that GIS enhances spatial operations and reasoning through data overlays, perspective switching, and simulations.

Meta-analytic evidence clearly confirms GIS’s substantial educational value across multiple contexts. Ma et al. [49] reported a moderate positive effect on learning outcomes, with significant gains in spatial thinking, motivation, and problem-solving, particularly when curricula explicitly emphasized authentic, meaningful spatial tasks over purely technical training. Cognitive research emphasizes the importance of scaffolding strategies, such as layering, directional cues, and scale guidance, to support accurate mental model construction and reduce common interpretive errors [50].

Immersive platforms, such as Virtual Geographic Environments (VGE), further enhance spatial construction and task efficiency by allowing students to explore and decide in low-risk contexts [51]. When paired with problem-based learning, contextual simulations, and cognitive scaffolds, GIS transforms students from technical operators into critical geographic thinkers. It enhances curriculum depth, fosters spatial literacy, and equips learners with reasoning and decision-making competencies essential for understanding and acting within complex geographic systems.

2.2.3. Three-Dimensional Simulation and Interactive Visualization

In the evolving context of digital learning environments, three-dimensional simulation and interactive visualization have become important tools for enhancing spatial understanding and cognitive reasoning in geography education. These technologies transform abstract concepts of landforms, spatial relationships, and geographic processes into concrete and manageable visual experiences. Consequently, they help students overcome the limitations of traditional 2D materials, enabling them to construct more comprehensive mental spatial representations and to develop higher-level spatial reasoning and geographic decision-making skills.

Anthamatten and Ziegler [52] argued that while traditional maps and static materials provide foundational information, they are often limited in helping students grasp complex landform and contour concepts. Their teaching experiments demonstrated that 3D models assisted students in recognizing elevation differences, observing spatial patterns from multiple perspectives, and exploring the meaning of geographic structures and distributions. They further emphasized that 3D simulations are effective for visual learners, serving as a powerful complement to conventional teaching.

Yin advanced this perspective by showing that integrating 3D simulations with GIS in instructional design enables students to engage in spatial planning and visual range analysis within virtual cityscapes [53]. This integration not only improved students’ geographic data interpretation but also strengthened their planning skills and decision-making logic. When tasked with resource allocation under constraints in virtual environments, students considered geographic factors such as topography, visibility, and spatial proximity, thereby translating abstract knowledge into real-world problem-solving ability and strengthening their practice-oriented spatial reasoning skills.

In secondary and higher education, Philips et al. [54] showed that immersive 3D visualization improved students’ understanding of geographic processes such as changes in land utilization, climate variation, and watershed evolution. Across German universities, students demonstrated higher motivation, stronger engagement, and better cross-scale integration, with results also enhancing confidence in professional geographic practice. Similarly, Hauptman’s Virtual Spaces 1.0 platform [55], featuring tasks such as mental rotation, spatial matching, and distance estimation, enabled experimental groups to outperform control groups in spatial recognition and reasoning, confirming the value of interactive virtual environments for developing spatial thinking skills.

Augmented Reality (AR) has proven effective as a complementary tool in geography instruction. Carbonell Carrera and Bermejo Asensio [56] used AR to overlay virtual buildings, terrain, and spatial data onto maps, enabling simultaneous engagement with real and virtual environments. Their results showed significant gains in spatial rotation, 3D image construction, motivation, and participation. More recently, Purwanto et al. [57] introduced the Geo-Virtual Reality (GVR) system, which combines VR and metaverse technologies to create immersive learning contexts. Students simulated geomorphic changes, completed spatial tasks, and recreated spatial events, leading to improvement in landform interpretation, spatial reasoning, and problem-based learning.

In a systematic review of VR applications in higher education, Radianti et al. [9] noted that immersive systems integrating real-time feedback, autonomous inquiry mechanisms, and clearly defined instructional objectives can increase student engagement and cognitive depth. However, they also cautioned that without effective pedagogical guidance and proper design, 3D simulations may lead to cognitive overload or usability issues, potentially impeding learning. Thus, educators must carefully consider students’ cognitive load and learning progression when adopting such technologies.

Three-dimensional simulations and interactive visualizations demonstrate strong potential to enhance geography learning by making spatial concepts more accessible and engaging. They not only enhance spatial thinking, increase motivation to learn, and bridge theoretical and practical knowledge, but also open diverse pathways for cultivating interdisciplinary geographic literacy. Prospectively, geography education that integrates emerging technologies—such as VR, AR, and GVR—alongside hands-on tasks and pedagogical scaffolding, can support learners more efficiently by advancing instruction from perception to understanding and from knowledge transmission to competency development.

2.3. Virtual Drones as a Learning Tool

With the expanding incorporation of technology in education, unmanned aerial vehicles (UAVs) or drones have emerged as highly mobile and elevated tools with significant instructional potential. In geography education, their aerial perspectives and terrain-sensing capabilities directly support the development of spatial literacy. However, classroom implementation of a physical drone faces challenges such as operational risks, high costs, weather dependence, and space limitations. Virtual drone systems provide a practical alternative by simulating drone flights through 3D modeling and task-oriented activities. These platforms enable realistic exploration and decision-making practice without the risks of using real drones, while promoting spatial thinking, geographic understanding, and interdisciplinary engagement. Their low entry barriers, high interactivity, and flexibility make virtual drones increasingly valuable in STEM education.

This section is organized into three parts: (1) a review of drone development and educational applications, emphasizing their role in STEM integration; (2) an analysis of key design elements and pedagogical implications of virtual drone systems; and (3) a discussion of the virtual drone system developed in this study, highlighting its task-oriented, interactive features and situating it within prior virtual flight research to highlight its innovation and practical relevance for geography education.

2.3.1. Applications of Drones in STEM Education

In recent years, drones have expanded from military and industrial applications into education, enhancing students’ technological literacy, spatial awareness, and problem-solving skills. Across secondary and higher education, drones are integrated into STEM curricula through flight missions, image analysis, and map construction, promoting interdisciplinary learning [11]. They integrate competencies in programming, spatial data analysis, engineering logic, and environmental observation. Compared with traditional maps, drones provide aerial perspectives and real-world data, enabling a concrete understanding of landforms and environmental changes [12]. Task-based flight planning and real-time feedback cultivate spatial reasoning, decision-making, and critical thinking. Both real and virtual drones are effective instructional media, with the Federal Aviation Administration (FAA) predicting drone operation as an emerging workforce skill [10]. Consequently, drone-based learning enhances practical skills, creativity, and career readiness.

Systematic reviews highlight the efficacy of drones in STEM education. Yeung et al. [18] found that problem-based learning and design-thinking approaches integrating drones, mathematics, physics, and programming enhanced spatial reasoning, computational thinking, teamwork, and engineering skills. Modular drone curricula combining flight and programming tasks further improve spatial understanding, problem-solving, and collaboration [58], guided by principles of task-oriented learning, simulation integration, and disciplinary alignment.

In geography education, virtual drones have been employed to simulate landform surveys, enabling indoor terrain observation and data recording [59], significantly improving motivation, exploratory behavior, and spatial concept construction. Cross-level studies indicate drone instruction fosters immersive learning, technological literacy, and inquiry-oriented exploration, benefitting low-achieving students, though teacher training and policy support are essential for large-scale adoption [60].

Research in higher education demonstrates drones’ impact on topographic observation, geomorphic interpretation, image analysis, and data collection, while advancing digital geography competencies [61,62]. However, drone curricula often focus on basic operation and photography, with insufficient integration of GIS, remote sensing, and data analysis [63]. Drones also contribute to interdisciplinary and sustainability-focused learning, supporting problem-based projects in environmental monitoring and forestry, enhancing technical skills and understanding of real-world issues [64,65,66].

Overall, drone education has evolved from aerial photography demonstrations to comprehensive designs integrating technical operation, curricular objectives, and competency-based learning. Virtual drones’ immersive, interactive, and mobile features make them ideal for authentic, task-based learning, fostering spatial thinking, geographic literacy, and technological competence. Future research directions include enhancing simulation-practice integration, developing interdisciplinary modules, and establishing assessment tools to track learning processes and outcomes.

2.3.2. Key Elements in Virtual Drone Systems

In this study, the term “virtual drone” refers to a computer-generated model that simulates the functions and behaviors of a real drone, including navigation, flight control, and aerial observation within a three-dimensional digital environment. In contrast, the term “virtual drone system” denotes the complete learning platform, which integrates the virtual drone with supporting components such as terrain visualization modules, interactive interfaces, instructional scenarios, and data recording functions.

While the descriptions of flight control and visualization pertain to the virtual drone itself, the discussions of instructional design, learner interaction, and performance assessment concern the broader virtual drone system as an integrated educational framework. To enhance the potential of virtual drone systems in geography education, advanced design must address three key dimensions: realistic simulation environments, stable flight control, and effective terrain visualization. Virtual drone systems should align with learner characteristics and instructional objectives to create an interactive, task-oriented platform that fosters cognitive development and achieves optimal learning outcomes.

High-fidelity 3D reconstruction forms the foundation for immersive learning and geographic understanding. By combining drone-based photography with point-cloud modeling and texture mapping, virtual drone systems can accurately reproduce landforms and spatial relationships [67]. Modular task design, including adjustable difficulty and scenario variation, allows alignment with different learning stages. For example, Szóstak et al. [68] incorporated wind-speed interference and visual obstructions to simulate realistic decision-making challenges in a safe simulation environment.

Flight control design requires real-time responsiveness, stability, and user-friendly operation. Integrating open-source platforms such as PX4 with simulators like Gazebo or AirSim enhances scalability and instructional flexibility [69], while novice-oriented interfaces, task prompts, and visual guidance reduce cognitive load and build user confidence [70,71]. In addition, collaborative and shared-perspective functions further support social interaction and spatial negotiation skills [72]. Adaptive mechanisms, including reinforcement learning–based adjustments of task difficulty and control sensitivity, enable personalized instruction and optimize learner performance [67]. Task-based activities integrated with landscape analysis and contextual simulations have also been shown to strengthen problem-solving and geographic reasoning [64].

Effective virtual drone systems combine high-fidelity 3D terrain modeling, interactive task-oriented learning, adaptive controls, and collaborative functionality for educational applications. By integrating these technological features with pedagogical goals, virtual drone systems can serve as powerful tools for cultivating spatial literacy, spatial reasoning, and cognitive skills in geography education.

2.3.3. Related Research in Virtual Drone Systems

Recent studies have explored the use of virtual drone systems as innovative tools for spatial learning, simulation training, and environmental visualization. Early research primarily focused on engineering education and flight skill development, emphasizing control precision, aerodynamic simulation, and hardware–software synchronization [73]. Later works extended these applications to science and geography education, using drone-based simulations to visualize topography, monitor land use, and promote spatial awareness [74]. Huang and Hu reported that interactive drone simulations enhance learners’ engagement and support spatial reasoning by providing a dynamic, manipulable 3D environment [75]. Immersion levels vary across systems—from desktop-based semi-immersive platforms to fully immersive VR drone experiences—each offering different degrees of presence and cognitive engagement.

Prior systems were mostly designed for professional training or environmental observation rather than pedagogical integration in formal geography curricula. The study by Albeaino et al. [3] demonstrated that virtual drone systems allow students to learn drone operation skills within a realistic environment and apply these skills for actual scientific observations. This interactive operation can enhance their learning motivation and sense of accomplishment. Virtual drone systems enable students to conduct flight training in a safe environment, stimulating their interest in scientific inquiry. When learners operate drones for aerial imaging, they gain a more intuitive understanding of scientific concepts such as atmospheric, geographical, and ecological phenomena.

Building upon these findings, the present study develops a virtual drone system specifically tailored for geography education. By integrating terrain simulation, drone control, and task-oriented learning activities, it aims to address existing research gaps concerning the application of virtual drone systems for enhancing spatial understanding and reasoning in school-level contexts. Students can observe landforms such as mountains, rivers, and forests, and learn how to collect and analyze spatial data using drones. This approach makes concepts in environmental science and ecological conservation more concrete and visualized, while also fostering students’ motivation and interest in learning.

2.4. Learning Outcomes and Learner Responses

Designing and evaluating technology-based learning systems requires understanding learners’ psychological and behavioral processes. To comprehensively evaluate the instructional effectiveness of the virtual drone system developed for geography education, this study focused on four core dimensions: learning outcomes, motivation, cognitive load, and learner experience. These aspects capture students’ achievements in knowledge, skills, and attitudes, while illuminating the cognitive and affective processes engaged during interactive learning activities.

This section examines each dimension in turn: the multifaceted definitions and assessment levels of learning outcomes; the categorization of learning motivation according to Self-Determination Theory (SDT); the instructional implications of Cognitive Load Theory (CLT); and the essential role of User Experience (UX) in evaluating digital learning environments. By integrating these theoretical perspectives, the study provides a rigorous foundation for quantitative evaluation and enhances understanding of both students’ learning trajectories and the system’s overall effectiveness.

2.4.1. Learning Outcome Assessments

Learning outcomes reflect students’ demonstrated knowledge, skills, and attitudes after completing a learning activity, serving as critical indicators of instructional effectiveness [76]. Bloom’s Taxonomy categorizes learning into cognitive, affective, and psychomotor domains, providing both theoretical and practical guidance for assessment. Krathwohl et al. refined the affective domain, highlighting levels from value comprehension to behavioral enactment [77]. Constructivist perspectives stress that assessment should measure internalized understanding aligned with instructional objectives [78,79].

Immersive technologies have been shown to enhance learning outcomes, particularly in spatially oriented and task-based domains [2]. In geography education, assessments should capture map interpretation, spatial reasoning, and problem-solving skills through integrated methods such as fieldwork, conceptual tasks, and scenario-based activities [80]. Performance-based assessment requires explicit objectives, measurable criteria, and feedback loops to inform instruction [81]. A combination of qualitative and quantitative tools, including tests, projects, observations, and reflective reports, supports authentic evaluation of knowledge, skills, and attitudes [82].

Effective learning objectives are observable, measurable, and aligned with disciplinary competencies, using action-oriented directions to guide assessment design [83]. Contemporary approaches emphasize assessing cognitive, affective, and hands-on performance outcomes with differentiated methods appropriate for each domain [84]. In line with these principles, the virtual drone system developed in this study employed pre- and post-tests for measuring cognitive gains, scenario-based operational tasks for skill assessment, and reflective feedback for affective engagement. This multi-layered approach enables comprehensive evaluation of students’ learning transformations and achievement within a virtual geographic environment.

2.4.2. Learning Motivation Theory

Learning motivation is the psychological drive that directs learners’ engagement, goal-setting, and persistence. Self-Determination Theory (SDT), widely applied in educational research, explains motivation by distinguishing intrinsic motivation—driven by interest and enjoyment—from extrinsic motivation, which stems from external incentives or pressures [85]. Intrinsic motivation is fostered when three psychological needs—autonomy, competence, and relatedness—are met [86]. Competence is strengthened when tasks provide appropriate challenge, timely feedback, and clear objectives, enhancing students’ sense of “I can do this” and sustaining engagement [87].

Extrinsic motivation can support learning if internalized; rewards framed in competence-enhancing, supportive ways increase engagement, whereas controlling incentives can undermine autonomy [88]. In digital and multimedia learning, these principles hold true: immersive and interactive systems that allow choice, adjustable difficulty, and social interaction effectively satisfy psychological needs, fostering intrinsic motivation [89]. Visual cues, task guidance, and feedback loops further support deep learning behaviors [90]. Teachers can enhance motivation by promoting autonomous inquiry, self-expression, and challenge, aligning with the Tripartite Model of Intrinsic Motivation [91]. Learning motivation is influenced by both intrinsic interest and the extent to which instructional environments satisfy learners’ psychological needs. Digital platforms and simulation systems aligned with SDT principles—such as offering choices and opportunities for social interaction—can effectively promote sustained engagement and deeper learning.

2.4.3. Cognitive Load Theory

Cognitive Load Theory (CLT) asserts that working memory has limited capacity, and poorly designed instruction can overload learners, reducing learning effectiveness [92,93]. CLT identifies three types of load: intrinsic, extraneous, and germane, and emphasizes balancing these by reducing extraneous load, regulating intrinsic load, and fostering germane load [94]. Strategies such as the modality effect and spatial contiguity help learners integrate information efficiently [95], while task difficulty, structure, and engagement should match learner capabilities and include guidance and self-regulation support [96]. Measurement of cognitive load increasingly integrates subjective ratings, physiological indicators, and performance outcomes to enhance validity [97]. Differentiated scales can guide instructional optimization [98]. In immersive VR contexts, segmented tasks, multi-channel design, and explicit guidance can reduce extraneous load and enhance germane load [99], offering insights into how instructional design can support more effective cognitive processing and deeper spatial learning in a virtual environment.

In geography education, CLT informs spatial and map-based instruction. Poorly designed GIS tools can overload learners and hinder outcomes [100], while 3D geographic technologies, though beneficial for spatial thinking, may increase cognitive demands [101]. Tasks and interface adjustments help learners allocate cognitive resources effectively. This study uses questionnaires to assess cognitive load, evaluating whether the virtual drone system reduces extraneous interference while promoting geographic reasoning. CLT provides an essential framework for designing and assessing virtual learning tools, enhancing cognitive efficiency and learning quality in geography education.

2.4.4. Learner Experience Evaluation

Learner Experience (LX) evaluates the effectiveness and acceptance of digital learning platforms, integrating cognitive, affective, and behavioral dimensions. Rooted in User Experience (UX), LX emphasizes not only usability but also enjoyment, satisfaction, and emotional engagement [102]. Zaharias and Poylymenakou [103] identified five elements central to LX assessment—learnability, usability, content quality, interactivity, and learning orientation—which are crucial in immersive environments where interface immediacy and responsive feedback help prevent cognitive overload.

Research indicates that while immersive technologies can enhance engagement, they may hinder learning outcomes without appropriate instructional guidance [104,105]. Interactivity and feedback, such as real-time responses, multimedia, and collaborative functions, positively influence motivation, efficiency, and self-directed learning [106,107]. Dynamic, adaptive evaluations more accurately capture learners’ evolving experiences [108]. Visual and interaction design also strongly shape LX. Simplicity, structural consistency, and clear guidance reduce uncertainty and operational anxiety, fostering engagement and persistence [109]. Effective LX evaluation combines functionality, aesthetics, interactivity, motivation, and emotional involvement. Educators and system designers should adopt dynamic, context-driven, and meaningful approaches to ensure digital platforms support meaningful learning and personalized development.

3. Materials and Methods

This study developed a virtual drone system for application in junior high school geography education, evaluating its impact on students’ learning effectiveness, motivation, cognitive load, and user experience. A quasi-experimental design was employed, with data collected through achievement tests and quantitative questionnaires to evaluate learning outcomes and student responses. This section outlines the research design and system development, organized into five subsections: Section 3.1 introduces the research framework and participants; Section 3.2 details the learning content and instructional design; Section 3.3 describes the development and operation of the virtual drone system; Section 3.4 explains the overall research design; and Section 3.5 presents the statistical methods used for data analysis.

3.1. Research Framework and Participants

This study develops a virtual drone system to enhance students’ understanding of settlement site selection and foster spatial thinking skills. It centers on three components: instructional design, system development, and empirical evaluation. A Design-Based Research (DBR) approach was adopted, progressing through iterative stages of conceptualization and planning, system development, pilot implementation, curriculum refinement, and formal teaching experiments to optimize the virtual drone system and its materials. To achieve these objectives, the research was structured into six phases: conceptualization and planning, system development, instructional design, pilot testing, formal experiment, and data analysis, as described below:

(1)

Conceptualization and Planning:

• Identification of research themes and issues;
• Determination of preliminary instructional objectives and framework;
• Execution of the literature review.

(2)

System Development:

• Investigation of technology and testing of tools;
• Collection of image and video materials through drone aerial imagery;
• Development and testing of system functions and user interface.

(3)

Instructional Design:

• Design of settlement site selection instruction, including task-oriented operations;
• Development of testing instruments (pre- and post-tests), covering learning effectiveness, motivation, cognitive load, and system usability;
• Design of instructional schedule and detailed activity plan.

(4)

Pilot Testing:

• Execution of a pilot test to verify system operability and instructional suitability;
• Collection of students’ learning outcomes and feedback on operational experience;
• Analysis of pilot test and questionnaire results;
• Review of pilot testing outcomes with an expert panel to revise test items

(5)

Formal Experiment:

• Selection of junior high school students as research participants, with assignment to experimental and control groups according to a quasi-experimental design;
• Conduct the teaching experiment with pre- and post-tests, collecting learning logs and system usage data.

(6)

Data Analysis:

• Processing and statistical analysis of collected data;
• Examination of the system’s effects on students’ learning outcomes and responses;
• Summarization of research findings and proposal of suggestions for improvement.

The participants of this study were seventh-grade students from a junior high school in Hsinchu City, Taiwan. A teaching experiment was conducted using the virtual drone system, and its effectiveness was compared with traditional instruction delivered through PowerPoint slides and maps. Participants and school selection were based on three principles: (1) students possessed sufficient foundational knowledge for the curriculum; (2) the school and teachers showed strong support for the experimental study; and (3) the school had prior collaboration experience to ensure smooth implementation. Based on these considerations, the final sample consisted of two intact, regularly assigned classes at the same grade level, totaling 85 students. Group assignment was conducted at the class level rather than through individual randomization. The experimental group comprised 43 students (21 boys and 22 girls) who received instruction supported by the virtual drone system. The control group included 42 students (14 boys and 28 girls), taught using traditional teaching materials, including PowerPoint slides and maps.

During the instructional treatment, both groups received the same content, instructional time, and were taught by the same teacher, ensuring consistency in curriculum delivery. The only difference was the instructional medium and the way of interaction. Specifically, the experimental group engaged in immersive and interactive learning activities using the virtual drone system, while the control group followed a traditional lecture-based approach supplemented with printed materials and maps to complete the same learning tasks. The quasi-experimental design was adopted due to practical constraints related to school scheduling and classroom operations. Using this approach, the study aimed to compare the effects of different instructional media on students’ learning outcomes and experiences while keeping the learning content consistent. A pre-test on settlement site selection concepts was administered to both groups to confirm no significant difference in prior knowledge, establishing group equivalence. This controlled setup ensured that observed differences in outcomes were attributable to the instructional approach rather than extraneous variables.

3.2. Instructional Design and Research Tools

In this study, the Smangus community was used as the instructional context to simulate historical migration and settlement expansion. Students evaluated potential settlement sites based on terrain, water resources, and transportation, enhancing understanding of human–environment interactions. To increase authenticity, the research team conducted fieldwork at the Smangus community and interviewed the tribal chief, obtaining first-hand relocation insights. During the intervention, students actively engaged in virtual drone operations, terrain observations, and annotated Google Earth layers, fostering spatial reasoning, geospatial manipulation, and information literacy.

3.2.1. Instructional Content

The instructional content was organized around the major theme of migration history and settlement site selection of the Smangus community. Four interrelated thematic units were designed to guide students from basic conceptual understanding to task-based practice, thereby enhancing their geographic reasoning and spatial thinking skills. The content of each thematic unit is outlined as follows:

• Importance of settlement site selection

This unit opens with the question, “If you were establishing a new settlement, where would you choose?” to engage students in considering geographic conditions and human settlement requirements. Using visual prompts and group discussion, students explore major survival factors: water sources, terrain, and climate, alongside safety considerations such as defensive topography and natural barriers.

• Migration history and geographic context

This unit presents the migration history of the Smangus community. Using slides and maps, students can trace the journey from Ruiyan through Siyuan Pass to Krasan and the current settlement, examining site-selection reasons and geographic influences at each stage. This approach fosters cultural–geographic connections, deepens the sense of place, and enhances awareness of Indigenous history. It is followed by a discussion of the relocation process, which considers topography, water availability, and accessibility, thereby illustrating the dynamic and informed nature of settlement decisions.

• Fundamental geographic knowledge

Interactive instruction in this unit introduces key geographic concepts, including river terraces, incision, watersheds, and ridgelines. Through slides and targeted tasks, students examine topography and hydrology while linking these concepts to settlement site selection. Guiding questions like “Why are natural barriers important?” and “How does wind direction influence relocation?” encourage critical thinking. The unit strengthens geographic reasoning and facilitates the transfer of knowledge to real-world contexts.

• Settlement site selection for tribal expansion

The final unit emphasizes applied practice. Working in groups of three, students use the virtual drone system to observe terrain, analyze landmarks, and select settlement sites, recording their choices and justifications on Google Earth 7.3.6.10201. This activity integrates spatial operations, geographic reasoning, and collaborative problem-solving to strengthen geographic literacy. Instructional materials were tailored for each group: shared content for all participants, virtual-operation aids for the experimental group, and paper-based aids for the control group. Grounded in Indigenous migration history, the learning sequence followed a thinking–understanding–applying–judging progression, designed to foster geographic reasoning and spatial interpretation skills (Figure 2).

In the experimental group, students used the virtual drone system to observe terrain and identify suitable settlement sites, switching between top-down and angled views to enhance spatial transformation and geographic analysis skills. The control group completed similar tasks using paper-based aids. The Google Earth activity was replaced with two printed maps: a near-range map showing landforms and settlements, and a long-range map highlighting watershed and ridgeline patterns. Based on the slides and prior discussion, students analyzed the strengths and weaknesses of potential sites and circled suitable expansion locations, providing justifications in a site selection task. Both groups received the same background knowledge and instructional guidance; the major difference was the mode of task operation during the second unit (

Table 1. Comparison of instructional design between the two groups.

Group	Experimental Group	Control Group
Pre-test	Assessing spatial and geographical concepts	Assessing spatial and geographical concepts
Basic concept instruction	Lecture: migration history and topographical features in the Smagus community	Lecture: migration history and topographical features in the Smagus community
Terrain marking task	Marking locations such as watersheds, water sources, transportation routes, and old villages on Google Earth.	Marking locations such as watersheds, water sources, transportation routes, and old villages on maps using markers.
Site selection activity	Students used the virtual drone system to explore terrains and selected suitable sites for tribal expansion.	Teachers distributed paper maps simulating the drone’s perspective. Students selected sites on maps and explained their reasoning.
Post-test and questionnaire	Evaluating learning effectiveness, learning motivation, cognitive load, and system satisfaction	Evaluating learning effectiveness, learning motivation, and cognitive load

). The instructional design enabled comparing the effects of using a virtual drone system versus traditional image-based operation on students’ spatial thinking and learning outcomes.

3.2.2. Research Instrument Design

The research instruments comprised two components: (1) learning achievement tests (pre-test and post-test) and (2) questionnaires assessing learning motivation, cognitive load, and learner experience. This mixed approach captured both cognitive and affective learning outcomes, enabling a more comprehensive evaluation.

(1)

Learning Effectiveness Assessment

The pre- and post-tests employed in this study were developed based on the instructional objectives and content of the curriculum. The purpose was to assess students’ acquisition of geographic knowledge and development of site-selection reasoning skills. A preliminary test version was administered, revealing that overall item difficulty was too low and several questions were not fully aligned with instructional objectives. Based on pilot results, feedback from geography teachers, and expert review, the research team evaluated the curricular context and instructional sequence, integrating Indigenous knowledge collected during fieldwork with community members. This iterative process produced the final test version, with revised items better reflecting the learning activities and students’ reasoning processes.

The final version of the achievement test consisted of 22 multiple-choice questions, with a total score of 110 points (converted into a percentage scale for subsequent analysis). The items covered four major thematic dimensions:

• Concepts and principles of site selection: Examining students’ understanding of settlement site-selection criteria and conditions.
• Migration history and cultural context: Focusing on the timeline, background, and cultural significance of tribal migration processes.
• Comparison of old and new settlement sites: Guiding students to analyze and compare the geographic conditions and suitability of various locations.
• Geographic knowledge and spatial concepts: Assessing students’ ability to interpret spatial features (e.g., river terraces, ridgelines, wind direction) and understand key geographic terminology.

To ensure alignment with instructional objectives and content validity, the test items were reviewed and revised by experts in the field of geography education. All items were phrased using student-accessible vocabulary to minimize the risk of semantic ambiguity or comprehension bias influencing test performance.

(2)

Learning Motivation, Cognitive Load, and Learning Experience Survey

Following the completion of instructional activities, the researchers administered a questionnaire survey consisting of three major dimensions: learning motivation, cognitive load, and learning experience, all measured using a five-point Likert scale [110]. Both groups responded to the sections on learning motivation and cognitive load, while the experimental group completed additional items evaluating system usability and satisfaction to examine the acceptance of the virtual drone system. The questionnaires included open-ended questions, inviting students to provide specific suggestions and feedback regarding the course content, task design, and system operation. The qualitative responses and qualitative feedback served as valuable references for subsequent curriculum refinement and instructional tool improvement.

3.3. System Development and Operation

To overcome the constraints of traditional geography instruction—particularly those related to time, space, and safety—this study developed a virtual drone system as an innovative learning tool for junior high school geography courses. The primary objective is to simulate the process through which students conduct aerial exploration and make site-selection decisions in authentic topographic environments, thereby fostering the development of spatial thinking and geographical reasoning skills. The system can be operated on mobile devices, enabling students to conduct virtual flights and participate in interactive learning activities that strengthen their sense of presence and engagement, while promoting experiential learning in geography education.

The system was designed to achieve high fidelity, low entry barriers, and strong interactivity, providing students with a safe, flexible, and contextually authentic learning environment that fosters a deeper understanding of landforms and human–environment interactions. This section presents the overall system design process, covering the development environment and tools, as well as the implementation procedures and key design considerations at each stage.

3.3.1. Design Process and Tools

The virtual drone system developed in this study provided students with a safe, low-cost environment for simulating drone flight operations. By integrating topographic observation and task-based learning, the system aims to strengthen students’ spatial thinking and geoscience literacy. The development process of the system comprised several stages, including needs analysis, material acquisition, model construction, interface design, and functional integration. Guided by instructional requirements, the research team defined the system’s core functions, which included drone flight simulation, camera perspective switching, topographic observation tasks, and flight control (e.g., takeoff, hovering, and landing). After that, fieldwork in the Smangus region of Hsinchu County, Taiwan, involved capturing high-resolution aerial imagery with a physical drone, which was subsequently processed for 3D modeling of the virtual scenes.

In this study, iTwin Capture Modeler 1.8 (Bentley Systems) was employed for 3D terrain reconstruction. The software conducted Structure-from-Motion (SfM) processing and image-based modeling to generate a three-dimensional point cloud and high-resolution terrain models. The completed models were exported in FBX format and imported into the Unity 2022 game engine, where C# programming was employed to implement an interactive user interface, integrate flight controls, and enable collision detection. This ensured that students could perform observation tasks within defined mission boundaries while maintaining the authenticity and integrity of the instructional context.

The 3D models of the virtual drone and its flight controller were created in Blender 4.0, incorporating components such as the fuselage, propellers, camera lens, and control buttons. Textures were created using Adobe Photoshop 23.0 to simulate realistic surface details and metallic gloss. The models were then exported as FBX files and integrated into Unity, together with the control scripts. The virtual drone system was exported as an APK for Android devices, enabling students to perform geographic exploration and learning activities within an immersive virtual environment.

3.3.2. System Development Process

The development process of the virtual drone system was divided into five stages: preliminary planning, material preparation, 3D terrain model reconstruction, system integration, and functional design. The following section outlines the implementation details and key considerations for each stage in sequence.

• Preliminary planning

At the initial stage of system development, the research team conducted a comprehensive technical feasibility analysis, evaluating widely used modeling and development tools for educational applications, including Unity (a cross-platform interactive development environment), Blender (a 3D modeling and animation tool), and iTwin Capture Modeler 1.8 (photogrammetry and 3D modeling software). Based on this evaluation, iTwin Capture Modeler was selected as the primary tool for terrain modeling to acquire authentic 3D landscape data. The terrain models were subsequently optimized and refined in Blender, where the drone appearance was also constructed, and ultimately exported into Unity as the main development platform for implementing user interaction, flight simulation, and task-oriented operation interfaces.

The research team conducted a comprehensive literature review and case study analysis of virtual drone systems implemented in educational contexts, both within Taiwan and internationally. This review enabled the identification of key instructional modules—such as free-flight exploration, landmark observation, target selection, and object delivery—and informed the design of the preliminary user workflow and operational logic, serving as the technical blueprint for subsequent development.

• Aerial imagery acquisition

To maintain geographic accuracy and cultural context, the research team collected aerial images of the Smangus region in early 2024. The site was selected as the research field in this study for its varied terrain and landforms—including ridgelines, catchment areas, river terraces, and core settlement zones—along with its rich history of Indigenous migration and settlement transformation, making it an ideal setting for geography instruction and place-based cultural education.

In this study, aerial imagery was captured using a DJI Mavic 2 Pro drone equipped with a high-resolution camera, following an automated flight path generated by Pix4Dcapture 4.11.0 (Figure 3). The aerial survey area covered approximately 1000 × 1000 m, with flight altitudes ranging from 100 to 120 m above ground. A multi-angle flight strategy, integrating top-down and oblique imagery, was adopted to capture settlement structures, roads, valleys, and hydrological features, with consecutive images overlapping 70–80% to enhance photogrammetry accuracy and 3D reconstruction.

Aerial imagery was acquired in sunny weather to reduce shadow interference and improve image clarity as well as texture recognition. Dynamic objects and human presence were deliberately avoided to minimize reconstruction errors and noise in the modeling process. The entire data collection spanned one full day and produced over 900 high-quality raw images, covering the Smangus region and its surrounding mountainous terrain. These image datasets served as the major source materials for creating the terrain model and instructional scenario, ensuring that the reconstructed environment retained a high degree of spatial fidelity and fine-grained texture details. Such authenticity is critical for reinforcing learners’ immersive experiences in terrain interpretation, site-selection reasoning, and geographic spatial cognition.

• Three-dimensional terrain model reconstruction:

Based on the aerial imagery collected from the Smangus region, this study employed iTwin Capture Modeler to create the terrain model and utilized Blender to design a virtual drone for integration into the virtual instructional environment. The modeling process comprised two major components: 3D terrain reconstruction and virtual drone modeling. To reconstruct the Smangus terrain model, over 900 high-resolution images captured by the DJI Mavic 2 Pro drone were imported into iTwin Capture Modeler for Structure-from-Motion (SfM) processing (Figure 4).

Through automated calibration, multi-view image stitching, and point cloud generation, the software transformed 2D aerial photographs into a 3D terrain model with textured surfaces (Figure 5). In the modeling process, the system automatically extracted GPS metadata embedded in the photographs to facilitate image alignment. The designer could adjust camera parameters or set Ground Control Points (GCPs) as needed to improve the geometric accuracy and fidelity of the reconstructed model.

• Virtual drone and controller models:

This study utilized Blender to create 3D models of the virtual drone and its controller. The drone body was modeled based on the appearance of commercially available civilian models, incorporating structures such as the fuselage, camera lens, propellers, and landing gear. Symmetrical components were modeled using the mirror function to reduce development time. The controller model was constructed by combining geometric primitives (e.g., cubes and cylinders) to represent buttons, joysticks, and shell recesses.

For texturing, UV unwrapping was applied to the model surfaces, and textures were designed to replicate the physical appearance of materials such as metal, plastic, and rubber. Glossiness, reflectance, and transparency were refined using the node editor (Figure 6). After completing both models, polygon optimization was performed, and the models were exported in FBX format and imported into the Unity game engine as visual assets for interactive objects and user interface design. Overall, the terrain and device models achieved a balance between realism and efficiency, providing a robust foundation for instructional applications.

• System integration and functional design:

The system was developed using the Unity game engine, integrating a 3D terrain model with a virtual drone to create an interactive learning environment (Figure 7). After importing the completed terrain model (FBX file from iTwin Capture Modeler) and the drone model (constructed in Blender), their scale and positioning were adjusted to match real-world proportions. Environmental lighting and a skybox were configured to simulate natural illumination, enhancing visual realism and immersion.

To enhance the drone flight experience, the Unity Rigid Body physics engine was employed to control drone motion and flight operations—including takeoff, hovering, turning, and landing—were implemented using C# programs. Additionally, a camera was integrated into the virtual drone, and the Render Texture function was employed to simulate aerial perspectives, allowing students to view terrain features from a first-person perspective. For user interaction, the system utilized a simplified control interface with touch-screen and viewpoint-switching functions, reducing the operational threshold. This design enabled students to perform drone simulations and observe geographic landscapes safely and cost-effectively within classroom settings.

• System optimization and deployment:

Following system integration, a series of performance optimization procedures was conducted to ensure stability across multiple devices and to support both classroom-based and independent learning. First, mesh simplification was applied to the 3D models, reducing polygon counts for both the terrain and drone to decrease computational workload and improve performance on mid- to low-end hardware. In addition, texture assets were compressed and converted into lightweight formats, and lightmap baking was employed to precompute lighting and shadow effects, thereby minimizing real-time rendering demands. After optimization, the system was packaged using Unity’s Build Settings and exported as an APK for installation on Android devices. This configuration provided a low-threshold operating environment. Designed for classroom instruction, the system supports both touchscreen and gamepad controls, enabling teachers to facilitate task-based exploration while allowing students to engage in independent practice, thereby enhancing geographic observation and spatial reasoning skills.

3.3.3. System Operation Guide

The virtual drone system was designed to simulate physical drone operation. Figure 8 highlights key control elements, including the virtual joysticks, takeoff/landing buttons, and perspective-switching options, to help readers better understand how students interacted with the system. In the virtual operating environment, A small window in the upper-right corner shows the drone’s front-facing camera feed in real time via the Render Texture function, allowing users to monitor the camera and closely replicate the authentic flight experience, including perspective changes during drone maneuvers.

When the virtual drone exceeds the preset altitude limit, the system automatically switches to a first-person perspective, showing the flight path over the terrain model by iTwin Capture Modeler. This design not only provides an intuitive operational experience but also enhances the realism of terrain visualization, improving learners’ spatial perception and learning outcomes. Users can freely control the drone to observe landscapes and terrain features during settlement site selection (Figure 9).

3.4. Research Design

This study employed a quasi-experimental, nonequivalent pre-test post-test design to examine the effects of a virtual drone system on junior high school geography education. Specifically, it investigated the system’s impact on students’ learning outcomes, learning motivation, and cognitive load. Teacher feedback was collected to evaluate the feasibility of classroom implementation and identify potential areas for improvement. The research design is illustrated in Figure 10, detailing the independent variable, dependent variables, and control variables, as described below.

(1)

Independent Variable: The independent variable in this study is the difference in teaching media, with students randomly assigned to two groups.

• Experimental group: Students engaged in immersive and interactive learning through the virtual drone system, integrated with task-based instructional activities. Learners operated the virtual drone from a first-person perspective to explore terrain features and conduct settlement site selection simulations.
• Control group: Students received traditional classroom instruction using printed Google Earth images and PowerPoint lectures. They analyzed terrain features and settlement site selection by examining static images and paper maps.

(2)

Dependent Variables: To comprehensively evaluate teaching effectiveness, this study established the following three categories of dependent variables:

• Learning outcomes: A researcher-developed Settlement Site Selection Achievement Test was used to assess learning gains in terrain interpretation, human–environment interaction reasoning, and geographic concept application. Pre- and post-test scores were compared through difference analysis.
• Learning motivation: Measured using a five-point Likert-scale questionnaire, this variable assessed students’ interest, engagement, and initiative toward course content, and learning activities. The analysis compared overall as well as intrinsic and extrinsic motivational components between groups.
• Cognitive load: Evaluated using an expert-validated cognitive load questionnaire, measuring students’ perceived mental load, mental effort, and overall task complexity during the learning activities.

(3)

Control Variables: To eliminate interference from non-instructional medium variables, the following conditions were controlled for consistency in the research design:

• Instructor: A researcher served as the instructor for both groups to ensure consistency in teaching style and instructional delivery.
• Instructional time: Both groups received equal instructional time and followed the same schedule for the intervention.
• Instructional content: Both groups received instruction on the same core concepts and activities related to settlement site selection, differing only in instructional medium and delivery approach.

3.5. Data Process and Statistical Analysis

The data collected in this study primarily consisted of paper-based test scores administered before and after the instructional intervention, as well as structured questionnaire responses gathered after completing the course. To ensure the integration of data sources and the accuracy of subsequent statistical analyses, the researchers followed established research requirements and statistical standards, conducting the following five procedures for data processing and analysis:

(1)

Evaluation of Learning Outcomes.

The learning achievement test in this study was administered in paper-and-pencil format and comprised 22 items, with a maximum score of 110 points. To standardize the scores and enable comparisons across groups, raw scores were converted to percentages using the formula: raw score × (100 ÷ 110), resulting in a 100-point scale. Pre-test and post-test scores were then recorded separately and organized by student identification number to ensure accurate matching, facilitating subsequent statistical analyses.

(2)

Data Consolidation and Standardization.

To enable comprehensive data organization and cross-variable analysis, datasets from multiple sources—including learning achievement tests, student backgrounds, and post-course questionnaires—were integrated. The integration process used the student identification number as the key variable, merging all data longitudinally. Variable naming conventions were standardized in accordance with SPSS 30 technical requirements: names were assigned primarily in English, limited to eight characters, and restricted from using special symbols or spaces. For instance, the gender variable was coded as Gender, map usage experience as Map Use, and items under the learning motivation scale were labeled as ExtMot_1 (Extrinsic Motivation, Item 1), IntMot_2 (Intrinsic Motivation, Item 2) to ensure consistency for subsequent subscale aggregation and analysis.

(3)

Questionnaire Data Coding and Labeling.

For the questionnaire data, all items were coded using a five-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree). All collected responses were converted into numerical data to facilitate statistical analysis. In addition, for the purpose of group comparison, a new variable, Group, was created in the dataset to indicate each student’s assigned instructional condition, with the control group coded as 0 and the experimental group coded as 1. This variable served as the primary categorical factor for subsequent difference testing and the analysis of covariance (ANCOVA). A self-constructed Variable Coding Sheet was developed to match variable names with questionnaire items, specifying each item’s code, description, construct, and scoring scheme. After data entry, the dataset was checked for errors such as missing values, duplicates, or out-of-range entries to ensure its correctness and accuracy.

(4)

Learning Outcomes and Learner Responses.

Experimental data were analyzed in SPSS, with the analytical strategy—guided by data type and research questions—employing the following statistical methods:

• Learning effectiveness analysis:

Descriptive statistics were used to present the means and standard deviations of the pre-test and post-test scores. An independent samples t-test was conducted on the pre-test scores to examine group homogeneity and confirm the absence of significant baseline differences. For the main evaluation, an analysis of covariance was conducted, using post-test scores as the dependent variable, pre-test scores as the covariate, and group (experimental vs. control) as the independent variable, to assess the effect of the instructional intervention on learning outcomes.

• Learning motivation and cognitive load analysis:

For the components of learning motivation and cognitive load, the means and standard deviations of intrinsic motivation and extrinsic motivation, as well as mental load and mental effort, were calculated, respectively. Independent samples t-tests were conducted to compare the differences between the experimental and control groups.

• User experience and system satisfaction analysis:

For the component of system satisfaction, which was evaluated only by students in the experimental group, descriptive statistics (means and standard deviations) were computed for each dimension and individual items. This analysis aims to examine students’ overall perceptions and acceptance of the virtual drone system, particularly in terms of operational convenience, interactivity, and task guidance.

The data processing and analysis followed rigorous procedures for dataset construction and transformation. Standardized variable naming, questionnaire coding, and group labeling ensured proper alignment and integration of all data. Descriptive statistics summarized patterns in learning outcomes and questionnaire responses, while inferential analyses—including independent samples t-tests and ANCOVA—examined the instructional intervention’s effects on cognitive performance, learning motivation, and system experience. This multidimensional evaluation confirmed the instructional effectiveness and usability of the virtual drone system and provided empirical evidence to guide future applications in geography education using VR technology.

4. Results and Discussion

This section examines the effects of the virtual drone system on geographic learning, focusing on students’ learning outcomes, motivation, cognitive load, and learning experience across instructional contexts. Adopting a quasi-experimental design, students were allocated to experimental and control groups, and their performance was evaluated using pre- and post-tests alongside questionnaire feedback. The analysis is organized into four parts. Section 4.1 compares the learning outcomes of the two groups in the settlement-site selection unit, using t-tests and ANCOVA to assess significant differences. Section 4.2 presents a quantitative analysis of learning motivation and cognitive load to evaluate the system’s impact on learners’ psychological responses. Section 4.3 reports the experimental group’s subjective feedback on system satisfaction and learning experience. Finally, Section 4.4 integrates the research findings to discuss broader implications and provide guidance for future instructional practice and research directions.

4.1. Learning Effectiveness Analysis

To investigate the effectiveness of the virtual drone system in geography instruction, this study employed a paper-and-pencil test to evaluate students’ learning outcomes in the Settlement Site Selection unit. The test consisted of 22 multiple-choice items, divided into two domains: the first part (12 items) focused on Indigenous historical and cultural contexts, including migration processes, settlement distribution, and geographical relationships; the second part (10 items) emphasized core geographical concepts such as landform interpretation, hydrological features, and spatial thinking. Each test item was worth five points, yielding a total of 110 points, which were subsequently converted to a standardized 100-point scale for analysis and visualization.

Both groups completed the same achievement test before and after the intervention. The analysis compared mean scores, standard deviations, and improvements from pre-test to post-test across groups. Paired-samples t-tests were conducted to examine within-group changes, while ANCOVA was applied to control for pre-test effects and assess the statistical significance of post-test performance. This section provides a detailed presentation of the results, using statistical tables to illustrate differences in learning outcomes between groups, followed by a discussion of the instructional factors, learning processes, and system characteristics that may have contributed to these differences, providing insights for future instructional design and research development.

4.1.1. Learning Outcome Analysis

To assess the impact of the virtual drone system on students’ geography learning outcomes, this study first compared the experimental group’s pre-test and post-test scores (

Table 2. Descriptive statistics of learning outcomes for both groups.

Group	Pre-Test		Post-Test
	Mean	SD	Mean	SD
Experimental Group	55.70	13.71	91.45	12.53
Control Group	46.87	12.23	72.51	28.42

). The mean pre-test score of the experimental group was 55.70 (SD = 13.71), which increased significantly to 91.45 in the post-test score (SD = 12.53), yielding an overall mean improvement of 35.75 points. This substantial gain indicates that students made notable progress after using the virtual drone system, particularly in their understanding of Indigenous culture, interpretation of geographical concepts, and integrated application of settlement site selection. While pre-test scores were distributed in the lower–middle range, post-test scores clustered at higher levels with reduced variability, reflecting both improved learning outcomes and greater consistency across students.

To further examine the learning effectiveness after the interventions, the pre-test and post-test results of the control group were analyzed. The control group obtained a mean pre-test score of 46.87 (SD = 12.23), which increased to 72.51 in the post-test (SD = 28.42), showing an average gain of 25.64 points. The findings suggest that students demonstrated gains in settlement site selection and spatial interpretation skills even under traditional instruction, without the use of the virtual drone system. In contrast, the experimental group showed a larger improvement with lower post-test variability, whereas the control group exhibited greater dispersion in post-test scores, indicating that traditional instruction produced more uneven outcomes among students.

To evaluate the impact of individual instructional methods, paired-samples t-tests were employed to compare the changes between the pre- and post-tests within each group. The paired-samples t-test results (

Table 3. Paired-samples t-test results of learning outcomes for both groups.

Group	Mean	SD	t	p
Experimental Group	35.75	14.83	15.81	0.000 ***
Control Group	25.65	25.90	6.41	0.000 ***

*** p < 0.001.

) highlight the relative advantage of the virtual drone system in enhancing learning outcomes while stabilizing the group performance. Students in the experimental group, after receiving instruction using the virtual drone system, demonstrated an average improvement of 35.75 points (SD = 14.83), with a t-value of 15.81, reaching a highly significant level (p < 0.001). The findings suggest that the virtual drone system significantly facilitated students’ mastery of knowledge and the integration of geographic concepts during settlement site selection tasks.

The control group, which received traditional instruction, also achieved significant progress (t = 6.41, p < 0.001), but with a smaller mean improvement of 25.65 points and less consistent learning gains among students (SD = 25.90). Compared with traditional instruction, these results suggest that the virtual drone system not only enhanced overall mastery of geographic concepts but also reduced variability in learning outcomes, likely due to its interactive and immersive design. This suggests that the virtual drone system can serve as an effective tool in geography education, especially for complex curricula that integrate cultural context, spatial interpretation, and decision-making skills.

4.1.2. ANCOVA Analysis

To further compare learning effectiveness between the two groups, an ANCOVA was conducted to assess the effect of the intervention on students’ learning outcomes (

Table 4. ANCOVA results of learning effectiveness for the experimental and control groups.

Source	Type III Sum of Squares	df	F	p	η2
Pre-test	5727.862	1	13.964	0.000 ***	0.147
Group	2561.806	1	6.245	0.014 *	0.072
Group × Pre-test	1341.49	1	3.270	0.074	0.039
Error	33,225.56	81
Total	620,157.59	85

* p < 0.05, *** p < 0.001.

). Pre-test scores were used as the covariate, post-test scores served as the dependent variable, and comparisons were made between the experimental and control groups.

ANCOVA results showed that pre-test scores significantly influenced post-test performance (F = 13.964, p < 0.001). After controlling for this effect, the group difference remained significant (F = 6.245, p = 0.014), indicating that the experimental group outperformed the control group. These results verify the virtual drone system’s independent contribution to enhancing student achievement. In addition, the interaction between group and pre-test was not significant (F = 3.270, p = 0.074 > 0.05), confirming the assumption of homogeneous regression slopes across groups. This result is critical for the validity of the ANCOVA, as it indicates that the effectiveness of the virtual drone system was consistent regardless of students’ prior ability.

From the perspective of effect size, the pre-test variable exerted a medium-to-large influence (η² = 0.147), while the group variable showed a moderate effect (η² = 0.072). These findings indicate that, beyond students’ initial ability levels, the intervention itself played a critical role in determining learning achievement. This analysis supports the earlier t-test results, confirming from multiple statistical perspectives that the virtual drone system effectively enhanced learning performance. In summary, the ANCOVA results indicate that the virtual drone system significantly improved students’ learning outcomes, even after controlling for prior abilities, highlighting both its statistical significance and practical relevance in geography education.

4.2. Learning Motivation Analysis

Learning motivation is a critical psychological factor influencing students’ engagement and learning outcomes, particularly in context-based and task-oriented instructional designs, where its role becomes even more salient. Based on Self-Determination Theory, this study differentiates intrinsic and extrinsic motivation to investigate the effects of instructional approaches on students’ motivation [85]. A self-developed learning motivation scale served as the measurement tool. It comprised six items, with three assessing extrinsic motivation and three assessing intrinsic motivation. All items were rated on a five-point Likert scale, ranging from strongly disagree (score = 1) to strongly agree (score = 5).

4.2.1. Overall Learning Motivation Analysis

To compare the differences in learning motivation between the experimental and control groups, an independent samples t-test was conducted to examine the impact of different instructional approaches (

Table 5. Independent samples t-test results of learning motivation for both groups.

Group	N	Mean	SD	df	t	p
Experimental Group	43	4.395	0.712	83	1.397	0.166
Control Group	42	4.175	0.745

). The experimental group obtained a higher mean score (mean = 4.395, SD = 0.712) on the overall learning motivation scale compared to the control group (mean = 4.175, SD = 0.745). However, this difference was not statistically significant (t = 1.397, p = 0.166 > 0.05). Although students in the experimental group showed slightly higher levels of learning motivation, the overall results indicate no significant difference in motivation between the two instructional approaches.

The results suggest that both groups already possessed a certain level of interest and engagement in the topic of settlement site selection. Moreover, since the course design incorporated authentic contexts, task-oriented learning, and cultural narratives, it was inherently appealing and meaningful. Even without the support of technological tools, the instructional design itself provided sufficient incentives for participation, explaining why no significant difference in overall motivation was observed between groups. In other words, the curriculum design alone offered adequate stimuli to foster student engagement. Nevertheless, from the perspective of Self-Determination Theory, different instructional approaches are more likely to generate differentiated impacts on intrinsic and extrinsic motivation. For example, immersive and interactive systems can increase enjoyment and the sense of challenge, while immediate feedback and visualized outcomes may enhance extrinsic motivation through performance pursuit or achievement display.

Therefore, analyzing only the overall motivation scores may not be sufficient to uncover the potential benefits of intervention. To address this issue, the present study further examined students’ performance on the two sub-dimensions—intrinsic motivation and extrinsic motivation—to determine whether the virtual drone system demonstrated more significant effects in specific aspects of motivation. These findings may also provide important implications for optimizing instructional design.

4.2.2. Dimensional Motivation Analysis

To further investigate whether the virtual drone system affected different types of learning motivation, the study compared the two groups on extrinsic and intrinsic motivation (

Table 6. Independent-samples t-test results of dimensional motivation between groups.

Group	Extrinsic Motivation			Intrinsic Motivation
	Mean	SD	p	Mean	SD	p
Experimental Group	4.33	0.873	0.624	4.46	0.716	0.037
Control Group	4.25	0.758		4.10	0.818

). For extrinsic motivation, the experimental group obtained an average score of 4.33 (SD = 0.873), while the control group scored 4.25 (SD = 0.758). The difference between the two groups was not statistically significant (p = 0.624 > 0.05). This result indicates that, regardless of whether a virtual drone system was used, students demonstrated a consistent level of emphasis on performance outcomes. In contrast, for intrinsic motivation, the experimental group reported an average score of 4.46 (SD = 0.716), which was significantly higher than the control group’s score of 4.10 (SD = 0.818), reaching statistical significance (p = 0.037 < 0.05).

The virtual drone system developed in this study features high interactivity, task orientation, and real-time feedback, enabling students to engage in missions, control flight operations, and observe terrain from a first-person perspective. The instructional design creates a learning experience that is both challenging and contextually immersive. Such characteristics align closely with the three basic psychological needs required in Self-Determination Theory—competence, autonomy, and relatedness—and thereby contribute to the stimulation and reinforcement of intrinsic motivation.

Questionnaire items such as “I feel happy when learning in this course” and “I enjoy interactive materials that are challenging” indicate that students’ engagement stemmed from genuine interest and identification with the learning content, rather than being solely driven by performance outcomes. While both instructional approaches had a similar effect on extrinsic motivation, the virtual drone system showed a significant advantage in enhancing intrinsic motivation. This finding emphasizes the impact of technology-assisted, context-based learning on students’ interest and engagement, providing valuable insights for future curriculum design and instructional strategies.

4.3. Cognitive Load Analysis

Cognitive load refers to the mental burden and the degree of working memory utilization that learners experience when processing learning content. According to Cognitive Load Theory, proposed by Sweller et al. [93], complex instructional design, excessive contextual distractions, or inappropriate time allocation may result in higher cognitive load, thereby reducing learning efficiency and outcomes. Based on this framework, this study developed a cognitive load scale for evaluating the cognitive load in both groups.

The questionnaire comprised eight items rated on a five-point Likert scale, divided into two sub-dimensions: (1) Mental Load (five items), reflecting a combination of intrinsic and extraneous load in CLT, including items such as “The learning content in this activity was difficult for me” and “The learning content was frustrating”; and (2) Mental Effort (three items), representing the amount of cognitive resources invested to complete the task, corresponding to germane load, with items such as “I had to put in a lot of effort to complete this learning activity.” This study compared overall cognitive load and its sub-dimensions between the experimental and control groups to examine whether the virtual drone system facilitates learning without imposing additional mental burden.

4.3.1. Overall Cognitive Load Analysis

To assess whether the virtual drone system increased learners’ cognitive burden, this study compared overall cognitive load scores between the experimental and control groups. The eight-item scale measured difficulty understanding content, frustration, anxiety, and mental effort, capturing students’ subjective cognitive load during the course (

Table 7. Independent samples t-test results of cognitive load between groups.

Group	N	Mean	SD	df	t	p
Experimental Group	43	2.795	1.355	83	0.106	0.916
Control Group	42	2.823	1.092

). The experimental group had a mean score of 2.795 (SD = 1.355), while the control group scored 2.823 (SD = 1.092). The difference between the two groups was negligible and not statistically significant (t = 0.106, p = 0.916). These results indicate that incorporating the virtual drone system into the learning task of settlement site selection did not impose additional cognitive load on students, suggesting that the system has potential instructional value while maintaining a stable level of learning burden.

These findings have significant practical and design implications. Cognitive Load Theory [93] suggests that inadequately designed technology-enhanced instruction may increase extraneous load, hindering effective cognitive resource allocation. Nonetheless, the experimental group, despite engaging in more interactive operations and scene exploration through virtual drone use, did not experience higher cognitive load than the control group. Possible explanations include:

• Intuitive interface and simple operations: The system’s clear operational flow and immediate feedback minimized additional extraneous load.
• Clear task structure: The task-oriented nature of learning activities guided students’ focus toward learning goals, thereby reducing distractions.
• Immersive experience: By engaging in first-person exploration and decision-making simulations, students directed their cognitive resources toward meaningful learning tasks without increasing mental workload.

4.3.2. Dimensional Cognitive Load Analysis

Cognitive load effects of instructional approaches were further examined using the Mental Load and Mental Effort dimensions [93]. The experimental group scored 2.698 (SD = 1.352) on Mental Load versus 2.710 (SD = 1.124) for the control group (p = 0.965), and 2.892 (SD = 1.416) on Mental Effort versus 2.937 (SD = 1.194) for the control group (p = 0.875), indicating no significant differences (

Table 8. Independent-samples t-test results of mental load and mental effort between groups.

Group	Mental Load			Mental Effort
	Mean	SD	p	Mean	SD	p
Experimental Group	2.698	1.352	0.965	2.892	1.416	0.875
Control Group	2.710	1.124		2.937	1.194

). These results suggest that the virtual drone system did not impose additional cognitive burden on students.

These findings support the feasibility and usability of technology-enhanced learning and demonstrate that immersive learning environments, when paired with well-structured instructional tasks and intuitive interface design, can avoid imposing excessive extraneous load while maintaining an appropriate level of learner engagement. The result is consistent with Cognitive Load Theory [92], which emphasizes that effective instructional design should minimize unnecessary extraneous load and guide learners to allocate their limited cognitive resources toward managing intrinsic load and fostering the germane load of the learning task. The virtual drone system clearly demonstrates such potential, and future development may further incorporate diverse inducement strategies to foster more in-depth cognitive engagement and knowledge construction.

4.4. System Satisfaction Analysis

A system satisfaction scale was developed to evaluate students’ experiences with and acceptance of the virtual drone system. This scale assessed students’ subjective perceptions and behavioral intentions across three dimensions: Perceived Usefulness, Perceived Ease of Use, and Behavioral Intention. The dimensions were derived from the Technology Acceptance Model (TAM) proposed by Davis [111]. Widely applied in educational technology research, TAM is appropriate for assessing students’ initial experiences with novel instructional tools such as the virtual drone system. The scale included 12 items, each rated on a five-point Likert scale, and was organized as follows:

• Perceived Usefulness (Items 1–5): Assesses whether students regard the virtual drone system as beneficial for enhancing their understanding of instructional content and improving learning effectiveness, particularly in mastering geographic concepts such as landforms, hydrology, and settlement site selection.
• Perceived Ease of Use (Items 6–9): Evaluates whether students perceive the system’s interface as intuitive, its interactions as smooth, and whether learning tasks and exploratory activities can be completed easily without technical obstacles.
• Behavioral Intention (Items 10–12): Measures students’ willingness to use the system in future learning and their acceptance of this instructional approach.

This scale captures students’ perceptions of technology acceptance to evaluate the system’s learning support, usability, and potential for broader application. Descriptive statistical analyses of survey data provide empirical evidence to guide system refinement, curricular adjustments, and instructional practice.

4.4.1. Perceived Usefulness

The Perceived Usefulness dimension aimed to evaluate whether students perceived the virtual drone system as beneficial for learning geographic concepts, understanding settlement site selection, and making environmental judgments. According to the results in

Table 9. Descriptive statistical results of Perceived Usefulness.

Dimension	Evaluation Item	Mean	SD
Perceived Usefulness (mean = 4.20, SD = 0.832)	1. I think the virtual drone interactive system helps me understand the factors affecting tribal settlement selection and terrain.	4.15	0.866
	2. I think the virtual drone interactive system enriches my learning experience of river terraces, watersheds, and terrain barriers.	4.19	0.824
	3. I think the virtual drone interactive system makes it easier to understand the geographical factors behind tribal migration.	4.20	0.828
	4. I think the virtual drone interactive system helps provide more useful information about terrain and environmental decision-making.	4.21	0.846
	5. I think learning geography concepts through the virtual drone system is more helpful than using traditional slides or lectures.	4.27	0.793

, the overall mean score for Perceived Usefulness was 4.20 with a standard deviation of 0.832, indicating a relatively high level of agreement. This suggests that most students affirmed the system’s helpfulness in their learning process.

Item-level analysis indicated that students perceived the virtual drone system as helpful for understanding geographic conditions and topographical factors influencing settlement site selection (mean = 4.15). They reported that operating the system provided concrete experiences of complex landforms such as river terraces, watersheds, and natural barriers (mean = 4.19), and that it clarified abstract migration processes and geographical cause–effect relationships (mean = 4.20). These outcomes supported the development of human–environment interaction and contextual reasoning skills. Notably, the highest-rated item was “Learning geographic concepts through the virtual drone system is more helpful than traditional slides or lectures” (mean = 4.27), indicating strong student endorsement of immersive, application-oriented instruction.

Overall, the virtual drone system demonstrated a high level of Perceived Usefulness in instructional applications, particularly in helping students master spatial-geographic knowledge, visualize abstract landforms, enhance geographic reasoning, and increase learning engagement. This finding aligns with the major predictions of the TAM, which posits that learners are more inclined to actively engage with and continue using the virtual drone system if they perceive it as beneficial for learning geographic concepts.

4.4.2. Perceived Ease of Use

The Perceived Ease of Use dimension assessed the intuitiveness of system operation, smoothness of interaction, and ease of completing tasks. As a core TAM predictor, ease of use promotes positive attitudes and the intention to continue using the system. As shown in

Table 10. Descriptive statistical results of Perceived Ease of Use.

Dimension	Evaluation Item	Mean	SD
Perceived Ease of Use (mean = 4.20, SD = 0.932)	1. The interface of the virtual drone system is easy to use.	4.19	0.994
	2. I quickly learned how to operate the virtual drone system.	4.15	0.838
	3. I can easily use the interactive materials to complete the tasks in the course.	4.20	0.884
	4. The learning activities using the virtual drone system are easy to understand.	4.25	1.011

, the mean score was 4.20 (SD = 0.932), indicating consistently high responses. Students found the system straightforward, interactive, and low in cognitive strain, suggesting that the instructional design minimized learning barriers and allowed focus on geographic content. The consistent responses further highlight the system’s reliability and potential for scalable classroom application.

Item-level analysis indicated that most students found the system interface easy to use (mean = 4.19) and they quickly learned to operate the system (mean = 4.15), reflecting a gentle learning curve suitable for classroom application. Students also reported that they could easily complete interactive tasks (mean = 4.20) and that the overall learning activity was simple and clear (mean = 4.25, the highest-rated item in this dimension). These findings suggest that the task-oriented design and visualized operation process of the virtual drone system effectively minimized difficulties, lowered learning barriers, and facilitated a seamless, engaging learning experience.

According to the TAM framework, a positive user experience enhances students’ acceptance of the system and their intention to use it in the future. The virtual drone system demonstrated strong performance in perceived ease of use, characterized by a low operational threshold and user-friendly interaction. These features effectively support students in smoothly engaging with geographic tasks and constructing knowledge. The following section examines students’ Behavioral Intention to present a more complete picture of their potential inclination toward continued system usage.

4.4.3. Behavioral Intention

Behavioral Intention reflects students’ willingness to continue using the virtual drone system and their motivation for self-directed learning, highlighting the system’s practical implications for future instructional use. As shown in

Table 11. Descriptive statistical results of Behavioral Intention.

Dimension	Evaluation Item	Mean	SD
Behavioral Intention (mean = 4.12, SD = 0.876)	1. When I need to learn related content in the future, I am willing to use this instructional material.	4.12	0.837
	2. I am willing to learn using this teaching method and material.	4.12	0.851
	3. I will take the initiative to use this material to learn the course.	4.11	0.939

, this dimension comprised three items, with an overall mean score of 4.12 and a standard deviation of 0.876, indicating a relatively high level of intention. The results suggest that most students held an open attitude toward using the system in future learning and exhibited a certain degree of continuity in learning motivation.

Item-level analysis further revealed that students generally indicated, “When I need to learn related content in the future, I am willing to use this instructional material” (mean = 4.12), reflecting the extensibility of the virtual drone system and its potential to generate positive learning effects in other instructional contexts. For the items “I am willing to learn using this teaching method and material” (mean = 4.12) and “I will take the initiative to use this material to learn the course” (mean = 4.11), students likewise revealed positive attitudes and a potential willingness to engage with the system proactively.

Collectively, the virtual drone system developed in this study demonstrates substantial potential for practical application and future extension. Based on the overall results across the three dimensions, the mean score of the scale was 4.18 (SD = 0.88), indicating that most students held a highly positive evaluation of the system. Students not only affirmed its supportive value in the learning process but also expressed willingness to continue using the system and recommend it to others.

4.5. Thematic Analysis

A thematic analysis of students’ open-ended responses revealed three dominant themes that further illuminate the quantitative findings. Many students emphasized that the virtual drone system enhanced their comprehension of geographic and environmental relationships, particularly the spatial patterns of tribal settlement, river terraces, and watershed formation. They described the 3D visualization and drone-flight perspectives as helping them “see the terrain more clearly” and “connect physical features with settlement decisions,” demonstrating that immersive visualization fostered contextual reasoning about human–environment interactions. Some students expressed high engagement and satisfaction with the system’s interactivity and intuitive design. They noted that the virtual joysticks, perspective-switching functions, and task-based missions made learning “more interesting and easier to understand than lectures.”

A smaller subset of responses reflected minor challenges, such as occasional difficulty in mastering drone flight control or maintaining spatial orientation during aerial navigation. Despite these issues, most students viewed such challenges as opportunities to improve their control skills and deepen spatial awareness. The qualitative findings reinforced the quantitative results, illustrating that students not only perceived the virtual drone system as useful and easy to use but also valued its immersive and exploratory learning environment. The system’s combination of experiential interaction and visual realism contributed to stronger conceptual understanding, higher motivation, and a sustained willingness to engage in self-directed geographic learning.

The teachers’ feedback additionally reinforced these findings, emphasizing the system’s instructional value and classroom applicability. They reported that the virtual drone system enriched geography lessons by providing vivid, context-rich representations that made landform and cultural concepts more tangible, thereby fostering greater student curiosity and collaboration. Teachers found it particularly effective for promoting inquiry-based and experiential learning compared with traditional lectures. However, they also noted challenges such as variations in students’ control skills, the need for initial technical familiarization, and management difficulties in larger classes. Despite these issues, most teachers agreed that the system holds strong potential for curriculum integration when supported by structured guidance and pre-lesson orientation.

5. Conclusions and Future Work

The virtual drone system demonstrated high instructional value and strong practical applicability in junior high school geography education, particularly in the context of settlement site selection. Students showed improvements in learning outcomes, intrinsic motivation, and system satisfaction, while maintaining a stable level of cognitive load, indicating that the system design effectively balances usability with instructional quality. Moreover, by embedding local contexts and cultural narratives, the system enhanced students’ understanding of geographical space and socio-cultural connections, fostering interdisciplinary integration and a stronger sense of place. However, limitations still remain regarding user familiarity, hardware constraints, and viewpoint control functions. Future improvements may focus on interactive learning design, providing step-by-step guidance and cross-platform deployment to further strengthen instructional outcomes and promote broader adoption in geography education.

5.1. Research Findings

This study addressed a critical gap in junior high school geography education, where traditional methods often inadequately support spatial reasoning in settlement site selection. To address this challenge, a virtual drone system was developed and evaluated using a quasi-experimental design, examining its effects on students’ learning outcomes, motivation, cognitive load, and system satisfaction. Based on quantitative results and student feedback, the study yielded the following conclusions:

(1)

learning outcomes and spatial reasoning

The experimental group that used the virtual drone system outperformed the control group on post-tests of settlement site selection, with ANCOVA results confirming statistical significance (p = 0.014). The system’s aerial perspective and task-based simulations effectively connected abstract concepts to real landscapes, enhancing knowledge transfer, spatial reasoning, and decision-making skills.

(2)

Improvement in intrinsic motivation

While overall learning motivation showed no significant group-level difference, experimental group students scored significantly higher in intrinsic motivation (p = 0.037). Immersive and interactive operations—including terrain exploration, aerial imaging, and spatial annotation—provided task-based challenges to enhance intrinsic motivation, reinforced learning achievement, and sustained student engagement.

(3)

Stable cognitive load with high usability

Despite engaging in multiple operations—including flight simulation, terrain observation, and task execution—the experimental group reported no significant increase in cognitive load compared to the control group. This suggests that the optimized and intuitive interface successfully reduced extraneous load, allowing students to concentrate on geographic exploration and reasoning without being hindered by technical complexity.

(4)

High system satisfaction and adaptability

Students expressed highly positive attitudes toward the virtual drone system, with mean scores of 4.20 in perceived usefulness, 4.20 in perceived ease of use, and 4.12 in behavioral intention, yielding an overall satisfaction score of 4.18. The system’s compatibility with mobile devices, VR support, and capability to export recorded task data demonstrate its strong adaptability across diverse instructional contexts.

(5)

Integration of local cultural context

By embedding tasks within the Indigenous context, the system linked migration histories and settlement decision-making with geographic knowledge, deepening students’ understanding of landscape–society interactions. This place-based design enhanced authenticity and cultural awareness while promoting interdisciplinary competencies and civic-mindedness, aligning with the primary goals of geography education.

5.2. Suggestions for System Improvement

Although the virtual drone system demonstrates significant potential for geography education, its effectiveness is constrained by user familiarity, hardware requirements, and limitations in visual perspective manipulation. Future improvements could focus on enhanced operational guidance, cross-platform compatibility, and multi-view functionality to optimize system performance.

(1)

Guidance for user interface operations.

Despite the simplified workflow and optimized user interface, some students—especially those with limited digital experience or weaker spatial skills—still faced difficulties with flight control, image recording, and task navigation during their initial use, indicating a learning barrier. Future developments should incorporate instructional videos, practice simulations, and step-by-step prompts to scaffold operational skills while reducing teachers’ immediate instructional burden.

(2)

Hardware constraints in system scalability.

The current system requires Android tablets or smartphones with sufficient computational capacity to support flight simulation and real-time interactive tasks. However, financial constraints, maintenance, and limited device availability in remote schools may hinder widespread adoption and restrict students’ post-class exploration. Future development should consider cross-platform compatibility or simplified modes for lower-end devices to improve accessibility and scalability.

(3)

Perspective and observation functions.

Currently, the system provides only a single viewpoint and fixed camera distance, offering a basic aerial experience with limited flexibility for detailed terrain observation, comparative analysis, and advanced interpretation. Students cannot freely change perspectives, adjust focal length, or zoom, restricting the development of three-dimensional spatial understanding. Future enhancements should incorporate multi-view switching, camera zoom, annotation tools, and recording functions to increase the system’s instructional value and applicability in geography education.

5.3. Pedagogical Implications

Based on the implementation experience and empirical findings of this study, future instructional design may benefit from promoting interdisciplinary integration and collaborative lesson planning to maximize the effectiveness of virtual teaching tools. Two specific pedagogical recommendations are proposed:

(1)

Promoting interdisciplinary integration.

Given its high modularity and extensibility, the virtual drone system can be applied to interdisciplinary curricula. Natural science courses could integrate geology, hydrology, and climate exploration; history classes could connect with Indigenous migration and settlement; and information technology courses could incorporate spatial data analysis and simulation tasks. Developing teacher professional learning communities and in-service training programs would familiarize educators with virtual drone operation and its integration into pedagogy. Such collaborative efforts would allow adaptation to diverse grade levels, student abilities, and curriculum requirements, enhancing classroom applicability and long-term sustainability.

(2)

Interactive design and learning feedback.

To enhance learner autonomy and provide timely feedback, the system should integrate real-time interactive features and guidance mechanisms, such as audio or visual prompts, to support task navigation. In addition, delivering individualized text- or image-based feedback based on students’ interaction logs and site-selection outcomes can promote self-correction and reflection to enhance learning outcomes. These improvements would reduce teachers’ monitoring demands while fostering students’ sense of control, engagement, and exploration, thereby supporting deeper geographic reasoning and developing effective and self-directed learning strategies.

5.4. Future Research Directions

To further increase the instructional applicability and research value of the virtual drone system in geography education, future studies could broaden its scope and refine methodological designs in the following directions:

(1)

Expanding sample size and observation time.

This study was conducted with a single grade level at one school, providing initial evidence of the system’s effectiveness and questionnaire responses. However, the limited sample size and short instructional duration restrict the generalizability of the findings. Future research should broaden the sample to include multiple schools across diverse regions, grade levels, and student demographics to enhance representativeness and external validity. Additionally, extending the intervention period and incorporating delayed post-tests or longitudinal assessments would enable evaluation of the system’s long-term effects on knowledge retention, learning transfer, and the internalization of spatial concepts, which may not be fully captured in short-term studies.

(2)

Qualitative and behavioral analysis.

While achievement tests and quantitative questionnaires provide valuable insights into students’ learning performance and perspectives, they cannot fully capture the detailed dynamics of cognitive processes and strategies employed during learning. Future research could incorporate in-depth interviews, student reflective portfolios, and classroom observations to uncover changes in reasoning, strategy adjustment, and decision-making. Combining system usage logs with screen-tracking analyses could further quantify behaviors such as terrain exploration, site-selection decisions, and task interactions, providing robust evidence for constructing more comprehensive learner models.

(3)

Personalized and adaptive learning design.

Given students’ variability in spatial ability, digital fluency, and learning styles, future research could leverage artificial intelligence (AI) to generate personalized learning content and scaffolds. Using generative AI and intelligent tutoring systems, task difficulty, sequence, and feedback could be adaptively adjusted in real time based on learners’ performance and interaction data. Such personalized systems would better address diverse learner needs, support differentiated instruction, enhance engagement, and enable analysis of performance differences across abilities, providing valuable insights for system optimization and precision instruction. Individual differences among students with varying digital/spatial abilities could also be used to examine how these individual differences may influence learning outcomes and interaction with the virtual drone system.