Smart Mobility Education and Capacity Building for Sustainable Development: A Review and Case Study

Abstract

Smart mobility has emerged as a transformative enabler for achieving the United Nations Sustainable Development Goals (SDGs), offering technological and systemic solutions to pressing urban challenges such as congestion, environmental degradation, accessibility, and economic inclusion. Realizing this potential, however, depends not only on technological maturity but also on robust education and capacity-building frameworks. This paper addresses two critical gaps: the absence of a systematic review of structured academic curricula, vocational training programs, and professional development pathways dedicated to smart mobility, and the lack of a formal approach to demonstrate how structured, research-oriented education can effectively bridge theory and practice. The review examines a wide spectrum of initiatives, including academic programs, industry training, challenge-based competitions, and community-driven platforms. The analysis shows significant progress in Europe and North America but also reveals important gaps, particularly the limited availability of structured initiatives in the Global South, the underrepresentation of accessibility and inclusivity, and the insufficient integration of governance, ethical AI, policy, and cybersecurity. A case study of the AI for Smart Mobility course, developed using a design science methodology, illustrates how research-oriented education can be operationalized in practice. Since 2020, the course has engaged hundreds of students and professionals, with project dissemination through the AI4SM Medium hub attracting more than 20,000 views and 11,000 reads worldwide. The findings highlight both the progress made and the persistent gaps in smart mobility education, underscoring the need for wider geographic reach, stronger emphasis on inclusivity and governance, and structured approaches that effectively link theory with practice.

1. Introduction

Developing the knowledge, skills, and capacities required to design, implement, and manage next-generation transportation systems is essential for preparing professionals to address the sustainability, accessibility, and efficiency challenges of future mobility. The urgency of such targeted education and capacity-building initiatives is amplified by rapid global urbanization and its associated impacts. According to the United Nations’ World Urbanization Prospects report [1], two-thirds of the global population is projected to reside in urban areas by mid-century, reversing the rural–urban distribution observed in the mid-20th century. This demographic shift, driven by social and economic forces, has intensified mobility challenges and produced negative externalities in contemporary cities, including heightened road safety risks, increased congestion, and elevated greenhouse gas emissions.

Addressing these challenges requires innovative, sustainable, and inclusive mobility solutions. Smart mobility represents a wide umbrella for a range of systems and services designed to meet diverse end-user needs without compromising the collective good of the society and the environment. These systems and services are built on advanced technologies such as softwarization, artificial intelligence, connectivity, and electrification, along with innovative business models such as digital economy, servitization, sharing economy, gig economy, experience economy, and circular economy. Smart mobility systems and services encompass a wide range of technologies and innovations designed to improve the efficiency, sustainability, and inclusiveness of transportation. Examples include Mobility-as-a-Service (MaaS) platforms that integrate public transit, ride-sharing, ride-hailing, bike-sharing, and car rental into a unified digital interface for seamless trip planning and payment. Real-time transit tracking systems enhance user experience by providing live updates on vehicle arrivals and service disruptions. Connected and automated electric vehicle technologies, smart roads, and smart intersections, and smart parking systems contribute to reducing congestion, emissions, and enhancing safety in urban environments. Additionally, demand-responsive transit services use data-driven algorithms to dynamically route shared vehicles based on passenger needs, increasing flexibility and coverage in underserved areas. These systems integrate advanced technologies such as softwarization, artificial intelligence, connectivity, and electrification with innovative business models including the digital economy, servitization, and the sharing economy [2].

The United Nations Sustainable Development Goals (SDGs) offer a comprehensive global framework for addressing critical societal, economic, and environmental challenges. Khamis and Malek discussed in [3] the essential role of smart mobility as an enabler for achieving numerous SDGs by leveraging advanced technologies and innovative business models to enhance accessibility, safety, inclusivity, and sustainability as shwon in Figure 1.

Smart mobility supports society-related goals such as eradicating poverty (SDG 1), ending hunger through resilient food delivery systems (SDG 2), improving health with mobile clinics and reduced emissions (SDG 3), expanding educational access (SDG 4), promoting gender equality through safer and more autonomous travel options for women (SDG 5), contributing to affordable and clean energy through vehicle electrification, smart charging infrastructure, and integration with renewable energy systems (SDG 7), and enabling sustainable, resilient, and inclusive urban transport systems (SDG 11). It also plays a critical role in economy-related goals, including fostering decent jobs via the gig and sharing economies (SDG 8), building sustainable infrastructure through connected and electrified transport (SDG 9), promoting equitable access to mobility (SDG 10), and enabling responsible consumption and production through circular economy practices, efficient resource use, and sustainable procurement (SDG 12). Environment-related goals are addressed through innovations in electrification, circular economy practices, and pollution mitigation technologies that contribute to clean water (SDG 6), climate action (SDG 13), marine conservation (SDG 14), and biodiversity protection (SDG 15).

However, realizing the full potential of smart mobility as an enabler for sustainable development requires more than just technological progress and policy ambition—it demands a skilled and informed workforce. The rapid evolution of smart mobility technologies and the emergence of innovative business models have created a pressing need for targeted education programs and capacity-building initiatives. These programs are essential for equipping current and future professionals with the interdisciplinary knowledge and practical skills required to design, implement, and manage next-generation mobility solutions. Despite the extensive contributions smart mobility can make to the SDGs, as discussed in [3], there remains a notable shortage of structured academic curricula, vocational training, and professional development pathways in this domain. Bridging this gap is critical to unlocking the full societal, economic, and environmental benefits of smart mobility and ensuring that its deployment is both effective and equitable across diverse global contexts.

From a theoretical standpoint, the emerging discourse on smart mobility education intersects with established frameworks in education for sustainable development (ESD), which emphasize equipping learners with the knowledge, skills, and values to address complex societal challenges through systems thinking and participatory approaches [4]. Technological literacy frameworks further highlight the importance of enabling individuals to understand, use, and critically evaluate advanced technologies [5], while transdisciplinary pedagogy advocates integrating knowledge across disciplines and engaging stakeholders beyond academia [6]. In the context of the use case presented in this paper, the design science methodology [7] is adopted as a structured research and teaching approach that emphasizes iterative problem identification, artifact design, and empirical validation.

This paper addresses two critical research gaps in the emerging field of smart mobility education. First, to the best of the author’s knowledge, there has been no systematic review or quantification of the availability of structured academic curricula, vocational training programs, and professional development pathways specifically dedicated to smart mobility. Second, there has been a lack of a formal approach on how structured, research-oriented education can effectively bridge theory and practice in this field. To respond to these gaps, the paper reviews a spectrum of initiatives that contribute to human capital development in smart mobility, including formal academic programs, professional training courses, hands-on competitions, and community-driven knowledge exchange platforms. The review is organized around three core pillars: (1) formal educational and professional training initiatives, (2) competitions and interactive learning activities, and (3) collaborative knowledge exchange and community engagement. In addition, the paper presents a case study on teaching AI for smart mobility, designed using the design science methodology, as a replicable model for bridging theory with practice and for scaling capacity-building efforts. By synthesizing current initiatives and demonstrating a structured pedagogical approach, the paper contributes to a deeper understanding of how education and capacity building can accelerate the adoption of smart mobility as a key enabler for achieving the Sustainable Development Goals and prepare the next generation of mobility innovators, policymakers, and practitioners.

The remainder of this paper is structured as follows. Section 2 reviews formal educational and professional training initiatives related to smart mobility. Section 3 discusses competitions and interactive learning activities that foster experiential and challenge-based learning. Section 4 highlights collaborative knowledge exchange platforms and community engagement efforts. Section 5 presents a detailed case study on teaching AI for smart mobility using a design science approach. Finally, Section 6 summarizes the key findings and future directions.

2. Formal Educational and Professional Training Initiatives

Formal educational programs and professional training initiatives are essential components for preparing a competent workforce in the rapidly evolving field of smart mobility. Given the multidisciplinary nature of smart mobility, which combines technology, urban planning, sustainability, and business innovation, structured educational frameworks and targeted industry training are crucial. These initiatives provide learners with both foundational theoretical knowledge and practical skills, enabling them to effectively engage with and contribute to emerging smart mobility solutions. This section distinguishes between academic programs, which are typically degree- or diploma-oriented and target students seeking foundational knowledge and research capabilities, and industry programs, which are non-degree certifications or short courses designed for professionals seeking applied, practice-oriented skills. The selection criteria for inclusion in this review were based on publicly available and formally established programs identified through targeted keyword searches (such as academic programs, professional development, and smart mobility training) as well as initiatives known to the author through professional engagement.

2.1. Academic Courses and Curriculum Development

Academic programs play a critical role in equipping students with the foundational knowledge and analytical skills needed to address emerging challenges in smart mobility. Their objectives typically include advancing theoretical understanding, building interdisciplinary research capacity, and training graduates who can contribute to innovation in the mobility sector. Pedagogically, effective programs integrate both instructional and constructional activities within active learning environments. The theoretical underpinning for such integration can be found in constructionism [8], an active learning approach where students engage in creating meaningful artifacts and solutions, thereby deepening their understanding through experiential learning [9]. Advanced courses in smart mobility systems should therefore balance theoretical concepts, such as artificial intelligence, electrification, connectivity, and servitization, with hands-on projects that apply these concepts to real-world mobility challenges.

Several academic institutions have recognized this need and established dedicated curricula, as shown in

Table 1. Examples of academic and professional development programs.

Program	Institution / Provider	Program Type
Smart Mobility [10]	Purdue University, USA	Study track
MSc Transport and Mobility Systems [11]	University College London (UCL), UK	MSc
MSc Smart, Connected and Autonomous Vehicles [12]	University of Warwick (WMG), UK	MSc
MA/MPhil/PhD Intelligent Mobility [13]	Royal College of Art (RCA), UK	MA/MPhil/PhD
Smart Mobility Data Science and Analytics (SMDSA) [14]	EIT Urban Mobility Master School	MSc
Sustainable Urban Mobility Transitions (SUMT) [15]	Aalto University, Finland	MSc
MBA Sustainable Mobility Management [16]	TU Berlin, Germany	MBA
MSc Advanced Air Mobility Systems [17]	Cranfield University, UK	MSc
Master’s in Mobility Studies [18]	University of Padua, Italy	Master’s
Smart Mobility and Sustainable Transport Systems [19]	London School of Business and Research, UK	Diploma
Information Technology for Smart and Sustainable Mobility [20]	EUNICE European University	Master’s

. These programs demonstrate best practices in interdisciplinary education by integrating engineering, urban planning, policy, and sustainability perspectives. However, most of the examples are concentrated in Europe and North America, highlighting a geographical imbalance. Institutions such as the EIT Urban Mobility Master School exemplify innovative formats that integrate cross-border collaboration, industry partnerships, and mobility-focused entrepreneurship. Limited information is available on structured curricula in the Global South, which underscores an important gap in global educational equity.

2.2. Industry-Oriented Training Courses and Certifications

In contrast to academic programs, industry-oriented training courses and certifications are shorter, more applied, and aimed at upskilling or reskilling current professionals. Their objectives are less focused on building theoretical foundations and more on providing practitioners with immediately applicable expertise to meet rapidly evolving industry needs. They often rely on partnerships between academia, industry, and public institutions, which facilitates technology transfer and accelerates adoption. Target audiences include engineers, planners, and policymakers who need to adapt to new mobility paradigms.

Table 2. Examples of professional training initiatives.

Initiative	Institution	Description
Foundations of Mobility [21]	University of Michigan	Introduces the systems, policies, and technologies shaping modern mobility.
Public Transit [22]	University of Toronto’s Mobility Network	Explores the design, operations, and policy frameworks of public transportation systems.
Shared Mobility Policy and Planning [23]	University of California, Los Angeles	Examines regulatory, planning, and implementation strategies for shared mobility.
Future Transportation Systems [24]	Massachusetts Institute of Technology	Investigates the role of automation, AI, and user-focused design in shaping future transport.
Self-Driving Car Specialization [25]	Coursera & University of Toronto	Focuses on the software and algorithms behind autonomous vehicle technology.
Batteries and Electric Vehicles [26]	Coursera & Arizona State University	Provides foundational and advanced knowledge of electric vehicle systems.
CIVITAS Sharing Mobility [27]	CIVITAS Initiative	Offers training and certification in shared mobility planning, policy, and implementation.
PAVE [28]	Partners for Automated Vehicle Education	Provides education and outreach to promote understanding of automated vehicles among policymakers and the public.
Urban Mobility Explained (UMX) [29]	EIT Urban Mobility	Delivers short-format online courses and materials to promote best practices and innovations in urban mobility internationally.
SDV 101: Introduction to Software Defined Vehicles [30]	digital.auto	Introduces the architecture, technologies, and development paradigms of SDVs [31].
OVIN Learn [32]	Ontario Vehicle Innovation Network (OVIN)	Offers workforce development programs to build talent in Ontario’s automotive and smart mobility sectors.
Global Challenges in Transport: Smart Mobilities [33]	University of Oxford, Transport Studies Unit	Explores global challenges in transport through the lens of emerging mobility technologies, social equity, and sustainable development.

summarizes representative initiatives, which span formats such as MOOCs, executive education, and professional certifications. Most of these programs are concentrated in Europe and North America, with limited comparable offerings in the Global South. Addressing this imbalance is essential for expanding global capacity in smart mobility education and ensuring more equitable access to training opportunities.

Beyond geographical gaps, several critical thematic areas remain underrepresented in current curricula. These include governance and regulatory issues such as risk, liability, and insurance frameworks for autonomous systems; standardization and interoperability of digital mobility platforms; funding models and public–private partnerships; and legislative frameworks for autonomous driving, urban air mobility, maritime automation, and cross-border mobility. Policy strategies for decarbonization and net-zero mobility, ethical frameworks for automated decision-making, and the social impacts of mobility innovations on employment and labor transitions also require more attention. Equally important are stakeholder and community engagement, accessibility and inclusivity, and legal frameworks for mobility services. Expanding both the regional reach and thematic coverage of smart mobility education therefore remains an urgent priority.

3. Competitions and Interactive Learning Activities

Competitions play a vital role in active and experiential learning, particularly when framed through challenge-based learning (CBL) [34]. CBL is a pedagogical approach in which learners work collaboratively to solve authentic, open-ended problems, often linked to real-world contexts. In the domain of smart mobility, CBL-driven competitions create immersive environments where participants translate theoretical knowledge into practice, thereby fostering technical expertise, creativity, and interdisciplinary teamwork. Beyond technical problem-solving, these competitions also encourage entrepreneurial thinking, communication, and leadership skills, which are essential for preparing a workforce that can address the complex challenges of sustainable and intelligent transportation systems. In this way, CBL connects pedagogy directly to the skills and mindsets required for the mobility sector of the future.

Smart mobility technologies are advancing rapidly, and international competitions have become important venues for innovation in autonomous vehicles, shared mobility, micromobility, and Intelligent Transportation Systems (ITS). These events operate as living laboratories where participants can experiment with new technologies, receive feedback from experts, and benchmark their progress against international peers. The competitions reviewed here were selected based on two criteria: (1) their international recognition and sustained relevance in the field of mobility, and (2) their focus on developing skills aligned with smart mobility education and capacity building. For each competition, contextual information such as the year, target audience, and expected outcomes is highlighted in

Table 3. Examples of competitions in the field of autonomous transportation (selection based on international relevance, audience, and outcomes).

Competition	Organizer(s)	Context, Audience, and Outcomes
Indy Autonomous Challenge (IAC) [35]	Energy Systems Network, CES	Established in 2019; targets university teams. Hosted the world’s first multi-car autonomous race at CES 2025 with speeds up to 163.6 mph. Outcomes include advanced AI algorithms for high-speed navigation, and training of students in safety-critical system design.
SAE AutoDrive Challenge II [36]	SAE International	Running since 2017; designed for multi-year student projects at North American universities. Participants build Level 4 autonomous vehicles for urban scenarios. Outcomes include integration of perception, planning, and control modules, along with exposure to industrial design practices.
Dubai World Challenge for Self-Driving Transport [37]	RTA Dubai, Dubai World Congress	Launched in 2018; targets global consortia of academia, startups, and industry. Offers a USD 3 million prize pool. Outcomes include field-tested prototypes for autonomous transit, logistics, and shared mobility solutions with direct policy relevance.
NXP Cup [38]	NXP Semiconductors	Global student competition with annual editions since 2003. Focused on building low-cost autonomous model vehicles using embedded hardware and sensors. Provides early exposure to embedded AI and mechatronics, supporting skills development at an accessible entry level.
CARLA Autonomous Driving Challenge [39]	CARLA Simulator Community	Ongoing since 2019; open to global academic and research teams. Simulation-first approach using the CARLA platform to test AI agents in realistic traffic scenarios. Outcomes include improved decision-making in ITS and reproducible benchmarks for safe AI in urban environments.
RoboRacer [40]	Roboracer AI	Launched in 2019; engages teams to develop AI control software for electric, driverless race-cars. Audience includes research groups and startups. Outcomes emphasize high-performance computing and public visibility of autonomous racing.

.

The first major milestone in this trajectory was the DARPA Grand Challenge in 2004, which inspired dozens of teams to build autonomous vehicles capable of navigating desert terrain. Although no team succeeded in the initial event, the competition catalyzed the birth of modern autonomous driving research and demonstrated the power of CBL to accelerate technological innovation. Since then, numerous competitions have emerged that advance both technical capabilities and educational outcomes. Table 3 provides examples of key international challenges.

Despite the growth of competitions focused on automated driving, there remains a limited number of challenges addressing other key dimensions of smart mobility, such as accessibility, inclusivity, and user-centered innovation. One example that fills this gap is the IEEE International Conference on Smart Mobility’s Barrier-Free Accessible Mobility Pitch-off Competition [41], which explicitly targets underserved populations including seniors, individuals with disabilities, and people with chronic conditions. The last two edition of this challenge attracted participants from 21 distinct countries across four continents showcasing innovative ideas and prototypes addressing diverse accessibility challenges in smart mobility. Contributions spanned multiple tracks, promoting inclusive transportation solutions for people with disabilities, seniors, and those with chronic conditions. Notable innovations included a smart cane for the visually impaired, accessible scooters, an online platform offering detailed accessibility information on parking, restrooms, and curb cuts, a multi-operator digital pass for streamlined access to MaaS features, and connected vehicle technologies to enhance the driving experience of disabled users, all aligning with the Sustainable Development Goals by fostering equitable, safe, and accessible urban mobility. By encouraging these solutions, this event illustrates how CBL can also be harnessed to advance equity and social sustainability in mobility.

These competitions and interactive learning activities combine technological experimentation with authentic, real-world challenges. Their CBL foundation ensures that learners not only gain technical proficiency but also develop problem-solving, collaboration, and critical thinking skills. These experiences strengthen the human capital pipeline for smart mobility and contribute to building more sustainable, inclusive, and intelligent transportation ecosystems.

4. Collaborative Knowledge Exchange and Community Engagement

Collaborative knowledge exchange and community engagement are foundational to advancing sustainable development through smart mobility technologies. As cities worldwide face mounting challenges related to congestion, emissions, accessibility, and infrastructure resilience, the smart mobility community has responded by cultivating vibrant ecosystems where researchers, practitioners, policymakers, and industry leaders converge to share ideas, align strategies, and co-develop innovative solutions.

Several major international conferences and exhibitions provide platforms for advancing knowledge, showcasing innovation, and fostering collaboration in smart mobility, as summarized in

Table 4. Examples of major international conferences and exhibitions on smart mobility.

Conference/Exhibition	Organizer(s)	Focus and Contributions
IEEE International Conference on Intelligent Transportation Systems (ITSC) [42]	IEEE	Interdisciplinary collaboration through workshops, technical papers, and policy discussions on intelligent vehicles, vehicle–infrastructure integration, and AI-driven traffic management.
Tomorrow.Mobility World Congress (TMWC) [43]	Fira Barcelona, EIT Urban Mobility	Brings together stakeholders from public and private sectors to explore urban mobility futures, including micromobility, shared transport, regulatory innovation, and urban design.
ITS World Congress [44]	ERTICO-ITS Europe	Leading annual venue showcasing connected and autonomous vehicle technologies, cooperative ITS systems, and next-generation infrastructure solutions across hundreds of sessions and demonstrations.
IEEE International Conference on Smart Mobility (IEEE SM) [41]	IEEE	Organized into three tracks, technology, city planning, and governance, reflecting the importance of governance frameworks and urban planning alongside technological maturity.
Consumer Electronics Show (CES) [45]	Consumer Technology Association (CTA)	A prominent global exhibition for unveiling smart mobility innovations, including electric vehicles, connectivity, autonomous driving, software-defined vehicles, and urban mobility solutions.

.

These events are not only platforms for presenting cutting-edge research and products, but also for building communities of practice, aligning international standards, and accelerating real-world deployment. By encouraging open dialogue between academia, industry, and government, they help break down silos and cultivate shared understanding of region-specific challenges and scalable mobility innovations.

Beyond technical advances, these forums increasingly emphasize inclusion, accessibility, and diversity—essential dimensions of sustainable development. Initiatives such as the Women in ITS & Mobility program by ERTICO–ITS Europe and Women in Mobility Symposium in IEEE SM highlight the importance of equitable representation in shaping future mobility ecosystems. Meanwhile, new formats such as innovation challenges and startup showcases bring in fresh perspectives from entrepreneurs and early-career professionals, expanding participation and enriching the collaborative landscape.

Complementing these conferences are digital platforms like the AI4SM Medium publication hub [46], which democratize access to knowledge and tools by sharing open-source code, datasets, and research insights related to smart logistics, mobility optimization, and AI-driven planning. These community-driven efforts empower a wider audience, from university students to city planners, to participate in the innovation cycle and apply emerging technologies to their local contexts.

These collaborative and community-centered initiatives reinforce a shared global commitment to sustainable mobility. They accelerate the development and responsible deployment of smart mobility technologies while nurturing inclusive networks that are essential for long-term societal impact.

The previous Section 2, Section 3 and Section 4 reviewed formal educational and professional training initiatives, challenge-based competitions, and collaborative platforms that collectively shape capacity building in smart mobility. Building on this broader landscape, the next section turns to a detailed case study of the AI for Smart Mobility course, which demonstrates how research-oriented education, grounded in a design science methodology, can bridge theory and practice and provide a replicable model for structured capacity building in this emerging field.

5. Case Study: Teaching AI for Smart Mobility Through Design Science

Artificial intelligence (AI) is one of the foundational technologies of smart mobility. Despite the critical role AI has played, plays, and will play in transforming mobility systems and services and enabling new business models, there remains a notably limited number of academic courses, professional development programs, vocational training, online certification programs, or corporate training explicitly focused on the intersection of AI with emerging smart mobility systems, services and business models. This gap leaves many students and professionals underprepared to address the challenges and opportunities arising from advancements in AI-driven mobility solutions. This section introduces a course designed to bridges this knowledge gap by integrating AI’s technical foundations with practical applications in smart mobility. This course was offered as a special topic in software engineering at the University of Toronto, as a special topic in operations research at KFUPM, and as an industry training course at the Canadian Technical Center of General Motors from Winter 2020 to Winter 2025, reaching a total of 300 students and 27 employees. The course was designed to explore the dynamic intersection between AI and mobility, examining how AI technologies are transforming systems, services, and business models in the mobility domain. This course was structured around three common patterns of AI application in smart mobility as illustrated in Figure 2.

• Situational awareness: AI is used to enhance situational awareness at various levels or perspectives, enabling better evidence-based decision-making. Valuable information can be derived from data to enhance situational awareness. While data is often referred to as “the new oil,” its true value lies in the information it generates—analogous to oil being refined into useful products. For instance, raw traffic accident data has limited value unless analyzed to extract actionable insights. By applying analytics at various levels, we can transform this data into meaningful information. Descriptive analytics can help identify patterns such as the most common times for accidents, the types of vehicles involved, or locations and hotspots with the highest accident rates. Diagnostic analytics delve deeper to uncover the root causes of high accident rates at specific intersections, considering factors such as lighting conditions, weather, road quality, traffic signal timing, or driver behavior. Finally, predictive analytics enable forecasting the likelihood of accidents in different areas under varying conditions, such as specific times of day or adverse weather scenarios. For example, Xu et al. propose in [47] a digital twin to enable real-time situational awareness and urban mobility management. Similarly, Zhang et al. [48] examine the mobility constraints of residents in marginal rural areas of megacities, providing spatial insights that can inform equitable transport planning. Yao et al. [49] review radar data representations in autonomous driving, offering comprehensive guidance for perception systems that enhance situational awareness in complex environments.
• Automation: AI acts as a key enabler for automation, encompassing both physical and digital processes. Examples of physical processes include mobility platform manufacturing tasks such as robotic welding, assembly, and painting; automated material handling; automated quality inspection; as well as applications like automated public transport, automated people movers (APM), smart parking systems, and automated driving. For digital processes, AI powers functionalities such as in-vehicle digital assistants, robotic process automation (RPA), and streamlined content creation. It also supports intelligent agent assistants and conversational agents, such as retail support chatbots, question-and-answer (Q&A) engines, and recommendation systems that generate personalized suggestions. Re-visiting the traffic accident example, prescriptive analytics can generate actionable recommendations for interventions, such as adjusting traffic light patterns, increasing police patrols, or improving road infrastructure in high-risk areas. Additionally, AI drives digital go-to-market (GTM) features, including online booking and shopping platforms, enhancing user experience and operational efficiency. For example, Donthireddy shows in [50] how AI techniques can enhance GTM strategies, resulting in significant improvements in sales growth, customer acquisition and retention, and return on investment (ROI). In the domain of connected and automated driving, Macioszek and Tumminello show in [51] how connectivity and automation can be integrated into smart intersections to improve traffic operations and enhance safety in complex urban environments. Moreover, Song et al. [52] model lane-changing spatiotemporal features based on human driver behavior generation mechanisms, enabling safer and more human-like automation in traffic flow.
• Optimization: AI also tackles complex optimization problems in diverse people mobility, logistics, and infrastructure contexts, including design, planning, and control. These problems are classified based on the expected quality of the solutions and the time available to find them [53]. Real-world design problems, often associated with strategic functions, prioritize solution quality over time. In such cases, users may be willing to wait, sometimes even several days, for optimal solutions. Examples include, but are not limited to, the optimal placement of EV charging stations [54] or micromobility stations [55]. Planning problems, or tactical functions, require solutions within a moderate timeframe, typically ranging from a few seconds to a few minutes. Proactive dispatching of emergency vehicles [56] and dynamic pricing of parking rates [57] are examples of planning problems. Control problems, representing operational functions, must be solved repetitively and within extremely short timeframes, often between a few milliseconds and a few seconds. In these cases, optimality is typically sacrificed to achieve the necessary speed for real-time or near-real-time decision-making. Examples include traffic light coordination at smart intersections [58] and lane-keeping assistance [59]. Recent advancements include the work by Luo et al. [60], which applies machine learning to optimize transport infrastructure connectivity and resolve conflicts.

This course was designed to explore these three patterns comprehensively. AI is presented as a general-purpose technology, empowering students to select problems within the domains of people mobility, logistics, or transportation infrastructure. Since its introduction in 2020, the course has maintained these themes while evolving the approaches and tools adopted by students in response to advances in AI and emerging mobility trends. Early cohorts primarily relied on classical machine learning and optimization techniques, whereas more recent projects have incorporated deep learning, reinforcement learning, and large language models. This progression fosters both technical proficiency and creative problem-solving, allowing students to tackle real-world challenges effectively while illustrating the course’s ability to preserve conceptual continuity and adapt to the rapidly changing technological landscape of AI for smart mobility.

5.1. Design Science Approach

Several pedagogical and instructional frameworks can guide the design of educational programs in emerging domains such as smart mobility. Transformative learning theory [61] focuses on fostering critical reflection that leads to shifts in learners’ frames of reference, thereby promoting deep, paradigm-changing learning experiences. Technology-enhanced learning [62] centers on using digital technologies to create interactive, adaptive, and student-centered learning environments. Other relevant frameworks include Experiential Learning Theory [63], which emphasizes learning through experience and reflection; practice-based learning [64] and challenge-based learning [65], which both promote collaborative problem-solving in authentic, real-world contexts; project-based learning [66], which engages learners in extended investigations and solutions; competency-based education [67], which focuses on mastery of specific skills; and Connectivism [68], which views learning as the ability to build and navigate networks of information and expertise. The Design Science approach [7] emphasizes the iterative creation and empirical evaluation of artifacts that address real-world problems while contributing to theoretical knowledge through rigorous and iterative processes. This approach is a well-established paradigm across multiple disciplines, including information systems, computer science and engineering, and various engineering fields such as mechanical, civil, architectural, and manufacturing engineering [69]. It offers a methodological foundation for systematically linking problem identification, solution design, and empirical validation. For this reason, it was adopted in this paper to structure the AI for Smart Mobility course as a replicable model for capacity building in smart mobility education.

Adopting the design science methodology in the AI for Smart Mobility course enables students to develop foundational and applied knowledge in smart mobility, equipping them with the skills to design and implement AI-driven solutions for real-world problems. This approach not only fosters a deep understanding of state-of-the-art AI methods but also bridges the gap between theoretical insights and practical applications in smart mobility systems, services, and business models. The methodology guides students through a structured process of designing innovative artifacts (e.g., algorithms, frameworks, or models) that address specific, context-driven problems while contributing to a broader understanding of smart mobility as a discipline.

Figure 3 illustrates the relationship between the problem domain and the solution domain, as well as the connection between theory and practice. Through this structured approach, students systematically analyze complex mobility problems, develop AI-based solutions, and evaluate their effectiveness in practical contexts. The following subsections describe the different components included in Figure 3.

5.1.1. Problem Domain

• Problem Instance(s): Students are presented with a curated set of design, planning, and control problems spanning transportation infrastructure, people mobility, and logistics. Additionally, they are granted the flexibility to propose their own problem instances, provided they align with smart mobility themes and demonstrate relevance to real-world problems.
• Problem Characterization: Students are tasked with defining the scope of their selected problem, articulating its significance, and identifying its challenging aspects. This step requires them to justify the problem’s relevance to smart mobility and highlight the complexities that necessitate innovative solutions.
• Problem Construct(s): Students formulate the problem using a structured modeling approach, such as mathematical representations, simulations, or conceptual frameworks, to precisely capture its key elements and constraints. This formalization serves as the foundation for subsequent solution design.

5.1.2. Solution Domain

• Design Construct(s): Students conduct a comprehensive literature review, evaluating the strengths and limitations of recent work in smart mobility and AI applications. This analysis enables them to identify gaps in the existing research.
• Instantiation: Students perform exploratory data analysis within the scope of their chosen problem, leveraging datasets relevant to smart mobility (e.g., traffic flows, electric micromobility, logistics networks) to test hypotheses, refine models, or uncover patterns that inform their solution development.
• Abstraction: This component focuses on distilling the solution’s novelty and potential for broader applicability. Students generalize their findings into design principles or frameworks that extend beyond the specific problem instance, contributing to the theoretical advancement of smart mobility.

5.1.3. Intersection (Bridging Problem and Solution Domains)

• Technological Rule(s): Technological rules encapsulate generalizable knowledge about the relationships between identified problems and their proposed solutions. Derived through abstraction and validated empirically, these rules articulate design principles or heuristics that practitioners can apply across similar smart mobility contexts.
• Solution Design: Positioned at the intersection, solution design involves students synthesizing insights from the problem domain (e.g., problem constructs) and the solution domain (e.g., design constructs) to create innovative AI-based artifacts. This process integrates theoretical foundations with practical applicability, guided by a literature-informed understanding of existing approaches.
• Empirical Validation: Serving as a critical link, empirical validation entails rigorous testing of the proposed solutions against real-world data or simulated scenarios. This step ensures that the artifacts are both effective in addressing the problem instances and robust in their theoretical contributions, closing the loop between problem and solution domains.

Through this iterative process, students not only develop actionable solutions for practitioners but also contribute to the theoretical foundation of smart mobility, fostering advances that are both practically relevant and academically rigorous.

The visual abstract template [70] serves as a tool for effectively communicating design science contributions. It helps highlight the key problem and solution constructs while also presenting the validity aspects of design knowledge. Figure 4 presents a visual abstract template proposed in [69]. This template captures three key aspects of design science contributions. First, it represents the proposed or refined theory in the form of a technological rule. Second, it illustrates the empirical contribution by depicting problem-solution instances along with the corresponding design and validation cycles. Finally, it provides a framework for assessing the value of the generated knowledge based on relevance, rigor, and novelty.

For instance, traffic congestion at urban intersections poses a significant challenge, particularly in rapidly growing cities like Hangzhou, China, where high population density and complex traffic patterns strain transportation infrastructure. This example explores the application of a design science approach to address this issue through the development of a deep reinforcement learning (DRL)-based traffic signal control (TSC) system. By integrating state-of-the-art AI techniques with real-world traffic data, the course guides students in designing, implementing, and validating a solution tailored to Hangzhou’s busy intersections [71]. The following visual abstract components illustrate how this process unfolds, from problem identification to solution validation, culminating in a technological rule with both practical and theoretical contributions to smart mobility.

• Technological Rule(s): a generalizable principle stating how a specific intervention achieves a desired effect in a given context. Formulated as: “To achieve $<Effect/Change>$ in $<Situation/Context>$, apply $<Solution/Intervention>$.” Example: to mitigate congestion at busy intersections in urban environments, apply deep reinforcement learning (DRL)-based traffic signal control. This rule encapsulates a design principle linking DRL to reduced traffic delays.
• Problem Instance: a concrete, real-world manifestation of the problem outlined in the Technological Rule, defined by specific data or stakeholder needs. Example: how can traffic signals at major intersections in Hangzhou, China, be controlled to reduce congestion, based on real traffic flow data and the needs of city planners to improve urban mobility?
• Problem Understanding: analytical studies or methods that deepen the conceptual grasp of the problem instance. Example: problem characterization identifies congestion as a critical issue at Hangzhou intersections, supported by exploratory data analysis of traffic patterns (e.g., peak-hour volumes, vehicle delays) to reveal underlying causes and complexities.
• Solution: the practical implementation of the intervention proposed in the Technological Rule. Example: a DRL-based traffic signal control (TSC) system, deployed to dynamically adjust signal timings at Hangzhou intersections, optimizing traffic flow in real time.
• Solution Design Approach: the theoretical foundation or investigative process guiding the core design choices of the solution. Example: the solution emerged from a literature review analyzing the strengths and limitations of prior adaptive signal control methods (e.g., fixed-time schedules, rule-based systems). By evaluating the applicability of DRL algorithms studied in the course, the design leverages DRL’s ability to adapt to fluctuating traffic conditions, addressing gaps in traditional approaches.
• Validation Approach: empirical studies or tests to assess the solution’s effectiveness in addressing the problem instance. Example: validation involves testing the DRL-based TSC system using real traffic data collected from Hangzhou intersections across diverse scenarios (e.g., rush hour, off-peak, incidents), measuring reductions in delay and queue length against baseline methods.
• Relevance: contextual factors influencing the solution’s applicability and value, identifying the stakeholders for whom the Technological Rule is pertinent. Example: this rule is relevant to city planners and traffic engineers in urban settings like Hangzhou, where dense populations and busy intersections create persistent congestion challenges, making adaptive, scalable solutions highly valuable.
• Rigor: attributes of the three knowledge-creating activities: problem understanding, solution design, and in-context evaluation. These attributes bolster the empirical credibility of the Technological Rule and indicate its maturity. Example: rigor stems from (1) a detailed analysis of congestion’s importance and challenges at intersections, (2) a comprehensive review of recent adaptive signal control approaches and DRL’s merits, and (3) robust validation using real-world data from four-way intersections in varied traffic scenarios, along with a plan to enhance maturity by extending the scope to diverse intersection types, such as five-way layouts, thereby refining the rule’s generalizability.
• Novelty: the Technological Rule’s contribution relative to existing knowledge, highlighting distinctions from comparable rules or solutions. Example: this DRL-based approach advances adaptive traffic signal control by outperforming traditional fixed schedules, offering greater adaptivity to fluctuating traffic scenarios, reduced congestion, and smoother state transitions. Unlike prior methods, it leverages DRL’s learning capacity to optimize signal timings dynamically, extending the precision and flexibility of existing frameworks.

This example demonstrates the use of a design science approach to tackle traffic congestion at major intersections in Hangzhou, China, through a deep reinforcement learning (DRL)-based traffic signal control (TSC) system. Students in the course identify the congestion problem, analyze real traffic data, and design a DRL solution that adapts signal timings to dynamic conditions. The process involves characterizing the problem, reviewing prior approaches, implementing the solution, and validating its effectiveness with Hangzhou-specific scenarios. The outcome is a technological rule: “To mitigate congestion at busy urban intersections, apply DRL-based TSC,” which offers reduced delays, enhanced adaptivity, and a novel contribution to smart mobility. This rule is relevant to city planners and can be extended to diverse intersection types.

5.2. Course Design and Development

The course is designed to address the transformative role of AI in smart mobility. It explores various AI techniques and their applications across different areas of smart mobility.

5.2.1. Course Objectives and Learning Outcomes

This course aims to provide students with a comprehensive understanding of smart mobility systems and services, focusing on their underlying technologies, key enablers, and potential disruptors. Through a combination of theoretical foundations and hands-on applications, students will explore advanced AI techniques and their role in solving real-world problems in transportation infrastructure, people mobility, and logistics following design science approach. The objective mirrors design science’s focus on creating practical, innovative artifacts (e.g., AI-driven solutions) while advancing theoretical knowledge. By adopting a design science methodology, the course structures learning around problem-solving and artifact development, emphasizing iterative design, implementation, and evaluation. Upon completion of the course, students will be able to do the following:

• Demonstrate knowledge of existing and emerging smart mobility systems, identifying their foundational technologies, enablers, and disruptive influences.
• Apply principles of geospatial data science to analyze and model mobility-related data.
• Implement graph search methods and evaluate their practical utility in smart mobility applications.
• Master search and optimization techniques, including trajectory-based algorithms, evolutionary computing, swarm intelligence, and machine learning, and adapt them to mobility contexts.
• Design, implement, and validate AI-driven solutions to real-world design, planning, and control problems, spanning transportation infrastructure, people mobility, and logistics.

The learning outcomes follow design science approach: understanding smart mobility systems (problem conceptualization), mastering AI techniques (solution design), applying geospatial data analysis, graph search, optimization methods and machine learning methods (instantiation), and solving real-world problems in transportation infrastructure, people mobility, and logistics (validation). This process fosters the generation of technological rules, generalizable design knowledge, bridging theory and practice, a core tenet of design science.

5.2.2. Lectures

Lectures cover topics such as ill-structured problems in smart mobility systems and services, geospatial data science, graph search methods, stochastic search, evolutionary computing algorithms, swarm intelligence algorithms, and machine learning methods, including supervised, unsupervised, and reinforcement learning. To understand students’ learning preferences and the classroom environment, the instructor can begin by asking them a few questions and share the results in the second lecture. For example, my students’ responses (

Table 5. Preferred teaching methods.

I Learn Best in Classes Where the Instructor	Percentage (%)
Provides clear and structured lectures.	36.6
Uses interactive methods such as discussions or group work.	9.8
Incorporates real-world examples and applications.	36.6
Offers hands-on learning opportunities like projects.	9.8
Encourages questions and open dialogue.	7.3

) indicate a strong preference for clear, structured lectures and real-world applications, while interactive and hands-on methods are less favored.

Table 6. Student perceptions of peer contributions to learning.

Other Students in the Course Help Me Learn When They	Percentage (%)
Actively participate in discussions and group work.	30.0
Share diverse perspectives and experiences.	20.0
Provide constructive feedback during peer reviews.	5.0
Collaborate on projects and assignments.	30.0
Respect different learning styles and contribute to a positive environment.	15.0

shows that students benefit most from active participation and collaboration, while peer feedback plays a less significant role in their learning experience.

Table 7. Factors that motivate student participation.

I Am Most Likely to Participate in Classes When	Percentage (%)
The class environment feels welcoming and inclusive.	32.5
The instructor encourages active participation.	15.0
The course material is engaging and relevant to my interests.	35.0
There are opportunities for collaborative learning.	15.0
Participation is linked to grades.	2.5

highlights that students are most likely to participate when the class environment is welcoming and the material is engaging, while participation tied to grades has minimal influence.

Table 8. Challenges that hinder student learning.

Something That Makes It Hard to Learn in a Course Is	Percentage (%)
Lack of clear instructions or guidance from the instructor.	30.0
A fast-paced curriculum that is difficult to keep up with.	37.5
Limited interaction or feedback from the instructor.	10.0
Distractions in the learning environment.	2.5
Poor organization or structure in course materials.	20.0

suggests that a fast-paced curriculum and unclear instructions are the biggest obstacles to learning, while distractions in the learning environment have minimal impact.

Table 9. Factors that enhance student learning.

Something That Makes It Easy to Learn in a Course Is	Percentage (%)
Clear and well-organized course materials.	37.5
Engaging and interactive teaching methods.	17.5
Frequent feedback and support from the instructor.	17.5
A positive and collaborative class environment.	10.0
Real-world applications and examples relevant to the subject.	17.5

indicates that clear and well-organized course materials are the most significant factor in facilitating learning, while a positive class environment plays a smaller role.

This learning preferences survey helps the instructor tailor teaching methods to students’ learning preferences, foster engagement, and create a positive, inclusive classroom environment.

Asking questions during the lecture also promotes active blended learning [72] by encouraging students to engage critically with the material, reflect on real-world applications, and articulate their thoughts. This approach fosters active class participation, critical thinking, and deeper understanding, making learning more interactive and meaningful.

Table 10. Sample of in-class participation questions.

In Your Opinion, What Are the Three Most Challenging Aspects of the Widespread Deployment and Societal Acceptance of Smart Mobility Technologies?	Percentage (%)
Maturity of the technology (e.g., safety, reliability, scalability and seamless integration between different transportation modes)	24.2
Proper city planning and infrastructure (e.g., road networks, charging stations, or communication infrastructure to support smart mobility)	21.2
Availability of a well-developed governance framework (e.g., regulations, policies, and legal standards that support smart mobility technologies and handle data security and privacy concerns)	28.8
Public awareness and trust (e.g., public concerns over safety, privacy, or ethical implications)	18.2
Economic feasibility (e.g., cost of deployment, maintenance, and overall affordability)	6.1
Environmental impact (e.g., the sustainability and ecological footprint of smart mobility technologies)	1.5

shows a sample of these questions. Students responses indicate that the maturity of the technology and the availability of a governance framework are considered the most significant challenges for the widespread deployment and societal acceptance of smart mobility technologies. Public trust and proper city planning also play key roles, while economic feasibility and environmental impact are seen as lesser concerns.

5.2.3. Course Projects

Projects are an essential component of the course and can be completed individually or in groups. These projects provide hands-on experience in applying AI techniques learned in the course to solve real-world problems. The course instructor provides guidance on conducting literature searches, analyzing the challenges of the selected problem, and utilizing relevant research. Sample codes and data are available through the Medium publication hub of the course [46]. The course standardizes on Python 3.8+ due to its accessibility, versatility, and extensive ecosystem of AI and data science libraries. Students are given the freedom to select any library, machine learning model, or architecture that best suited their project needs. The project requirements are as follows:

• Students explore the applicability of the algorithms studied in the course to solve the selected problem.
• Students select an AI approach, implement it in Python, and justify their choice.
• Students define a set of evaluation metrics and conduct experimental analyses to assess the performance of the implemented algorithm under different parameter settings, identifying the strengths and limitations of their solution.
• Students perform a comparative study between their selected approach and alternative methods, evaluating both quantitatively and qualitatively using well-defined metrics and a benchmark dataset.

Each group must select a single problem for the course project. The chosen problem can focus on design, planning, or control within one of the three subdomains of smart mobility: transportation infrastructure, people mobility, or logistics. A list of suggested project ideas for each category is provided below:

• Design Problems—Infrastructure: Optimal placement of physical assets (e.g., EV charging stations, city bike terminals, walking routes, cycling lanes, air taxi takeoff and landing sites, traffic sensors); districting problems (e.g., school bus routing, postal or package delivery, waste collection, sales territory design, public transport service areas, emergency service districting, maintenance districting).
• Design Problems—People Mobility: MaaS bundling, Manufacturer’s Suggested Retail Price (MSRP) optimization, EV incentive optimization, incentive allocation/optimization.
• Design Problems—Logistics: Optimal allocation of distribution points.
• Planning Problems—Infrastructure: Optimal deployment of physical assets (ambulances or fire trucks, mobile healthcare units, mobile speed cameras/radar, patrolling units, mobile sensors, machinery for projects, etc.), dynamic pricing of public transit (e.g., roadway congestion pricing, public transit peak pricing, parking rates, railway and bus pricing, electric vehicle charging stations, freight transportation, shipping fees, etc.).
• Planning Problems—People Mobility: Multicriteria routing, optimal dispatching for ride-hailing or ridesourcing services, deadheading problem, fitness planning, multimodal seamless trip planning, dynamic pricing (e.g., ride-sharing platforms, airline tickets, ferry services).
• Planning Problems—Logistics: Eco-routing, last-mile delivery dispatching, deadheading problem, dynamic pricing (e.g., freight transportation, shipping fees).
• Control Problems—Infrastructure: High Occupancy Vehicle (HOV) and High Occupancy Toll (HOT) management, traffic light control, variable speed limit control, smart intersection control (e.g., emergency modes).
• Control Problems—People Mobility: Longitudinal and lateral control in assisted/ automated vehicles, lane-keep assist, automated valet parking, autopilots, elevator dispatching, Automated People Mover (APM) dispatching.
• Control Problems—Logistics: Communication relaying, instant order assignment and tracking, truck platooning.

This list serves as a set of suggested project ideas only. Students may identify another problem of academic, industrial, or commercial significance related to smart mobility for which no viable solution with reasonable capabilities currently exists and attempt to address it using an AI algorithm covered in the course.

The course project follows the design science methodology, adhering to its iterative, problem-solving framework. Students begin by identifying real-world smart mobility problems (e.g., traffic congestion in Hangzhou), conceptualizing them through problem characterization and data analysis, which aligns with design science’s emphasis on understanding the problem domain. They then design AI-driven artifacts, such as a deep reinforcement learning-based traffic signal control system, reflecting the methodology’s focus on creating innovative solutions. Through implementation and empirical validation using real data, students test these artifacts’ effectiveness, mirroring design science’s evaluation phase. Finally, by abstracting generalizable technological rules (e.g., “To mitigate urban congestion, apply DRL-based control”), they contribute to theoretical knowledge, fulfilling design science’s dual goals of practical utility and scientific rigor.

5.2.4. Assignments

Two assignments are designed following the design science methodology, addressing key phases of artifact development. The first assignment, problem characterization, formulation, and modeling, corresponds to design science’s problem conceptualization phase, where students define and structure a real-world smart mobility issue (e.g., traffic congestion) and lay the groundwork for solution design through mathematical models. The second assignment, a critical survey and taxonomic classification of state-of-the-art approaches, supports the solution design phase by enabling students to analyze existing methods, identify gaps, and establish a theoretical foundation for their innovative AI-based artifact. Together, these tasks initiate the iterative design science process of understanding the problem domain and crafting a rigorously informed solution.

5.2.5. Midterm Article

The midterm report is a short article to be published in the course’s Medium publication, AI for Smart Mobility—AI4SM [46]. Students receive guidance on writing for AI4SM. The article, designed as a brief Medium post (5–10 min reading time), should outline the problem and detail the exploratory spatial data analysis (ESDA) conducted on the problem dataset. It must describe the ESDA process, highlight key observations, and explain how these findings deepen the understanding of the problem. The dataset and source code should be shared via Google Colab. The article should include the following components:

• Problem Dataset: students compile a dataset tailored to their selected problem. Depending on the problem, this dataset may include map location data (e.g., vehicles, passengers, eBike/eScooter stations, depots, parking spots, bus stops), traffic data, time windows, road data (e.g., lane boundaries, intersections, crosswalks, stop signs, traffic lights), delivery schedules, base and surge service pricing, etc. Publicly available open datasets may be used if they align with the problem’s requirements. A concise description of the dataset’s components should be included in the article.
• Exploratory Spatial Data Analysis (ESDA): students visualize and analyze the problem data to uncover patterns and relationships, thereby enhancing their understanding of the smart mobility challenge addressed in the course project.

The midterm article aligns with the design science methodology by contributing to the problem conceptualization and early solution design phases. Through compiling and describing the problem dataset, students define a real-world smart mobility problem (e.g., traffic congestion), grounding it in concrete data. The ESDA process enhances problem understanding by uncovering patterns and relationships, fulfilling design science’s emphasis on rigorous problem characterization. By sharing their findings and code via the AI4SM Medium publication and Google Colab, students initiate the artifact development cycle, setting the stage for designing and evaluating AI-driven solutions while contributing to generalizable knowledge, a core design science objective.

5.2.6. Final Project

The final project report is prepared as a scientific paper that aligns with the design science methodology by encompassing its core phases of artifact creation, evaluation, and knowledge contribution. The Introduction and Problem Formulation sections of the paper define and model a real-world smart mobility problem, fulfilling the problem conceptualization phase. The Literature Review situates the work within existing knowledge, guiding the solution design phase, where the Proposed Solution details an AI-driven artifact (e.g., a traffic control algorithm). Performance Evaluation tests the artifact’s effectiveness through experiments and metrics, reflecting design science’s empirical validation requirement. Conclusions abstract generalizable insights or technological rules, contributing to theoretical knowledge, while the Python code shared via GitHub or Colab ensures practical utility—together embodying design science’s dual focus on rigor and relevance.

The course has produced impressive student projects, published with open-source code and data, covering a wide range of topics related to people mobility, logistics, and infrastructure. These topics include: leveraging large language models (LLMs) for reinforcement learning agents in intelligent driving scenarios, traffic flow prediction, optimal placement of EV charging stations, food delivery optimization, estimated time of arrival (ETA) prediction, optimizing traffic sensor placement, personalized cycling path routing, revamping Toronto’s fire districts, ride-sharing pricing, enhancing school bus routing, predicting bike availability, districting and routing for waste collection, bus stop optimization, dynamic pricing of public transit, student boarding and routing optimization, hotel recommendations and route optimization, optimal placement of public parcel lockers, and predictive response and optimal allocation of healthcare facilities with an accessibility score. Some projects have been published as Medium articles, while others were presented as scientific papers at international conferences, such as the IEEE International Conference on Smart Mobility.

The selection of projects for dissemination is guided by a set of quality-oriented criteria. Specifically, projects are evaluated based on the general quality of the article, including clarity of presentation and rigor of analysis; the originality and level of innovativeness; the significance of the contribution for advancing theory and/or practice in smart mobility; the robustness of experimental results, including both quantitative and qualitative analyses; and the quality and reproducibility of the implemented code along with the availability of the datasets used. These criteria ensured that the published projects met high academic and professional standards while maximizing their impact and relevance.

5.3. Lessons Learned

Student feedback on the AI for Smart Mobility course, as reflected in the overall mean course evaluations from Winter 2020 to Winter 2025 (

Table 11. Overall mean course evaluation out of 5 (Winter 2020, Fall 2020, Fall 2021, Fall 2022, Fall 2023, Fall 2024 and Winter 2025).

Question	Course Mean	Division Mean
I found the course intellectually stimulating.	4.4	4.3
The course provided me with a deeper understanding of the subject matter.	4.5	4.3
The instructor created an atmosphere that was conducive to my learning.	4.6	4.3
Course projects, assignments, tests, and/or exams improved my understanding of the course material.	4.5	4.3
Course projects, assignments, tests and/or exams provided opportunity for me to demonstrate an understanding of the course material.	4.3	4.3
Overall, the quality of my learning experience in this course was:	4.5	4.2
Overall Mean	4.5	4.3

), indicates a consistently positive learning experience. Students rated the course highly for being intellectually stimulating, suggesting that it effectively engages their curiosity and critical thinking. The course also excels in deepening subject matter understanding, aligning with its objective to explore smart mobility systems and AI applications. The instructor’s ability to foster a conducive learning atmosphere received the highest rating, reflecting strong facilitation of student engagement. Course projects and assignments were rated solidly for enhancing understanding and providing opportunities to demonstrate mastery, indicating their relevance and effectiveness within the design science framework. These ratings, consistent across course, department, and division levels, affirm the course’s success in delivering a rigorous, practical, and well-supported educational experience, though minor adjustments could elevate the overall perception further.

The data on views and reads for the AI for Smart Mobility (AI4SM) Medium publication hub (Figure 5) reveals distinct engagement rhythms over the 22-month period from August 2023 to June 2025 and links directly to the course learning outcomes (CLOs) mentioned in Section 5.2.1. Peaks consistently coincide with assessment milestones when students publicly post project artifacts. The November 2023 surge (2400 views, 1400 reads; read-to-view ratio 58%) aligns with midterm publications in which learners demonstrated CLO 1 by surveying existing and emerging smart mobility systems, CLO 2 by presenting geospatial analyses and models, and CLO 3 by explaining graph search implementations with performance evidence. Subsequent steady engagement from January to April 2024 (about 1300 views and 700–750 reads per month) reflects continued discovery and sustained reading of optimization and AI posts that operationalize CLO 4 and CLO 5 in design, planning, and control contexts. The November 2024 spike (1930 views, 943 reads) coincides with final project dissemination, where teams reported end-to-end AI-driven solutions and validation results, thereby evidencing attainment of CLO 5.

In the five months following the course introduction at KFUPM (January–May 2025), the hub stabilized at roughly 600 views and 250 reads per month with an average conversion of about 45%. Although absolute traffic is lower than the November 2024 peak, the low variance and a modest rise in May, driven by new project posts, indicate an emerging core readership beyond a single institution. Educationally, these patterns suggest that public dissemination supports knowledge consolidation and reflective articulation of methods mapped to CLOs 1–5, while the sustained read share is a proxy for depth of engagement with technical content.

The medium’s native metrics, including views, reads, and read-to-view ratios, were also used as a feedback mechanism in subsequent course offerings. Articles with higher engagement were presented to new students as examples of effective project communication and dissemination, helping them refine how they framed and shared their own work.

From a community impact perspective, recurring baselines outside peak academic periods point to continued access by external audiences such as students, educators, and practitioners who consult the posted articles, code, and datasets. Overall, the evidence links publication analytics to concrete educational outcomes, namely demonstrated mastery of course competencies through public scholarship, and to broader community benefit through ongoing open access to reproducible artifacts.

Looking ahead, future iterations of the course may also explore opportunities for international collaboration, for example through joint projects with students from other universities, in order to broaden perspectives, enhance global engagement, and further strengthen the interdisciplinary learning experience.

6. Conclusions

The success of smart mobility as a transformative force depends not only on technological maturity but also on education and capacity building that prepare the next generation of innovators, policymakers, and practitioners. This paper examined the critical role of education in advancing smart mobility as a key enabler of the Sustainable Development Goals and addressed two research gaps: the lack of a systematic review of structured academic curricula, vocational training programs, and professional development pathways, and the absence of a formal approach to demonstrate how structured, research-oriented education can effectively bridge theory and practice.

Through a structured review of academic programs, professional training initiatives, competitions, and community-driven platforms, the study showed how diverse learning formats contribute to human capital development in this rapidly evolving domain. While progress is evident, particularly in Europe where initiatives such as EIT Urban Mobility lead the way, significant gaps remain. These include the limited availability of structured programs in the Global South, insufficient attention to accessibility and inclusivity, and underrepresentation of critical themes such as governance, ethical AI, policy, and cybersecurity. Expanding both the geographical reach and thematic scope of smart mobility education is therefore essential. The case study on teaching AI for Smart Mobility, developed using a design science approach, illustrates a replicable model for bridging theoretical instruction with hands-on innovation. Over five years, the course has engaged hundreds of learners and fostered public engagement through open knowledge sharing, with student projects gaining broad visibility and relevance via the AI4SM Medium hub.

Future research should explore the expansion of structured, research-oriented smart mobility courses to regions with limited access to such educational opportunities, particularly in the Global South where mobility challenges are often most acute. In addition, longitudinal studies are needed to evaluate the long-term impact of these educational models on learners’ professional trajectories, industry innovation, and policy development. Future iterations of the course may also incorporate international collaboration, for example through joint projects with students from other universities, in order to broaden perspectives, enhance global engagement, and strengthen the interdisciplinary learning experience. Such efforts will be essential to ensure that capacity-building initiatives not only address current skill gaps but also create sustained, inclusive pathways toward achieving the Sustainable Development Goals through smart mobility.