Children’s Perceived Ease of Use of a Projected Augmented Reality Game Designed for Balance and Coordination Training

Abstract

Developing balance and coordination skills is essential for children, especially those aged 4–8, but limited Health and Physical Education (HPE) programs in schools hinder effective training. Game-based learning and Augmented Reality (AR) technologies offer promising ways to enhance these skills by providing immersive HPE experiences. An AR exercise-game prototype was developed to train children’s balance and coordination, with 19 children aged 4 to 9 testing the prototype. Post-activity surveys revealed high engagement and interactivity ratings. The difficultly of the current prototype was found to be appropriately challenging for 4- to 6-year-olds. Feedback emphasized a preference for interactive, challenging elements, suggesting improvements in difficulty customization, visuals, and technical stability. Competitive play between multiple children testing the prototype encouraged repeated attempts, but also highlighted the need for improved tracking solutions and lab setup. Overall, the innovative design shows educational potential but requires further large-scale testing on a refined version to assess its effectiveness in balance and coordination training.

1. Introduction

Children are at particularly high risk of falls because they are still developing and have inadequate motor skills. The World Health Organisation [1] has highlighted falls as the leading cause of childhood injuries worldwide. Fall-related injuries place a huge financial burden on the world’s public health systems. At the same time, childhood obesity rates have been increasing worldwide and much of this increase is linked to the sedentary lifestyle of today’s children, which also increases the risk of diabetes, cardiovascular disease, hypertension, and cancer in children [2,3]. Children with better motor skills are more confident participating and staying active, while children with poorer motor skills tend to be less physically active [4,5]. Balance and coordination improve children’s motor skills, which not only leads them to spend time exercising, thereby reducing the risk of falls and injuries in their daily lives, but also improves children’s physical fitness and contributes to the prevention of non-communicable diseases [6,7,8,9].

The best time to develop motor skills is from birth to 8 years old, and from ages 4 to 6, children’s motor skills focus heavily on balance and coordination [10,11]. For children between the ages of 4 and 8, the main setting in which they engage in physical activity is at school [12,13]. Health and Physical Education (HPE) primary school teachers strive to improve children’s physical fitness and motor skills, yet the HPE subject is marginalised in schools and considered less academic and less important to the overall educational mission, which creates additional challenges for HPE teachers [14]. Although student enrolment has increased over the years, the proportion of professionally qualified teachers in HPE subjects in primary schools is small compared to other key subjects [15,16]. It is common for primary level teachers to be juggling multiple roles, such as having to teach both Arts and HPE lessons [17,18]. The quality of HPE lessons in primary school is also affected by factors such as weather, safety policies, student health, and resources [19]. Coordinating time among planning, teaching and assessment is the most difficult challenge faced by inexperienced HPE teachers [20,21]. Due to these barriers, which are often out of their control, it can be difficult for primary school teachers to deliver a comprehensive HPE lesson.

The root cause of a strong barrier to teaching traditional HPE may lie in the broader context of the HPE itself. Students are confronted with recurring themes centred on traditional games in HPE lessons, and this teaching appears to be focused on the performance of narrow motor- and object-control skills, rather than on balance, coordination or other skills that promote the holistic development of children [22]. Furthermore, such teaching fails to fully engage all students and ensures success only for those who are capable, while the focus of education shifts from providing all children with the skills for active adult life to that of pure athletic performance [23,24]. Kirk [25] has long warned that physical education needs to be reformed and that if multi-activity, sport-based forms of physical education are allowed to continue to dominate, physical education will die out. However, there is a misconception that HPE is the same as sport. Compared with teaching sports, HPE requires more diverse teaching methods, and the input and inclusion of educators, parents, students, and society as a whole [26,27].

Recent research highlighted the potential benefits of gamification in HPE: children are more enthusiastic and engaged in games than in fitness or motor-skills training [28,29,30,31]. These studies have found that gamification of the HPE can improve motor learning and control and enhance physical skills if the game is designed and implemented correctly. Furthermore, game-based learning assists with children’s cognitive development and allows them to fail gracefully [32].

Gamified education is undergoing a profound transformative, driven by the integration of emerging technologies such as Virtual Reality (VR) and Augmented Reality (AR) [33]. VR opens up unprecedented possibilities for education by creating fully immersive 3D virtual environments [34], while AR overlays virtual information onto real-world scenes in real time, providing a more interactive and relevant experience for learners [35]. While children’s toys in the past may have been cars or blocks, for today’s Alpha Generation (born between 2010 and 2024), their toys are likely to be smart devices [36]. For this generation of learners, technology is not an accessory, but a way of life [37]. The embracement of technologies is motivated by the positive effects on learner motivation and the need to adapt to a technologically immersive society [38].

There is potential for combining emerging technologies with gamified education and applying them to HPE for healthy children aged 4 to 8. This paper focuses on investigating the ease of use of an AR prototype application in challenging children’s balance and coordination skills while engaging their interest. VR and AR technologies have made significant progress in the field of rehabilitation, e.g., for stroke patients, elderly people in need of rehabilitation and children with cerebral palsy, helping to improve their motor function, balance and hand–eye coordination [39,40,41,42,43]. However, research on the use of gamification to improve motor skills (such as balance and coordination), which are important components of HPE subjects, is still relatively lacking for healthy children, especially for the age group 4 to 8 years. Compared to VR, AR is more representative of the real world in terms of immediacy, environmental awareness, device convenience and practicality [44], enabling children to learn in a real physical environment, in turn helping them to better relate learning content to the real world [45,46,47].

AR technology is ubiquitous, with examples ranging from the popular Pokémon GO smart phone game that continues to receive millions of dollars in revenue annually [48], to the Van Gogh Alive art exhibition that attracted more than 25 million visitors in 200 cities around the world [49]. Of relevance to this study is projection-based AR, which projects augmented digital content onto a wall [50]. Projection-based AR enables a spatial display to be separated from the body of the user, so there is no need for users to wear or carry any specialised equipment. Such freedom has been shown to provide a greater naturalness for the user’s movement and interaction with the surroundings [51,52,53]. The combination of AR technology and games encourages physical activity and engaging gaming experiences, and could positively impact physical health [54,55]. Piumatti et al. [56] emphasised that projection-based AR technology allows players to naturally interact with the game, and is the most effective way to provide them with strong physical engagement, with the support of immersion.

The purpose of this study was to develop a game prototype which uses AR technology to project virtual objects onto a wall, allowing children to develop their balance and coordination skills by interacting with the virtual objects by stretching their limbs. The main objective of this study was to investigate children’s perceptions of a projection-based AR balance and coordination game prototype to measure the value of the prototype regarding interactivity and engagement. This evaluation result not only shows the potential of combining AR and gamified education in the field of HPE, but also helps us to identify the strengths and weaknesses of the current projection-based AR prototype, which will serve as a reference for further development and large-scale testing in the future, to improve its usability in helping children train their balance and coordination.

2. Materials and Methods

2.1. Intervention

The prototype developed for this study focused on developing children’s balance and coordination skills. Balance and coordination play an important role in sport, providing the fundamental movement skills (FMSs) required for meaningful participation [57,58]. Static balance is the ability to maintain the centre of mass on a base of support, such as standing [9]. People still need to maintain balance in dynamic situations, such as walking and running. Even the simplest motor tasks require us to coordinate our movements throughout the body in time and space. For example, to achieve a smooth and efficient moving gait while walking, we need to coordinate the simultaneous movements of our arms, legs, torso and head [59]. Balance and coordination are complementary, and studies have shown that children’s coordination is related to their balance, and that improved balance enhances children’s coordination during physical activity [60,61].

In designing the prototype, the researchers considered FMSs for 4 to 8 years old, such as object control skills of catching, kicking, and stationary reach; locomotor skills such as walking and jumping; and stability skills such as static and dynamic balance [62]. An exercise game (exergame) is defined as technology-driven physical activities, such as video game play requiring participants to be physically active to play the game [63,64]. The prototype designed for and evaluated in this study requires gentle physical activity; its main purpose was to encourage children to stretch their limbs actively during the game, to maximise their range of motion, and to develop their balance and coordination.

During testing the prototype, the child stands in front of the wall, so they can see their shadow within the game projection (Figure 1). During play, coloured and icon-coded objects fly randomly from the left and right directions, from random heights. The child must coordinate their limbs while maintaining balance and scoring points by touching objects associated with specific limbs. For example, the right hand is required to touch the orange object with a sun icon, while the left foot must touch the white object with a smiley faced icon. Touching an object with the correct limb earns 100 points and touching an object with the incorrect limb loses 50 points. The duration of each round is 1 min, and the exergame shows the total score when each round is finished.

The game is built using the Unity engine. The child’s movements are captured by an Intel depth-tracking camera, which detects collisions between limbs and projected icons in virtual space by recording the 2D coordinates of skeletal joints in each frame.

At the beginning of the game, the camera detects the children’s heights to adjust the maximum vertical height at which objects could appear within the boundaries of the projected game image. This customises the experience for players of varying heights, so they are e not disadvantaged in reaching high objects. Additionally, an upbeat soundtrack accompanies the game, and whenever a virtual object is hit correctly or touched by mistake, a corresponding sound effect will play.

2.2. Methods

This exploratory research aimed to evaluate the usability of an AR-based exergame prototype in enhancing balance and coordination skills among children aged 4 to 8 years old. The study involved testing the game in a controlled lab environment, where children’s movements were tracked using depth-sensing technology, and their experiences were assessed through a post-activity survey.

The prototype shown in Figure 2 was set up for testing in a controlled lab with an area of approximately five-metres squared, with an empty wall located directly in front of the children, a data projector to project the game onto this wall, and an Intel Realsense D435 (Intel, Santa Clara, CA, USA, 2018) depth-tracking camera behind the children to track their limb movements. The depth camera was located in the same position as the projector, about five metres from the projected image on the wall of the room. The participant is approximately one metre from the projected image on the wall (see Figure 2a). Children needed to stay around the centre of the display to ensure their movements were correctly tracked by the depth-tracking camera. Calibration process of the depth camera: the depth component of the 3D coordinates of the skeletal joints is used for this prototype, which is recorded in a log file for future analysis. This prototype only uses the 2D coordinates of the skeletal joints (see Figure 2b). The estimation of 2D coordinates are calibrated by adjusting the size of a “bounding box” in the captured video stream that represents where the projected image is located. The ratio of the total image size compared to the configured bounding box size is then used to calibrate the estimated 2D skeletal-joint coordinates. Height adjustment: the Unity application constantly monitors the highest point of any pose joint, and if that value exceeds the current maximum height (which defaults to that of a small child), a new maximum height is set, and the locations of the arm-height object spawners and leg-height spawners (which determine where the game targets spawn from on either side of the game) are adjusted accordingly as the game progresses to ensure children of different heights can reach all objects.

The prototype tests were conducted on a family basis, with a minimum of one child participating at a time and a maximum of four children. Children would watch a brief tutorial together to familiarise themselves with the objectives and mechanics of the game before playing the game individually. Each child would play multiple rounds, but switch after each round (1 min), with a maximum of 10 min allotted for each child. The testing process was supervised by parents and the research team throughout, and at the end of the test each participating child filled out a questionnaire about the game prototype, and some younger children (e.g., 4 to 6 years old) received assistance from their parents, if necessary, to comprehend the question correctly.

Although previous research has demonstrated the usefulness of AR technology and Exergames in the rehabilitation of FMSs such as balance and coordination [65,66], research combining AR technology and Exergames to help train balance and coordination in all healthy 4- to 8-year-old children is difficult to find. The Davis’s Technology Acceptance Model (TAM) is a widely used framework in the early stages of system development, originally developed to assess end-user acceptance of IT products [67,68]. TAM (Figure 3) recognises that perceived usefulness and perceived ease of use are key factors influencing user acceptance, and that users’ perceptions of the ease of use of a system influence their perception of the usefulness of the system [69,70].

In the initial development phase of the prototype, the research team used TAM as a guide to test the prototype’s functionality by evaluating children’s perceived ease of use of the projection-based AR exergame prototype in developing balance and coordination skills. This will help us improve the prototype in the future, improve its usability and usefulness, and conduct larger-scale testing. This explains why we focused on children’s perceived ease of use of the AR game designed for balance and coordination training, and how the research questions were determined.

2.2.1. Research Questions

The overarching research question that guided this study was the following:

“How can the ease of use be improved in a projected augmented reality exergame designed to help children train their balance and coordination skills?”

The following sub-questions aided with answering the research question:

• What are children’s overall perceptions about ease of use and usefulness of the projected AR exergame prototype to practice balance and coordination?
• How can projected AR exergames be made more engaging and motivating for children in training balance and coordination skills?

2.2.2. Sample

As previously described, age 4 to 8 is the most critical time for children to work on their balance and coordination skills [10,11]. Accordingly, the research team recruited a total of 19 children aged 4 to 9 years old (N = 19, M = 6.53, SD = 1.54), through convenient sampling, to participate in the survey. Among them, 18 children met the target age (4 to 8 years), and one was 9 years old, who was also included in the data because of their proximity to the target age. When recruiting children to participate in the survey, we sent parents two plain language information statements, which were different: one for parents and one for children. Before the children participated in the survey, the parents had to ensure that they had fully explained the content of the research project to the children and that they and the children had received satisfactory answers to any questions before signing the consent form. Throughout the testing of the prototype, a FedUni qualified first-aider was present throughout the test site to ensure the safety of the participants. Ethical approval from the Federation University’s Human Research Ethics Committee was granted (no.: 2022-166) for this research.

2.2.3. Survey Design

As mentioned previously, this survey required children to complete a questionnaire after testing the prototype. This questionnaire began with each child creating their own pseudonym used for anonymity. It then consisted of six single-choice questions generated from a five-point Likert scale and five open-ended questions. The Smileyometer (Figure 4) was also incorporated into a five-point Likert scale and the children were asked to rate the prototype’s game elements, difficulty levels, and their general feelings about the gameplay.

The Smileyometer uses emotional facial expression to gauge levels of agreement and is often used in surveys with children because it is easily understood by them [71,72]. We also used age-appropriate language to supplement each smiley emotion ranging from awful (rating of 1) to awesome! (rating of 5). There were different rating scales regarding the difficulty of the game and how often they would like to play the game, which are explained in detail in the results section. Open-ended questions were used to explore the children’s rating criteria in the Likert scale section, as well as their views on the ease and usefulness of playing the prototype.

For younger children who cannot read or write (e.g., 4 to 6 years old), parents can provide assistance if necessary, but their role is limited to helping the child understand the question, and not influence the answer. Parents should only provide assistance when necessary (e.g., when the child has to write a long sentence in response to an open-ended question) and should record the child’s verbal response as it is. The whole process of completing the questionnaire is supervised by the researchers, the level of parental involvement is recorded, and the researcher checks the data again after receiving the questionnaire.

2.2.4. Analysis

Embedded Sequential Design (ESD) described in (Figure 5) was used to assist in analysing the data from the questionnaire. The quantitative analyses were able to provide an overview of children’s perceptions of the prototype, helping us to identify some key findings and trends, with qualitative analyses contributing to an in-depth understanding of these findings and trends [73,74].

The embedded sequential design process employed in this study is outlined below:

• Quantitative data collection: Likert scale questions, recording each child’s rating for analysis.
• Quantitative data analysis: descriptive, frequency and cross-tabulations were used to generate a series of visualisations to identify key findings.
• Qualitative data collection: thematic analysis was used to collate and analyse children’s responses to open-ended questions. Words or phrases were compared and synonyms were considered to categorise answers and generate a keyword-frequency table.
• Qualitative data analysis: a keyword-frequency table was used in the formation of themes that explain and complement the results of the quantitative data.
• Comprehensive analysis: combining the results of the quantitative and qualitative analyses with observations of the children’s performance during their participation in the test provided a more complete picture of the children’s view of the prototype.

3. Results

By analysing questionnaires completed by 19 children aged 4 to 9 years old after taking the prototype test, the researchers were able to gain a general understanding of the perceptions of children aged 4 to 8 years old regarding the use of the prototype to exercise their balance and coordination skills; this information provides recommendations for the future design of AR Exergames and related tests to engage children of the target age group in active and sustained participation in balance and coordination exercise.

3.1. Likert Scale Results

When analysing questions one, three, four and five, the Likert scale of 1 = Awful, 2 = Bad, 3 = OK, 4 = Good, and 5 = Awesome (presented as Smileyometer), was used to describe the child participants’ perception about a particular aspect of the game.

The researchers were interested in the children’s first impressions of the game after testing the prototype. All 19 children (N = 19) answered every question. Question one asked children to give an overall appreciation score of the game and the results were M = 4.37, SD = 0.761 (Figure 6). More than 75% of the children (N = 16) who participated in the game-prototype test rated the game between 4 (Good) and 5 (Awesome!). The children seemed to enjoy the game and were interested in the gameplay of using their limbs to interact with objects projected onto the wall (corresponding to question 4), with the results M = 3.84, SD = 0.898, and more than 50% of the children (N = 12) rated the gameplay as a 4 (Good) or a 5 (Awesome!). Examining question 5 and 6, the trend in their ratings of the images in the game (M = 3.89, SD = 0.809) was similar to question 4, but more children rated the sounds (M = 3.84, SD = 0.765) as 3 (OK) or 4 (Good).

It is important to note that the second Likert-scaled question used different adjectives. In all other questions a high rating is desirable, but not for the second question. The second question asks the children how they feel about the difficulty of the game, with scores of 1 to 5 corresponding to Too Hard (sad face), A Little Hard (neutral face), Just Right (happy face), A Little Easy (neutral face), and Too Easy (sad face), and the results were M = 3.11, SD = 1.049. In terms of the difficulty rating (which corresponds to question 2), the median score was 3 (Just Right), and the mean score of 3.11 is also very close to the ideal score of 3, which shows that the game is almost correct in terms of difficulty. While there is a clear trend in the ratings of the questions we mentioned earlier, the distribution of children’s ratings of the difficulty of the game is spread out and symmetrical, which means that there are still children who think the game is too hard and children who think it is too easy. This can be explained by the fact that the prototype had a single difficulty, and the natural development of children from 4 to 8 emphasized the fact that older children felt less challenged. Understanding that the current difficult is suitable for 4- to 6-year-olds will help us in the future to balance the game for multiple ages, with dynamic difficulties.

The third question asked children how often they would play this game if available at their school, with scores of 1 (sad face) to 5 (happy face) corresponding to Never, A Little, Sometimes, Lots, and Always, with the results M = 4.00, SD = 0.943. The children presented interest in continuing to play the game at school in the future, with at least 75% rating the level of interest between 3 (Sometimes) and 5 (Always!).

3.2. Open-Ended Results

While quantitative analysis of the data from the Likert scales already gives us a general idea of the children’s combined views on the prototype, analysis of the responses to the open-ended questions can corroborate or refute the results of the quantitative analysis and provides an explanation of the children’s rating scale [75,76]. Children often tend to give more information when answering open-ended questions [77], and we found that children’s response to one question often also contained lots of information about other questions, and using thematic analysis can help the researcher to disaggregate and categorise the responses, helping us to identify the patterns and themes that underlie them [78,79].

Singular words or short sentences from children’s responses were compared, and synonyms were considered to code responses into different categories, resulting in

Table 1. Response Frequency per Key Theme.

Q7: Likes?	N	Q8: Dislikes?	N	Q9: Suggestions?	N	Q10: Favorite Games?	N
Interactive	7	Bugs	6	Progressive challenges	4	Sandbox video games	6
Movements	5	Short-Playtime	2	Themes	4	Board games	5
Instructions	4			Different shapes	3	Action RPG video games	4
Challenge	3			Better tutorials	2	Sport games	3
Shadow	2			More colourful	2	Puzzle games	3
Score	2			Use more of the body	2	Schoolyard games	2
Educational	2					Platform video games	2
						Exercise video games	2
						Racing video games	1
						Simulation video games	1

.

The prototype is aimed at children, so understanding the types of games and game elements they enjoyed will help to design the AR Exergame to be more in line with children’s preferences, and to consistently engage them. Therefore, in question 10, children were asked what games they liked and why. In response to this question, most children filled in more than one game and some of these video games were categorised into more than one game type. The most popular type of game for electronic devices such as computers, mobile phones and game consoles was Sandbox video games (N = 6), with multiple citations of Roblox (Roblox Corporation, San Mateo, CA, USA, 2006) and Minecraft (Mojang Studios, Stockholm, Sweden, 2009), followed by Action Role-Playing Game (RPG) video games (N = 4). Reflecting on these popular games, we found that the two types of games had three significant things in common. Firstly, most of the games included a combat mode; secondly, players can create an avatar; and thirdly, the games were somewhat exploratory. We also note that other games also allow players to create avatars; for example, in some Platform video games (N = 2) and Simulation video games (N = 1) children can dress up their characters. After research, we found that having avatars in the game is particularly important, because even without a reward and punishment system, people will imitate the behaviour of others through learning and observation, and the more similar the avatar is to the user, the more likely the user is to succeed in learning the avatar’s behaviour [3,80]. As such, we should consider how to include avatars that resemble the images of children in all Exergames, which would motivate them to participate in training or exercise.

Not all children are interested in traditional video games, and some are more inclined to engage in other forms of game activities. Of all the game types, Board games are the second most popular game type for these kids. They mentioned that they liked Board games (N = 5) because they could choose their own character and use strategies in the game, while Puzzle games (N = 3) also corresponded to this reason for strategy. The children also mentioned their favourite Sport games (N = 3): football, basketball, football and skipping. Exercise games (N = 2) also featured in children’s responses; for example, Loo (aged 6) said he liked to play Wii Boxing, where he could stand in front of a monitor connected to a console and practise boxing with the controller in his hand. In terms of why children enjoyed these sports, their responses related to physical movement, strategy and running. Among the children’s responses were also Schoolyard games (N = 2) like Tag, Hide and Seek, and Rock Paper Scissors. They expressed enjoyment in engaging with these games alongside their peers, emphasizing that is a crucial factor, which was also mentioned by some of the children who liked video games or board games.

In question 7, children were asked what they liked about the game. The most common answers revolved around the theme of the Interactive nature of the game (N = 7), followed by different types of Movements (N = 5). Children enjoyed the unique interactive gameplay of the prototype. Through observation, researchers identified that children were interested in punching and kicking objects and being active. When some children discovered that they could use their shadow to interact with objects, they would assume fighting poses in front of a wall, imagining themselves as a martial arts hero and hitting objects with their punches. Participant Body Movement (aged 7) indicated that they felt like a ninja when kicking objects. Charli (aged 8) associated their movements with doing ballet. In view of the features of children’s favourite game types, the inclusion of a combat mode and allowing children to create their own avatars according to their imagination could be considered in future game designs.

Encouragingly some children enjoyed this game because they found it challenging (N = 3) for their balance and coordination. For instance, Eric (aged 5) was observed to initially struggle to distinguish left from right, but after several rounds, had learned to associate the colour and shape of objects with the correct limbs. Furthermore, Lily (aged 7) and Mia (aged 7) found the game Educational (N = 2). It is worth noting that although only two children explicitly indicated that they wanted a higher Score, the researchers observed that when there was more than one child present, such as siblings, in the testing lab, then they became competitive in terms of achieving a high score.

In question 8, children were asked what they did not like about the game. Almost one-third of children mentioned Bugs (N = 6), stating that sometimes when they touched objects with the correct limbs, they did not score or were deducted points, as if the game was not working properly.

This can be attributed to four possible reasons due to the games design and technology:

• Some children were wearing clothes or shoes in a similar colour to the wall, which prevented the depth camera from tracking their movements well, due to lack of contrast.
• Sometimes children moved too close to the wall and too far from the depth camera for the camera to perform accurate tracking.
• At times, features on the back of the children’s heads (such as hair movement, hair colour, or artifacts introduced during the image-compression process) were misinterpreted by the depth camera and would sometimes incorrectly identify the wrong hand or foot, resulting in points being deducted.
• The differences in lighting in the game room occasionally affected the depth camera’s ability to accurately track the children’s movements.

In question 9, the children were asked how to improve the game. Four children explicitly stated that they would like it to be more challenging and have different levels of difficulty. They also suggested how to make the game more challenging. For example, Loo (aged 6) wanted the game to have more than four objects so that they need to memorise more objects and think about which hand or foot to use. Charli (aged 8) wanted multiple objects flying across the screen at the same time. Two other children mentioned they would like to use more parts of their bodies to hit objects. In the prototype, objects were set to only appear within reach of the child’s height. Body Movement (aged 7) stated that she would like to be able to jump up and use their head to hit objects. A notable finding is that from age 6 onwards, after one or two rounds of the game, children begin to ask for more challenges, but it is still challenging for children aged 4–5, which coincidentally suggests that the prototype is adequate for children aged 4–5. We believe there are two reasons why the prototype is appropriate for this age group. One reason for this is that some children in this age bracket have difficulty distinguishing between left and right, and another reason is that they are slower than older children in understanding the gameplay [81,82]. These children need more instruction and time to understand game play. Princess G (aged 4) thought it would be good if the researchers demonstrated how to play the game. Charli (aged 8) suggested that perhaps when the left-handed/right-handed object appears, there should be an in-game tone that says Left Hand/Right Hand.

The children also made other suggestions. They wanted the game to include a variety of Themes (N = 4) or to make the background of the game more Colourful (N = 2). Lily (aged 7) stated that they would like the background of the game to be more colourful, and not just white. Princess G (aged 4) stated that they would like the background to have more pink, and Eric (aged 5) stated that they would like to use pictures of tractors or trucks. Design-wise, it is best to keep the background of the game as simple as possible, to enable the depth-tracking camera to track children’s movements more accurately. Nevertheless, some background variations could be achievable. It is important to note that, although rate of motor development varies widely, there is a developmental difference across the age band in this cohort with the youngest aged 4 years and the oldest aged 9 years [83]. All the participants in this study played the same game at the same level. Given the developmental difference across the age band, it is unsurprising that the younger ones found it challenging while the older ones wanted to see more levels that would challenge them.

4. Discussion

This study investigated children’s perceptions of the usability and usefulness of the prototype of a Projection-based Augmented Reality (AR) Exercise Game (Exergame) for improving children’s balance and coordination skills. The aim was to conduct this study as a “proof of concept” to investigate how well children would interact with the game. The intention was to take feedback from the children, address any technical issues with the hardware and software, and then, in the future, to further develop the tool and test again with a larger cohort. As a result, the participant numbers are low, but this is reasonable for an initial pilot study. According to the questionnaire results, we found that there are also children who play commercial Exercise Games at home. To distinguish them from the Exergames developed based on the research, they are more often referred to as Active Video Games (AVGs). Previous studies have shown the effectiveness of AVGs in improving children’s motor skills (including, but not limited to, balance and coordination) and attitudes towards physical activity, but the AVGs can only be guaranteed with professional on-site guidance [84,85,86]. The current research results show that the overall evaluation of the AR prototype is favourable. The children were particularly interested in the gameplay of using physical movements to interact with virtual objects projected onto the wall, and expressed willingness to continue playing such games.

Previous research [6,7,8,9,87,88] has shown good balance and coordination skills can improve children’s fundamental movement skills (FMSs), enable children to participate in physical exercise on a sustained basis, and reduce the risk of children getting injured, being overweight, and suffering from non-communicable diseases, while also providing children with the skills, knowledge and attitude to stay active for life. The ages of 4 to 8 are the most critical period for children to develop balance and coordination skills [10,11], so we need to determine whether the game is appropriate for children of this age. The researchers had pre-assessed the difficulty of the game for the target age group of 4 to 8 years old before the prototype test and believed that by having children test it, the difficulty and game mechanics could be addressed in future iterations. What was found was that the current game prototype was somewhat difficult for children aged 4 and 5 because they needed to distinguish left from right or spend more time understanding the gameplay [81,82]. Children aged 6 and above become proficient after one or two rounds of play and demand increased difficulty, but children of the same age have different perceptions of the difficulty of the game, which shows that the difficulty of the game is not only related to age, but also to their own gaming/sports experience. In developing games, varying difficulty levels should be considered and subdivided, and when applying the game, to ensure that each child is assigned the appropriate individual difficulty level.

The children showed great enthusiasm for using the game to learn new things, as evidenced by the fact that the children emphasised the word strategy when referring to their reasons for enjoying a particular game, and some felt that the prototype was educational in that it allowed them to learn to differentiate between left and right. During the game testing, we also observed that the game has a competitive atmosphere, and the children appear more relaxed and active in the group setting, which in turn encourages children to make multiple attempts to achieve higher scores. This may cause concern among some educators, as an excessive focus on winning may encourage unacceptable behaviour in the school environment [89,90,91]. However, competition is an important concept in sport and education programs, and positive competition motivates children to make progress [92,93,94]. The guidance of the teacher, combined with virtuous competition, instils in the children the notion of “striving together” and “respect for the opponent” [95,96,97]. Just as children have repeatedly emphasised that they like a particular game because they like playing with their peers, this game is suitable for use in schools and, with the right guidance, may further encourage children to participate in training.

During the prototype testing, the projector was installed behind the children and below their line of sight, casting their full shadow on the wall. This was a deliberate design decision, so that it felt like the player was controlling the shadow and did not require players physically approaching the wall to hit objects on a hard surface with bare hands. Children shuffled left and right while facing the wall, so their vision was not affected by the light of the camera, due to them not facing it. Additionally, brightness settings from the projector were dimmed to ensure that children turning around during play would not be affected by high brightness levels. With the children’s shadows appearing on the wall, this visual feedback provided by the shadows helps the children observe their movements during the game [98,99], thereby assisting their training. This is also evident in game testing, where children use their shadows to interact with objects or strike various poses.

As mentioned in the results section, the depth-tracking camera occasionally failed to accurately track the children’s movements. The main reason for this is that some of the limitations of depth-tracking cameras and associated software when tracking human skeletal models, were only discovered during testing, and is a primary reason for testing ongoing prototypes. Factors such as insufficient contrast between the child’s clothing colour and the game background, and the child being too close/too far away from the depth-tracking camera, all affect the accuracy of tracking the limbs. The researchers took immediate action as soon as the problem was discovered: firstly, the colour of the game background was modified, according to the specific situation of each child; secondly, a boundary was drawn on the floor to keep the children at an optimal distance from the camera. These are only temporary solutions. In the future, the accuracy of movement tracking can be improved by improving the game design; for example, the game will sound an alarm when the depth-tracking camera detects that the child has left a fixed area. For the issue regarding misinterpretation of data (such as incorrectly detecting the child facing the wrong direction), more testing needs to be carried out to find the optimal room layout, spacing between camera, child and wall projection, and perfect lighting to reliably view the player, without reducing visibility of the game projection. An additional tracking camera from a different angle may also help with data collection and during poor lighting conditions.

5. Conclusions

An Exergame for children needs to appeal to their interests to attract them to the training of balance and coordination on a sustained basis. While the sample size of this study was small, with a total of 19 children participating in the tests (18 aged between 4 and 8 years, and one aged 9 years), it provided valuable exploratory insights. The study primarily aimed to evaluate children’s acceptability and perceived experience of a prototype Projection-based Augmented Reality Exergame and to gather preliminary data to guide larger research. The children’s suggestions will not only help in the further development of the prototype, but also serve as an important reference for improving the usability and attractiveness of this type of game.

According to the results of the questionnaire, the children’s ratings of the in-game sounds were lower than their ratings of the game technology and visuals. Some children said that they were focused on scoring points in the game, and did not pay much attention to the background music and sound effects. This may be because the overall immersive experience masks the audio in games, or because the game soundtrack is repetitive and lacks variety [100,101]. The children also expressed a desire for a more colourful game background and wanted the game to have different themes. Some wished for the game background to have pictures of tractors or trucks, while others wanted more pink elements in the background. When testing the prototype, they imagined themselves as a superhero, ninja or ballerina. The types of games they liked were diverse: some liked sports that emphasised physical activity, while some preferred board games or puzzles that emphasised strategic thinking.

These findings highlight the diverse personalities and strengths of children. When further developing the AR Exergame in the future, consideration should also be given to incorporating the features of various types of games. This includes visual and auditory themes that appeal to children, and in-game tasks that challenge not only their FMSs, but also their cognitive and strategic-thinking abilities. This would make the AR Exergame more inclusive and relevant to the interests of more children, while responding to the key objective of Health and Physical Education (HPE), which is to provide all children with the skills, knowledge and attitudes to be physically active for life through physical activity for holistic development [102].

Although this study did not include a control group to compare Projection-based AR Exergames with traditional training methods, its findings underscore the need for further investigation. Future studies should incorporate control groups, to rigorously evaluate the relative effectiveness of AR Exergames in improving balance and coordination skills. Addressing this gap will strengthen the evidence base for integrating AR technology into HPE subjects.

Standardized protocols for assessing improvements in FMS, particularly balance and coordination, should also be developed. This would facilitate more meaningful comparisons between AR Exergames and traditional methods, providing stronger support for their educational and developmental value. Such research would enhance the integration of AR technology into kindergartens and primary schools, ensuring games are both engaging and effective.

While evidence supports the effectiveness of AR technology and Exergames in rehabilitation and motor-skill training, gaps remain in understanding the specific benefits of projection-based AR Exergames for children. Future research should include a larger and more diverse cohort of participants and more sports scientists, to add their knowledge to the existing team of technologists and researchers. This will help design a projection-based AR Exergame that better fits the HPE environment, ultimately supporting children’s physical and cognitive development.