Motivating Elderly People to Exercise Using a Social Collaborative Exergame with Adaptive

Motivating Elderly People to Exercise Using a Social Collaborative Exergame with Adaptive Difficulty.

Dale Cantwell, Daire O Broin, Ross Palmer, Greg Doyle

Institute of Technology, Carlow, Ireland

{dale.cantwell, daire.obroin, ross.palmer, greg.doyle}@itcarlow.ie

Abstract: The European population is ageing and physical decline of elderly people has become a significant issue. A lack of physical exercise results in decreased strength, flexibility, and the loss of bone mass. These are risk factors that cause elderly people to fall, which can result in injury and mortality. The risk of falling can be reduced by partaking in physical activities such as walking. Many elderly people don’t exercise as they suffer from social isolation. This affects physical and mental health, causing illnesses such as depression and may result in low motivation to exercise. This paper proposes a solution to these issues: a social collaborative online walking game, the goals of which are to engage and motivate the elderly to exercise frequently and to increase their cardiovascular fitness and muscular strength. We have developed a prototype in HTML5, designed for ease of use for elderly people, which uses the Microsoft Kinect motion camera, the preferred controller selected by the user group.

The principal advantage of this solution is it uses adaptive difficulty to provide dynamic balancing, allowing users with different skill levels to play together. User group tests suggested that static difficulty settings are not flexible enough to suit the group’s range of skill levels. Each player has an individual difficulty level which is conducive to providing an enjoyable experience for all players. In particular, this facilitates intergenerational gameplay: elderly users can play with their children and grandchildren without comprising their individual enjoyment. Another advantage is that it promotes a gradual rate of improvement that reduces the risk of the user overexertion during exercise. The effects of the walking game are currently being evaluated with an elderly user group.

Keywords: Exergames, Social collaborative games, Adaptive difficulty, Serious games, Microsoft Kinect

1.  Introduction

The European population is ageing, by the year 2060 it is projected that the sixty-five to eighty demographic will increase from 18% to 30% of the population (European Commission, 2011). Elderly people suffer from diminished physical fitness which results in muscle weakness and reduced strength, balance, flexibility and coordination; these are risk factors causing elderly people to fall (Scott et. al. 2004). It is possible to prevent falls by partaking in exercises such as walking training (Sherrington et al., 2011). Social isolation, caused by factors such as living alone, death of family or friends and health issues, is a growing concern in the elderly demographic. Elderly people who suffer from social isolation may be less motivated to exercise (Burbank and Riebe, 2001).

A solution to this problem is being developed as part of Join-In, a project of the Ambient Assisted Living (AAL) Joint Programme is an online social exercise game where a primary goal is to motivate the users to exercise frequently (Join-In, 2012). The main considerations that have emerged from user groups relate to social interaction, game balancing and usability.

Social interaction is a primary motivation for elderly people to exercise (Stead et al., 1997). Another social aspect that may motivate elderly people to play the exergames frequently is facilitating intergenerational gameplay, so that a user may play with children and grandchildren (Khoo et al., 2006). Games developed for the Join-In project promote social interaction by featuring competitive and cooperative game modes.

Because of the variation in users’ skill levels, a game that supports intergenerational gameplay provides the added challenge of balancing the game to be suitable for each user without compromising individual enjoyment. Usually a game is balanced by adjusting difficulty statically or dynamically, although the static model may lead to less than optimal uptake of the game in the case where players do not fit well with the expected stereotypes (Westra et al., 2008). The Join-In project implements a Dynamic Difficulty Adjustment (DDA) system to address this issue.

Usability is a vital element in the design of a video game for an elderly audience (IJsselsteijn et al., 2007). An example of this is the choice of controller, as the target audience are inexperienced at playing games, the game’s usability may be improved by selecting a controller that is easy to use. User group sessions informed the choice of the Microsoft Kinect as the input device for this game.

The remainder of the paper describes related work and the requirements gathering design of the game, before wrapped up with an outline of future work.

2.  Related work

Commercial exercise games and research projects featuring walking challenges designed for elderly people include SilverFit (Rademaker et al., 2009), which uses a time of flight camera, and The Horsepower Challenge which is a collaborative walking game that records the number of steps each user takes in a group, and allows users to compare steps using an online social platform (Eiriksdottir et al., 2010).

DDA is a key feature of the Join-In project allowing users with different skill levels to play together. SiN Episodes: Emergence is a commercial first person shooter which features a DDA system comprised of four key parts: statistics, advisors, the decision maker, and gameplay variables (Kazemi, 2008). These parts identify when the player is failing and the advisors suggest a method of adapting the difficulty to assist the player. The Hamlet DDA system uses statistical metrics to analyse game data and extrapolate the player’s future state (Hunicke and Chapman, 2004). The goal of the system is to detect when a player is at risk of repeatedly entering an undesirable state, not having the necessary resources to progress and intervene in an attempt to return the user to engaging interaction states.

3.  The Game

The primary goals of developing this serious game are to improve the physical condition of elderly users, motivate the users to exercise frequently, and provide a method for social interaction and facilitate intergenerational gameplay.

To develop a game that suits the needs of the target audience, a series of user group gaming sessions were held. It emerged from the sessions that walking is a primary hobby for the group, and requirements gathered included the possibility for cooperative and competitive play, the ability to communicate with other players and adjustable difficulty levels so that the users do not feel pressured. The user group demonstrated a preference for the Microsoft Kinect as they found it quick to learn and easy to use.

On the basis of the preliminary user testing results, a decision was made to develop a collaborative walking game. The game was developed in JavaScript and is playable in HTML5 compatible web browsers, and uses the Microsoft Kinect as the primary input device. Skeletal data is streamed from the Kinect on the client’s machine to a HTML5 compatible browser which is mapped to an in game avatar in the game world. Each user is represented by a customisable in-game avatar. This self-avatar representation is used to increase the user’s immersion in a virtual environment, achieved by translating the user’s sense of proprioception from the real world to the virtual world (Slater and Usoh, 1994).

The basic goal of the game is to walk a set number of steps during each play of the game (Figure 1).The user’s steps are recorded using the Kinect. To add an element of challenge (necessary for any game), a user must attempt to walk to a rhythm and avoid obstacles such as potholes. Every incorrect or missed step results in the user’s avatar slowing down which is indicated by changing the colour of the avatar’s body from green to red.

Each player attempts to walk at the same pace as the other players in the game; players of a high skill level will see additional obstacles to those presented to a player of a low skill level (Figure 2).

Figure 1: In-game screenshot of two players walking. The white bars along the path represent the rhythm and each black hole is an obstacle. The bar on the left represents the progress towards the goal. The two users must take a step to avoid each pothole.

Figure 2: In-game screenshot of a high skilled player (left) and a low skilled player (right). The target goals and number of obstacles differ.

The adjustment system developed for this project is similar in design to the SiN Episodes Personal Challenge system (Kazemi, 2008). There are several methods of possible adjustment each represented by an advisor. An example of an advisor used is rubber band AI, a method of difficulty adjustment typically found in racing games such as Pure and Mario Kart (Jimenez, 2009, Olesen et al., 2008). Other advisors adjust different game settings in an attempt to increase usability. One such advisor adjusts a threshold value which is used to determine how high the user raises their knee in order to perform a step. This is adjusted based on the user’s motions, which allows the intensity of the exercise to be adapted to suit each individual. Another advisor checks if the player is making the correct motion but missing the time, if that is the case this advisor recommends reducing the precision required for taking a step.

The game is currently in the testing phase, the purpose of the investigation is to assess the effect the adaptive component has on the game’s motivational impact, as well as evaluating the game’s usability and the effect on the user’s physical condition and social connectedness.

Each participant is issued a set-top box with the game and any necessary software set up as well as a Microsoft Kinect for playing the game. Half of the participants are issued a version of the software with the adaptive difficulty system and the other half are presented with the same game with three static difficulty options.

The system tracks user data such as how often the game is played, how much time is spent per session, usability data such as the success rate for completing tasks, the time taken to complete each task and keeps track of adjustments made by the difficulty adjustment system. In addition to the recorded data the participants are asked to fill out questionnaires at regular intervals to keep track of their progress. The questionnaires include the Intrinsic Motivation Inventory (IMI), the Software Usability Measurement Inventory (SUMI) and the Social Network Index (SNI). The user’s physical improvements, specifically any improvements to balance can be measured using the Berg Balance Scale (BBS).

The comparison of the two versions of the game should reveal the effect of the adaptive difficulty system. The aim of the study is to investigate whether the group using the adaptive version will be motivated to play the game more frequently and for longer periods of time

4.  Conclusion and future work

This paper presents an overview of the design of a collaborative walking game for an elderly user group that utilises an adaptive difficulty system to increase the motivation to exercise. The Dynamic Difficulty Adjustment system developed aims to provide a personalised level of difficulty tailored for each individual to promote cooperative, competitive and intergenerational gameplay.

Future work includes improving the adaptive system by increasing its ability to assess the user’s skill level, provide additional methods of adjustment and improve the chances of the system selecting the best method. New modes may also be added in later versions of the game to specifically improve physical skills, such as the addition of balance challenges. This game will be embedded into the Join-In online social platform along with other cognitive games and exergames all designed for elderly users (Join-In, 2012). The results of the user testing phase will influence any further changes to the software and any modifications to future studies with the user group.

References

Burbank, P. M. Riebe, D. R. (December 2001). Promoting Exercise and Behavior Change in Older Adults: Interventions with the Transtheoretical Model. Springer Publishing Company; 1 edition. pp. 36-37.

Eiriksdottir, E., Kestranek, D., Miller, A., Poole, E. S., Xu, Y., Catrambone, R. and Mynatt, E. (2010). Stepping Outside the Classroom:Fitness Video Games For K-12 Settings. CHI 2010, April 10–15, 2010, Atlanta, Georgia, USA.

European Commission. (September 2011). The 2012 Ageing Report: Underlying Assumptions and Projection Methodologies. European Economy 4 | 2011.

Hunicke, R. Chapman, V. (2004). AI for Dynamic Difficulty Adjustment in Games. Challenges in Game Artificial Intelligence AAAI Workshop, San Jose.

IJsselsteijn, W. Nap, H. H. de Kort, Y. Poels, K. (2007). Digital game design for elderly users. In Proceedings of the 2007 conference on Future Play, pp. 17-22.

Jimenez, E. (2009) The Pure Advantage: Advanced Racing Game AI, [Online], Available: http://www.gamasutra.com/view/feature/3920/the_pure_advantage_advanced_.php?print=1 [19 Apr 2012].

Join-In. (2012). [Online], Join-In Senior Citizens Overcoming Barriers by Joining Fun Activities Available: http://www.helmholtz-muenchen.de/en/join-in/home/index.html [19 Apr 2012].

Kazemi, D. (2008). Metrics and Dynamic Difficulty in Ritual's SiN Episodes, [Online], Available: //orbusgameworks.com/blog/article/70/metrics-anddynamic-difficulty-in-rituals-sin-episodes-part-1 [20 Apr 2012]

Khoo, E.T. Cheok, A. D. Nguyen, T. H. D. Pan, Z. (2008). Age invaders: social and physical inter-generational mixed reality family entertainment. Virtual Reality, 12(1), pp.3-16.

Olesen, J.K. Yannakakis, G.N. Hallam, J. (2008). Real-time challenge balance in an RTS game using rtNEAT. Computational Intelligence and Games, CIG '08, IEEE Symposium. pp.87-94.

Rademaker, A. van der Linden, S. Wiersinga, J. (2009). SilverFit, a virtual rehabilitation system. Gerontechnology; 8(2);119

Scott, V. Peck, S. Kendall, P. (January 2004). Prevention of Falls and Injuries Among the Elderly: A special report from the Office of the Provincial Health Officer. Victoria, B.C.: Ministry of Health Planning.

Sherrington, C. Tiedemann, A. Fairhall, N. Close, J. Lord, S. (2011) Exercise to prevent falls in older adults: an updated meta‐analysis and best practice recommendations. New South Wales Public Health Bulletin, vol. 22, no. 3-4, pp. 78–83.

Slater, M. Usoh, M. (1994). Body Centered Interaction in Immersive Virtual Environments. Magnenat Thalmann and Daniel Thalmann, Artificial Life and Virtual reality, John Wiley,

Stead, M. Wimbush, E. Eadie, D. Teer, P. (1997). A qualitative study of older people’s perceptions of ageing and exercise: The implications for health promotion. Health Education Journal, 56, pp. 3–16.

Westra, J. van Hasselt, H. Dignum, V. Dignum, F. (2008). On-line Adapting Games using Agent Organizations. 2008 IEEE Symposium on Computational Intelligence and Games (CIG'08)

1