|Creation of a Virtual Reality Game for Training in a Virtual School Environment
Timothy T. Smith, Computer Science, DA&S
Janice T. Searleman - Advisor
There are many groups that could stand to benefit from an easy to access wheelchair training simulator. Although this affects a large demographic, the primary target audience of this work is children with Cerebral Palsy (CP). Virtual reality provides the opportunity for the creation of an environment that is entertaining while simultaneously training a patient in an every day task. One example of this would be learning the techniques and control necessary for successful use of a power wheelchair (PW/C).
Cerebral Palsy is a congenital condition in which there is damage to the part of the brain that controls muscle movement. There is not currently a cure for this condition, however with treatment a patient can improve. Common effects of CP are lack of coordination in voluntary movement, stiff or tight muscles, exaggerated reflexes, one foot dragging, “scissored” gait, or a crouched gait . CP has four major types; the first is Spastic, which means the patient has a stiffness/tightness of muscles. The second type of CP is Athetoid CP, which is when the person experiences uncontrolled movements that are often erratic. Ataxic CP leads to a lack of balance and coordination, however it is the least common form of CP. Finally, the fourth type of CP is some combination of the previous three forms, known as Mixed Type CP. In some cases mobility is still possible, however in others a wheelchair is the best solution . When applicable, it is important that a patient be able to obtain a Power Wheelchair (P/WC), which helps the patient achieve the maximum possible level of mobility.
One problem with P/WC’s is that they are very expensive. It is unrealistic to expect the patient to cover the cost alone, so they typically require assistance from insurance companies. Although this seems like a straightforward solution, many insurance companies are not willing to cover the costs for a P/WC until the client can demonstrate that they know how to use one. Since the patient does not own a PW/C to practice on, or to demonstrate competency with, this poses a problem . There are facilities in existence that allow a patient to train in the use of a P/WC for this reason.
These facilities are not ideal in that they need sufficient staff to accommodate people seeking to get training for a P/WC. Also, taking the time to go to these facilities can be tasking on the patients themselves, and naturally limits the amount of time they can devote to training . Additionally, another problem is cause by the fact that the P/WC that the patients are training on has to be customized for the patient each time, which can result in a waste of even more time and effort.
A solution to these problems would be to move a portion of the training from the physical world to a virtual one. Virtual reality technology is capable of creating a realistic, 3D training environment that can help patients polish their P/WC skills. This is ideal because it provides a safe and friendly environment in which to learn P/WC operation . Additionally, it would require less hands-on work from physical therapists and could be performed from the comforts of the patient’s own wheelchair. This is important because it gives the physical therapists the ability to work with more patients simultaneously, without sacrificing effectiveness. Instead of constant attention from a physical therapist, the program could handle most mundane details, leaving the physical therapist with a more specialized role.
Advantages of Virtual Reality
First, virtual reality has the advantage of added safety. This is especially important to beginning users who are not familiar with P/WC’s and so have a high risk of collision. In these situations, there is the potential for a collision that risks damage to the P/WC, or even worse, injury to the patient or others. Virtual collisions however would have no adverse physical effects. In short, virtual training provides a safe world where mistakes can be made without fear of dangerous repercussions .
It is also possible that if a patient were to be placed in a virtual world, then they would need less individual attention. An electronic “trainer” could be incorporated into the program, with the goal of providing an automated learning aid for the patient. This program could do things such as: assigning tasks, providing useful tips for correcting mistakes, and collecting data about the patient’s performance. These tips could be geared toward the actions of the player, and could be used in situations where the patient was repeatedly hitting a wall moving backward. The program could detect this and then alert the patient to progress more slowly backward and to check behind himself or herself more . Additional prompts could also be used, in which when moving backward for a certain amount of time, the patient could be reminded to check behind them. This data can later be reviewed by a physical therapist to assess the patient’s progress, and so isolated areas that the patient needs work.
Virtual reality also possesses a portability that the current system lacks. If the program is correctly written then the patient will be able to access it from their home anytime they want via the Internet. This is part of the reason why Virtools was selected. The Virtools software package posses a web applet feature that would allow for easy hosting, and access to any program created. This would allow for a much larger potential practice time, requiring that the patient physically enter the facility only occasionally, and to take a final assessment for the insurance companies. This saves time and effort on travel, and does not require that the P/WC’s at the facility be reconfigured for the individual patient as much. Instead they can practice from the home, from the comfort of their own, standard wheelchair.
There are some minor problems with possibly home use, however the fact that it can be used from the home at all is still a significant advantage. When using the system at home, there would be some loss of immersion, which would detract from the sense of involvement in the virtual world. Because of this, the home version would not provide as convincing an environment. Also some of the customizations by the physical therapist would be harder to implement via the online program, however they would not be impossible to implement. The current design for the game system is to take all of those customizations before the game program is even launched, and using that information create a data file. The program would then use this data file to customize itself, so with this structure online customization would be possible. A final problem is the patient would need to buy some hardware to utilize the joystick training, either his or her own joystick or some acceptable alternative such as a sip and puff system
A sip and puff system allows a user who lacks the manual dexterity to use a joystick an alternate means of controlling the wheelchair by blowing or sipping into tubes. These are ideal for creating things such as alternative wheelchair or computer interfaces and can simulate things such as a mouse with the addition of head tracking .
Previous work has been done on the use of virtual reality as a training tool for physical rehabilitation. Two specific examples are Kyle Burdick’s Virtual Reality House , and Chad Homan’s Virtual tutor . Both of these serve as training tools for use in a P/WC, utilizing a virtual environment.
Kyle Burdick’s house provides a home-oriented environment where the patient can maneuver around a two-story house and interact with various objects in the rooms. It is equipped with an access ramp, and an elevator to simulate practice on both devices. In order to make the 3D world more entertaining, a virtual reality scavenger hunt game was created in which the player is give the task of locating 6 objects based off of clues . Data that is gathered from this work is the number of collisions occurring during the game, and the total score . This original concept was later expanded into a more comprehensive game, where varying clue difficulties were added, along with a user interface that allowed for the display of objects missing and objects already found. Additionally an introduction and sound track were added .
There are a few problems with Kyle’s house, and the subsequent scavenger hunt game that if addressed, would make a stronger product. Firstly, Kyle’s house doesn’t follow physics constraints, which are important to provide a realistic and immerse experience. Also, more accurate and useful data collection would be important to accurately analyze the performance of the patient and present a solid case to the insurance companies. Furthermore the color scheme is not accurate to a real house, giving the distinct impression of the virtual world. Higher levels of immersion will create a more believable simulation, and in doing so would provide for a more intense and enjoyable environment.
Chad Homan’s virtual tutor is a series of different training rooms, each designed to teach a specific skill to a user of a virtual P/WC. Each of these rooms provides a solid base for the user, and also actively attempts to assess their progress and provide useful suggestions to improve performance. Additionally a notice is given whenever a wall is touched or an object is hit, meaning that the user knows when they have made a mistake . This type of instruction is most useful in providing a strong foundation of training for a basic user, however as far as playability goes it is not a lot of fun.
At the same time that Kyle was creating the house; Ajay Sonar created a virtual version of a wheelchair that is physically accurate, enabling realistic movement and momentum constraints . These wheelchairs were employed in both Kyle and Chad’s projects.
Another project is Cassandra Dearborn’s design of a virtual school. She utilized an engineering design process to create a robust 3d virtual model of a typical middle or high school. Her model provides a framework that can be expanded by others, and imported into a virtual world. It contains 26 rooms, and 2 stairwells for potential expansion to a second floor. There are also standard school rooms such as a gym, cafeteria, nurses office, and so on.
Another group of students, working with Cassandra’s model, imported the model of the school into Virtools so that it could be utilized in a virtual environment. Additionally, more objects were created and placed to enhance the virtual experience .
This virtual school was further expanded to allow for an intelligent guide system, represented by an arrow that directs the user to a specific location. Frank Ashcraft is still developing this system. Additionally, Bill Bushy is created a virtual planning agent inside the virtual school, using artificial intelligence techniques to create more realistic interactive characters, utilizing a virtual planning agent, by Bill Bushy .
To ensure replay value, and to help motivate patients, the training simulation must have some value as an entertainment device. If the tutorial is disguised as a serious game, and were enjoyable to play, then it could be expected that the patient would seek to play it even when they were not required to. Since the game could have free online access, and would still possess most of the key features, it could prove valuable in helping players become acquainted with the manner of movement as well as the nuances of a P/WC.
The purpose of this thesis is to create such a virtual reality game, for use with the existing virtual school. The design is for a serious game, meaning that in addition to being enjoyable to play, it is also designed to teach the player useful important skills, in this case about the proper operation of a P/WC. The key to the creation of a serious game is that there has to be a blend of playability and functionality. Too much information and not enough game play means that the game will become boring and have little replay value. Whereas too little information means that the game will not fulfill its original purpose of helping the user learn the operation of a P/WC. So the key to creating a serious game is balancing game design with accuracy of information, which is no easy task .
Games provide a useful medium for instructional products that is, as of yet, untapped. There are theories in existence that the chemical reactions brought on by game play can leave the player in a state that actually heightens their ability to learn . Additionally, by crafting a fun game that contains the techniques that we hope the patient will master, we should be able to create a method of training that will not only be enjoyable, but that the patient will actually seek to play. Since they will want to play the game, and so may spend part of their own time practicing, it would be possible to expect faster improvement.
It has been shown that a virtual training environment is in fact quite feasible. An example of a particularly successful one would be DARSWARS. This is a military training simulator in which two squads undergo ambush training, one squad escorting a caravan and the other ambushing them. One of the major reasons that this project was so successful was the fact that it enabled easy modification to the game design, allowing for customizable scenarios for each individual squad. This is important since it creates an atmosphere which can adapt to a changing demographic easily .
To move onto this stage, we need a robust, fully functional model of the virtual school, which enables realistic movement through the system. My first goal is to modify the current virtual school to satisfy this. Once this has been tested to ensure that the wheelchair can successfully navigate the school, while retaining realistic and precise physics, and then the game can be created.
The game needs to be carefully designed in order to ensure that all aspects are both fun and educational. To create a compelling game, there needs to be a source of conflict so that the user feels challenged, and eventually accomplished. This motivation can be reinforced by two major techniques, attaboy’s for tasks well done, and demerits for failure .
The plan is give the patients something to see from these rewards, such as unlocking games that they can just have fun with. Current ideas include things such as a food fight, a variation on rock, paper, scissors, and other such small games. The idea is that the first time a task is completed it will unlock the next task, but it can be repeated to allow for multiple play through. Every successful completion of a task will unlock a token, which can be redeemed for a round at a fun game.
Additionally, ideas for demerits have been considered, and in addition to a simple restarting of the incomplete task. There are three major classifications of punishment, which apply to our game, loss of progress, loss of powers and loss of time . Loss of power seems to be the best option to pursue. For instance when the patient is experiencing too many collisions, the wheelchair could be slowed down. The user would then have an easier time controlling the wheelchair, but it would still hinder the user’s performance. Good performance afterwards could then restore the speed.
We also want this game to be customizable, to reflect a patients specific needs, so the game utilized a mini-game structure. The physical therapist selects which mini games are open in the session before it starts, and so can best emphasize the points that the patient needs work on. This allows for customizability of the game and keeps a good amount of control in the hands of the physical therapist. This also enables the story to be vague and easier to understand.
One aspect of the training will be the simple mobility thought the school. Simple things such as navigating the halls, pulling up to and away from desks are the kind of training that is needed to be able to justify a P/WC, so these actions will be an integral part of the game. Currently, the game centers on class activities, and after each class the player is required to navigate to the next class. The games themselves should emphasize cognitive skills as well because this will allow the game to server a double purpose in terms of wheelchair training and cognitive development. The goal of this project is to create two functioning training games, as well as two reward games for the player.
Since the target audience is children with CP, the varying nature of the disorder needs to be taken into account. Some of the patients may have a very difficult time understanding complex plots or storylines, and in this case those elements will detract from the game play. However, to other patients this type of immersion might be just what they need to have an enjoyable experience and to continue playing the game. So a balance needs to be established in order too keep the game interesting for all. So another task is to detail exactly how much plot and challenge patients can be expected to take, so that a full game design can be developed. This will be best accomplished by creating varying difficulty settings that loads different prompts for different people. Ideally this setting will be configured by the physical therapist.
This ensures that each patient has as close to a tailored experience as is possible. This will help keep the game interesting and engaging. Players are more likely to continue against an adequate challenge, but will grow bored if it is either too hard too easy. And this will also prevent players from becoming lost in plots they cannot follow.
However to be useful this game also needs to accumulate data on the performance of the patients. This data can then be reviewed by the physical therapist in place of active note taking, and used to determine what aspects the patient still needs work on. Also, this can help assess where problems are and lead to more constructive feedback from both the physical therapist and the program.
This project necessitates the building of such a data collection structure, inside the game. Information that will be collected will be number of collisions, direction of motion when the collision occurred, speed at the time of collision, and possibly more if it is later thought of. This information will be saved to a file that the program will create, separated by patient data and padded with line spacing.
To create a realistic 3D world, it is going to need a working physics simulation. This will be achieved utilizing the Virtools Dev Physics pack, which is software developed to run physics rules in the Virtools virtual environment.
The ideal instructional game is one in which the user desires to play, even on their own time. To this effect this project hopes to create engaging mini games in which the patient is presented with a challenge or task that they have to overcome. It is hoped that basic tasks can be both beneficial for, and enjoyable to the patient.
Current ideas for mini games include, hallway hazards, dropped notebooks, missing objects etc. When a hallway hazard mini-game is active, unlike other mini games, it modifies the actual game play. This option provides dangers and challenges in the navigation of the hallways. Specifically these ideas include stationery people or objects who narrow the width of the hallway and make it harder to navigate, wet floors that change the traction of the wheelchair, any mobile obstacles such as doors that open when a player gets a set distance in front of them. Another stationary obstacle idea was to have some of the lockers remain open. This would give the patient either the option of simply moving past them, or of driving up and closing them to widen the hallway.
These hazards will provide the opportunity to assess performance under abnormal conditions and can help assess response time of the patient. With this type of data it can be determined exactly how far away a patient needs to be able to react and safely stop before an unexpected event (the doors). It can also provide information about their judgment of the size of the P/WC and its ability to navigate tight or enclosed spaces. Finally, the water hazards provide the opportunity to gain experience in conditions that may not be the P/WC’s ideal operating environment.
A second task idea is the retrieval of a dropped notebook. The patient begins in front of their desk, and is asked to navigate out of it and to their next class. Directions to this class will be given. When they arrive and pull-up to their desk, they will be notified that their notebook is missing! They mush have left it in their last class, and then they are instructed to retrieve it. This time, directions and guides are not given. They must return to their classroom and retrieve their notebook, then go back to their new class, again without directions.
Another game is a more of a “racing” type game. Late for class involves the patient receiving a notice that the first bell has rung, and that they have to hurry to make it to their next class. Guidance to the room will be provided, however they have to navigate to the next room in a time frame of less than 2 minutes. Different difficulty settings could adjust this time. The task is completed when the patient pulls into their desk in the new room.
An additional task would be to perform a science experiment, where they receive a demonstration of the task for a finite time, and then must repeat it. This could use particle effects to create different variations on the experiment and would provide an interesting alternative to most of the other tasks. The P/WC practice in this task involves the patient leaving the previous class, and driving up to the table.
Another game idea would be to put the patient into a situation that involves them practicing backing into bathroom stalls. In this game, the patient is told that they need to go to the bathroom and back into a stall or pull in and turn around, and close the door. They then must open the door and go back to class. This is to ensure they have proper training navigating tight spaces such as bathroom stalls.
The next idea centers on arriving at school where the student needs to enter the building and arrive at their first class. This is just a basic introductory part, and can be used to provide the basic story line and tasks to the student.
Similarly an end of the day scenario could involve the student walking to their locker, opening it up and getting their backpack. They would then have to navigate to the bus out front in order to head home.
A searching game could be implemented in the library where the participant would need to go and find a book in the library. It would begin in a classroom, requiring them to go to the library. Once there they would navigate the shelves and find the desired book. The book would then need to be taken to the counter to be checked out, and the student would need to return to their class.
A final idea was a missing item around the school. The patient would be requested to help locate an item that their best friend had lost. They would then have to search the school looking for it, and would be given clues about it from the friend such as “I remember having it at lunch.” If the patient cannot find the ring in the time limit, to keep the game moving the friend will approach them, thank them for their help and tell them they have found the ring themselves.
Although these ideas are not fully developed each one helps the patient work on their skills in the P/WC and provides a challenge to be overcome. They need a little more development in terms of being effective games, however they provide a strong baseline. Once these are perfected the resulting game should have solid replay value and provide a unique experience each time. The finalized game that I’m creating will involve two such games, created specifically to aid the patient, and two games that are rooted in fun but still have some training value. These fun mini games will be rewarded with satisfactory or exceptional performance on the training games, and can be stored up and redeemed by the patient any time.
The overall design will focus on a user-centered design process, which will enable an early incorporation of the user into the development process [13, 14]. Additionally, this will ensure that the final product is specifically geared to solving the problems of the user and will thereby increase its effectiveness. The major tenants of the user-centered design processes are an early focus on users and tasks, empirical measurement, and iterative design. This means that throughout the whole process user needs will be sought after, and the product will be tested to make sure those needs are satisfied. If the products solution to those needs is found lacking, then a reassessment will occur, with changes to fix the problem. From there the cycle will repeat.
September. – Completion of the physicalized school, and moveable wheelchair within that environment.
Optimize for speed and rendering time. Plan structure and detail for the mini-games
Including story ideas, and ideas for rewards and demerits
October: - Data collection system established and working, as well as front-end
user interface to enable physical therapist input.
Mid-semester break - one therapy game functioning as well as a functioning reward system and
basic storyline. Implementation of one reward game as well.
January – Testing and assessment of functioning games
End of senior year – completion of assessment of all four games and final paper, as well as a final version of all four games
National Institute of Neurological Disorders and Stroke. "NINDS Cerebral Palsy Information Page." National Institute of Health. 6/24/08. <http://www.ninds.nih.gov/disorders/cerebral_palsy/cerebral_palsy.htm#What_is>
Better Health Channel. 6 September 2008. Cerebral Palsy – causes and implications. Scope. 9/7/08 <http://www.betterhealth.vic.gov.au/bhcv2/bhcarticles.nsf/pages/Cerebral_palsy_causes_and_implications>
Homan, Chad. Design of an Intelligent Tutoring System for Teaching Power Wheelchair Operation in a Virtual Environment. Clarkson University, 2005.
ATWiki. March-3-2008. assistivetech.net. 7/7/2008 [http://atwiki.assistivetech.net/index.php/Alternative_wheelchair_control].
Burdick, D. Kyle. Virtual Environment Design for a Virtual Reality Wheelchair Simulator. Clarkson University, 2005.
Savage, Samuel, and Timothy Smith. Susan E. Conry and Janice T. Searleman advisors. “Improvements to a Scavenger Hunt Game for Virtual Reality Wheelchair Training.” Clarkson University SURE: Symposium on Undergraduate Research Experiences, July 2005
Sonar, A. V., Burdick, K. D., Begin, R. R., Resch, E. M., Thompson, E. M., Thacher, E., Searleman, J., Fulk, G., Carroll, J. J. “Development of a Virtual Reality-Based Power Wheel Chair Simulator,” Proceedings of the IEEE International Conference on Mechatronics and Automation, August 2005.
Edmands, Max, Alana Searleman and Katelynn Wilton. Janice T. Searleman, advisor. “Design of a Virtual School Environment for Power Wheelchair Training.” Clarkson University SURE: Symposium on Undergraduate Research Experiences, August 2, 2007.
Bushey, William. Janice T. Searleman, advisor. “Design of an Artificially Intelligent Planning Agent for Virtual Wheelchair Simulator using Hierarchical Task Networks.” Clarkson University SURE: Symposium on Undergraduate Research Experiences, April 2008.
Kelly, H; K Howell; E Glinert; L Holding; C Swain; A Burrowbridge; M Roper "How to Build Serious Games" Communications of the ACM 50.7 (2007): 45-49.
Chatham, Ralph. "Games for Training" Communications of the ACM 50.7 (2007): 37-43.
Dille, Flint, and John Zuur Platten. The Ultimate Guide to Video Game Writing and Design. New York: Lone Eagle Publishing Company, 2007.
Sharp, Rogers, and Preece. Interaction Design beyond human-computer interaction. England: John Wiley & Sons, Ltd, 2007.
Isbister, Katherine and Noah Schaffer. Game Usability. Burlington, MA. Morgan Kaufmann Publishers, 2008.