This study has successfully created a virtual reality training environment for wheelchair users aimed to serve children with severe cerebral palsy and cognitive impairments. The proposed system integrates motion tracking with virtual reality technologies in real time for rehabilitation purposes. The system aims at improving the trainee motor functions allowing the caregivers to directly interact with the child using visual cues. The live motion trajectory recorded from caregivers could provide a strong interaction with patients to enhance their engagement, comfort, and enjoyment of training. The proposed system can potentially help improve the lives of dozens of children who suffer from the effects of cerebral palsy in a safe, cost-effective, and entertaining enviorenment.

Cerebral palsy (CP) is one of most common diagnosed childhood motor function disabilities in United States. It affects a person’s movement and coordination of muscles to maintain balance. Though many treatments do exist on the market, they can be expensive and time-consuming activities which are often beyond the means of many families There is often the problem of limited space available for practicing the control of assistive devices in a safe manner. Virtual reality (VR) technology has gained attentions as an intervention to improve motor function for the children
with CP. It can provide end-users the feel that they are engaged with their surroundings with far more realistic conditions than a training program which involves only screen-based action. Meanwhile, the level of task difficulty can be easily modified to effectively improve their motor function, based on the real-time feedback from task performance. In 2022, the report from Birth Injury Help Center indicated that a large percentage of children with CP have some level of hearing impairment. It directly leads to their speech and language problems, which greatly results in their difficulties to follow instructions and may result in training ineffectiveness. Accordingly, motion tracking technology can be used to incorporate visual cues from the therapist within VR sessions when interacting with individuals affected by CP. Unity 3D software was used to create the VR environment in this study. Considering that the users will be children with CP, the training “game” was produced to simulate a moving wheelchair in an urban environment. The different therapy training levels are provided by changing the driving speed of the wheelchair and the number of obstacles in the VR environment. The individual can control the wheelchair by only steering left or
right. Additionally, Xsens MVN system was used to capture the full-body movement of one volunteer, which was selected as the caregiver’s action for patients to follow. Given the unique compatible feature of Xsens and Unity 3D, the live motion of the caregiver recorded by Xsens was successfully projected into the VR environment to effectively provide real-time instruction/command and obtain a reaction from children with CP. Our study has successfully created a VR motor function training environment. This is the first time in which motion tracking technology is used within VR for the training purpose of children with CP. This technique solves the limited space issue for mobility
training avoiding the risk for the trainee of getting physically injured in the realistic circumstances, by not being able to control a powered wheelchair. It also adds a layer of connection between the caregiver (therapist or parent) to directly interact the trainee within the VR environment. This is specifically important for children who have hearing impairment or cognitive deficit. The caregiver can directly interact with the child indicating the direction to go and providing a familiar human interaction. The combination of VR and motion tracking technologies makes for a comfortable training environment with the interaction between patients and caregivers that can greatly improve the confidence of the child.

Keywords
Cerebral palsy, Virtual Reality, Rehabilitation, Unity 3D and Xsens MVN