A group of students from Ho Chi Minh City University of Technology has applied AI and virtual reality to assist stroke patients in rehabilitation through games like rowing and cycling.
The project was developed by Nguyễn Anh Quốc, Nguyễn Võ Tâm Toàn, and Trần Thế Lực from the Mechanical Engineering Department, starting in June 2020. After more than a year, the team created a rehabilitation training system consisting of three main components: training equipment, a virtual reality screen, a monitoring camera, and data management software.
The arm and leg training equipment is equipped with sensors to measure speed and torque, providing metrics on the user’s performance. The screen displays a virtual reality interface for an engaging gaming experience. The monitoring camera tracks the user’s movements, detects errors, and issues alerts for corrections.
“The training device utilizes motivational technology through games, showcasing their recovery through metrics that help users stay encouraged,” Quốc, the team leader, explained the motivation behind the product. He also cited statistics from health authorities indicating that over 80% of stroke survivors experience physical impairment. Therefore, the demand for rehabilitation among these individuals is significant. Currently, rehabilitation training follows traditional methods using simple equipment and requires medical supervision. The training process is often challenging, and many individuals recover slowly due to improper methods, which can lead to feelings of frustration.
The virtual reality screen simulates a rowing game with notifications for users to adjust their movements. (Image: Provided by the authors)
Based on consultations with rehabilitation physicians and scientific publications, the team developed a training regimen tailored to each patient’s physical condition.
The team also created a software interface where users input their name, age, gender, and health metrics such as FIM (Functional Independence Measure), arm strength, and leg strength, allowing the system to generate a personalized training plan regarding intensity and duration.
The exercises are presented in the form of rowing and cycling games. During the training, if the user does not meet the required standards, the system automatically sends notifications for adjustments on the screen. For instance, in the rowing game, if the rowing speed is slow, the system prompts the user to increase their pace. If the arm angle is below standard, the system alerts the user to raise their elbow.
According to Quốc, the training device is equipped with motors to assist patients with weak arms and legs, helping them complete their exercises. After each session, a display shows the completion level of the exercises, allowing the system to assess the patient’s recovery progress through each activity. All data collected during the training sessions are saved for users and doctors to monitor and evaluate the patient’s daily recovery.
Users experiencing the use of the machine. (Image: Provided by the authors)
By mid-2021, the product was completed, but testing was delayed due to the Covid-19 pandemic. The team has now contacted a healthcare facility, which has agreed to test the product on patients.
According to Associate Professor Dr. Nguyễn Trường Thịnh, head of the Mechanical Engineering Department at Ho Chi Minh City University of Technology, the product’s advantage lies in combining a physical therapy training system with an AI and augmented reality guidance system. The team’s system is fully automated, handling guidance, assessments, and data storage throughout the training process. This helps patients feel less pressured while training with the device through engaging physical games integrated into a virtual environment.
“Such medical devices must be evaluated or tested with patients and approved by an ethics committee. In the near future, the university will support the team in testing on patients and providing assessments based on scientific data,” said Associate Professor Thịnh.
The estimated production cost for one system is around 3 to 4 million VND, not including the cost of the screen. The research team mentioned that most households already own televisions or laptops, which can be utilized. Compared to conventional training devices on the market, “this cost is not significantly different, and most people can access it,” a representative of the team stated.
