Smart Rehabilitation System for Motion-Impaired Patients
Parkinson?s disease is a neurodegenerative disorder that affects the part of human brain that controls body movements. More than 10 million people are affected by motion diseases around the globe which is alarming. The proposed system targets the diagnosis of Parkinson?s disease through machine
2025-06-28 16:35:40 - Adil Khan
Smart Rehabilitation System for Motion-Impaired Patients
Project Area of Specialization Computer ScienceProject SummaryParkinson’s disease is a neurodegenerative disorder that affects the part of human brain that controls body movements. More than 10 million people are affected by motion diseases around the globe which is alarming. The proposed system targets the diagnosis of Parkinson’s disease through machine learning techniques and facilitates the doctors in the rehabilitation process by utilizing Inertial Movement Unit (IMU) Sensor-based suit and leap motion controller. The system will predict the disease stage of the patient based on perceived movements and historical records through deep learning algorithm. It will provide multiple goal-oriented tasks based on physiotherapy interventions and criteria for the recovery process. The smart system will enable doctors to design personalized therapy for the patient. Moreover, the system will predict the best suitable tasks for the patients. Medical history and reports of assigned tasks will be stored in real-time database and rehabilitation progress will be analyzed to evaluate the patient’s recovery.
Project ObjectivesThe main objectives of the project are as follows:
- Design and implement an intelligent system to predict the best goal-oriented task to improve the rehabilitation process of Parkinson’s disease patients.
- Automate the analysis of major factors in Parkinson’s disease i.e. gait and posture analysis for early diagnosis and assistance in rehabilitation by assessing real-time tracking of each body joint using IMU sensors.
- Assess the motor dysfunction in patients based on the MDS-Unified Parkinson's disease rating scale using leap motion controller to help boost the rehabilitation process.
- Quantify bradykinesia using a leap motion controller for the treatment and monitoring of Parkinson's disease.
- Assist the doctors with continuous performance tracking and measuring the range and agility of each body joint in real-time.
- Identify delicate changes in the patient’s disorder progression, and to evaluate the efficiency of rehabilitation in a consistent way.
- Provide a 3D virtual avatar to replicate the patient's body motion in real-time to give visual feedback.
- Implement a cost-effective and efficient system.
For the implementation of the project, a combination of SCRUM model and agile methodology will be deployed:
SCRUM Model:
Scrum is a lightweight framework used to manage knowledge work using agile methodology. It is mainly used for software development. We are using this approach because it emphasizes teamwork, makes work faster, maximizes responsiveness, makes changing easy and is iterative so makes developing complex systems faster and easier.
Agile Methodology :
To implement these requirements, agile methodology is used as it assists developer teams in constructing software. This uses incremental and iterative work sequences that are called sprints. A sprint is a dedicated time period for each phase in the project. As there can be some differences between team members whether the development is acceptable enough or not. But the other phases of the project will be continued within their time period.
The proposed system is beneficial for both doctors and patients. It will provide dedicated assistance in the rehabilitation of motion impaired patients with help of IMU sensor-based suit and deep learning algorithms integrated into the system.
Games and interactive tasks have proven positive impact on patients’ rehabilitation and therapeutic processes because patients are engaged in interactive activities. Therefore, the proposed system will increase the rehabilitation rate of motion impaired people and motivate the patients toward therapy.
The system will acquire data through sensors which are precise to reduce the margin of error that could be caused by normal human observation and helps to perform tasks accurately.
Deep learning-based Parkinson's disease identification mechanisms will assist doctors to diagnose the disease accurately. The system will help in making future decisions regarding task assignments for rehabilitation purpose by gradually learning from the patient’s data, recovery process and past decisions made by doctors.
Technical Details of Final DeliverableThe proposed system will be developed in Unity 3D which is a cross-platform development tool. The backend logic and scripts will be written using C# programming language. Data will be managed and stored in a Firebase cloud service which will be linked to the system using rest-client API and Firebase SDK. Python will be used to implement machine and deep learning based techniques to make the proposed system intelligent. A wireless router will be used to establish a connection between the system and motion capturing suit. After establishing the connection, the suit will provide raw motion data in numeric format to the system. This data will then be analysed and pre-processed to extract features from it. After feature extraction, data will be split into training and testing datasets using the class balanced split technique. For disease diagnosis, supervised learning will be used as required input and output is known. To use supervised learning the training and testing dataset needs to be completely annotated by domain experts during preprocessing. Since the datasets are numeric so regression predictive modeling will be used to train the machine learning algorithm and build the model. The test data will then be used to evaluate the performance of the model. The model will learn with time and improve its performance. The system will offer 3D animations of various goal-oriented tasks which will be created by acquiring motion data of a healthy person. The patient will have to wear the IMU sensor-based suit and perform those tasks to be a part of the rehabilitation process. Leap motion controller will be used to acquire the hand gestures of the patient. This real-time motion data of the patient will be compared with a healthy person's motion and undergoes the evaluation process. The evaluation process is based on physiotherapy interventions and criteria such as exercises and treadmill training with different outcome measures. These tasks will have real-time 3D visual representation rigged on a humanoid character such as a virtual avatar that will give visual feedback of the patient's movements. The progress made by completing those tasks will provide sufficient information and data to the system, which will then evaluate the patients’ recovery rate and help the model to learn and predict the best suitable task for the rehabilitation process.
Final Deliverable of the Project HW/SW integrated systemCore Industry ITOther Industries Medical , Health Core Technology Wearables and ImplantablesOther Technologies Artificial Intelligence(AI), Internet of Things (IoT)Sustainable Development Goals Good Health and Well-Being for PeopleRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 80000 | |||
| Leap Motion Controller | Equipment | 2 | 15000 | 30000 |
| Programming Boards | Equipment | 2 | 1500 | 3000 |
| Wireless Router | Equipment | 1 | 2000 | 2000 |
| IMU Sensor Kit | Equipment | 1 | 35000 | 35000 |
| Cloud services, printing, stationary, overheads | Miscellaneous | 1 | 10000 | 10000 |