Research And Design Of Hand Rehabilitation Evaluation System Based On Multi-data Fusion

Stroke-induced hand dysfunction has a significant impact on patients’ daily lives.Traditional hand rehabilitation treatment plans face issues such as high equipment costs and insufficient medical resources.To assist doctors in conducting hand rehabilitation assessments and developing appropriate rehabilitation training plans,allowing patients to receive long-term hand rehabilitation assessment guidance at home,this paper designs a hand rehabilitation assessment system based on multi-data fusion.The system includes a glove specifically designed for hand function rehabilitation assessment in stroke patients.This glove can collect angle signals,gyroscope signals,acceleration signals,and pressure signals from the patient’s hand.The unscented Kalman filter with system parameter adaptive estimation is used to process angle data for posture calculation,and the joint range of motion of the hand is computed;the SG(Savitzky Golay Filter)smoothing filter is used to process pressure data to compute finger muscle strength information.Additionally,virtual reality technology is employed to guide patients through the assessment process more easily.A cloud-based client system is established to upload assessment results to the cloud,allowing patients and rehabilitation physicians to view and analyze the results at any time.(1)To achieve low-cost and high-precision hand information acquisition,this paper designs a data glove based on JY901 S and pressure sensors.According to research on stroke patients and rehabilitation assessment requirements,a hardware design scheme for the rehabilitation assessment system is proposed,which realizes the acquisition of hand joint range of motion,hand tremor data,finger muscle strength,and hand flexibility indicators.(2)To accurately restore hand information,this paper processes the posture angle data collected by the data glove using the unscented Kalman filter with system initial parameter adaptive estimation,and performs posture calculation on the processed data.Experimental results show that the accuracy of posture calculation after data processing is much higher than that of unprocessed data.(3)To ensure the orderly operation of the hand rehabilitation assessment system,this paper carries out multiple designs in software: the wristband controller is based on the Free RTOS(Free Real time operating system)real-time operating system,constructing multiple tasks to achieve system functions,and decoupling the embedded system as much as possible;the virtual scene is built on the mature Unity platform for VR(Virtual Reality)development,using Leap Motion for virtual hand generation,and displaying the virtual scene with Meta Quest VR headset,realizing the guidance function for hand rehabilitation assessment;assessment data is uploaded to the cloud for easy access by rehabilitation physicians and patients.(4)To verify the feasibility of the system designed in this paper,the following experiments are conducted: performance analysis and experiments on the unscented Kalman filter with system parameter adaptive estimation and SG filtering;testing of the VR virtual scene and virtual hand model restored using Leap Motion;inspection and testing of the designed hardware system and corresponding sensors to ensure they meet system operation requirements;multi-person experiments on the assessment algorithm designed in this paper.The final results show that the hand rehabilitation assessment system designed in this paper can accurately assess hand joint range of motion,hand tremor degree,finger muscle strength,and hand flexibility.