Mobicrus – Classification of Lower Limb Actions Based on EMG Signals Using Machine Learning

One of the most crucial types of care is rehabilitation, which helps patients regain or increase their capacity for daily living. Particularly, there has been a significant rise in the usage of robots in rehabilitation and medical applications in recent years. Numerous studies are being done in this area for a variety of reasons, including the upper and lower limbs of the human body. According to studies, robots can greatly assist physiotherapists in this regard. The physical or physiological treatment of the body known as “physiotherapy” tries to enhance muscular function and stimulation. By stimulating the neuromuscular and muscular impulses, a therapist works to restore the wounded limb's capacity for normal movement. One of the drawbacks of this approach is the challenge it presents to doctors in doing a repetitive error-free action without growing bored or lowering the standard of care provided to the patient. The restoration of patients' motor abilities with muscular weakening is made possible with the help of a practical framework provided by rehabilitation expertise. Establishing a framework for handling and assessing lower limb electromyography (EMG) signals that are ready to be sent to a rehabilitation robot is the main goal of this effort. Data cleaning steps are undertaken on data collected from various sensors. Also various Machine Learning techniques are used to build the system. During the model selection and parameter tuning techniques, various algorithms like random forest, decision tree, support vector machine for EMG signal will be used and the best performing model will be selected. Finally the trained ML model is used to classify the lower limb actions based on the sensor data. With the help of this work's findings, prosthetic limbs or rehabilitation robots can be designed and programmed to perform as closely as possible to real human limbs.