Skip to content
Conference

Adaptive Sensor Fusion and Machine Learning Framework for Movement Detection in Rehabilitation System

Aug 2026 · International Workshop on Artificial Intelligence and Cognition · pp. 1494-1500 · 0 citations · 15 references

Abstract

With proper movement detection, the rehabilitation process becomes more successful as evaluation of the progression of the patient and corrective feedback is necessary. This paper presents a machine learning and adaptive sensor fusion algorithm that has been used to detect and classify movement patterns during rehabilitation. The system combines multi-sensor information of wearable inertial sensors with emphasis on the properties of the joint kinematics, gait, and energy-related characteristics to group movement performance into three categories, namely, Optimal, Slight Deviation and Deviation. Complex movement data is analyzed using different Support Vector Machines, Decision Trees, Random Forests and Gradient Boosting models. The findings have indicated that the system is extremely highly rated with the largest scores of XGBoost (99) followed by the random forest (98) and SVM (97) and Decision Tree (96) machine learning algorithm. Findings indicate that the system was very high with the highest score of XGBoost (99) and the second and third most with the highest score of the fourth being the SVM (97) and the last being the Decision Tree (96). The feature selection techniques are used to diminish the dimensions so as to be efficient and to have better performance of the models. The proposed system will provide an effective solution to the customized rehabilitation monitoring by offering the real-time feedback and scalability of the solutions to meet the needs of different patients.

View source

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.