A Four-Stage AI Framework for Gait Disturbance Detection and Remote Alerting in Parkinson's Disease

Department of Computer Science and Engineering KPR Institute of Engineering and Technology Coimbator, India
Department of Artificial Intelligent and Machine learning Engineering KPR Institute of Engineering and Technology Coimbator, India
Department of Electrical and Electronic Engineering KPR Institute of Engineering and Technology Coimbator, India
Department of Biomedical Engineering KPR Institute of Engineering and Technology Coimbator, India
Department of Mechatronics Engineering KPR Institute of Engineering and Technology Coimbator, India
Department of Computer Science and Engineering KPR Institute of Engineering and Technology Coimbator, India

This email address is being protected from spambots. You need JavaScript enabled to view it.
This email address is being protected from spambots. You need JavaScript enabled to view it.
This email address is being protected from spambots. You need JavaScript enabled to view it.
This email address is being protected from spambots. You need JavaScript enabled to view it.
This email address is being protected from spambots. You need JavaScript enabled to view it.
This email address is being protected from spambots. You need JavaScript enabled to view it.

Abstract

Parkinson's disease functions as a progressive neurological disorder which destroys a person's ability to walk and maintain balance and move about. The conventional walker device supports users with basic mechanical assistance while it fails to identify or forecast walking pattern disturbances. The research develops an AI-basedsmart walker system which operates through a four-part framework which combines vision- based sensing and pose-driven feature extraction and analytical gait processing and multimodal feedback. A smartphone- mounted camera captures lower-limb motion and streams video wirelessly to a laptop-based analytical engine. The system uses pose estimation technology to obtain anatomically normalized gait features which allow for identification of freezing of gait and .slow gait and fall initiation. The system shows three types of responses which include rhythmic auditory cues and emergency alerts and monitoring of remote caregivers. The experimental tests show the system achieves realtime operation between 20 to 25 frames per second while maintaining accurate detection performance. The systemprovides an affordable and non-invasive and scalable support system which matches medical practices that use AI technologies.