Research On Human Activity Detection And Assistive Robot Technology

With the coming of aging society,people pay more and more attention to the health and life problems of the elderly.Developing a safe and reliable elderly assistance system becomes a very urgent social need,which would improve the life quality of the elderly and alleviate the pressure of medical and nursing resources.The elderly assistance system mainly achieves two functions: activity detection and behavior assistance.It can not only detect the daily activities of the elderly,but also assist them to complete Sit-To-Stand(STS)and walking activities which occur frequently in their daily life.After full investigation,this thesis proposes the schemes of sensor-based activity detection and rehabilitation-assistance-robot-based behavior assistance.Hence,we have designed a wireless wearable sensor system and an assistive robot.This thesis respectively introduces the wireless wearable sensor system and assistive robot from the aspects of overall structure,critical module design,and software system.And it helps readers to deeply understand the design process,program flow and usage method of the wireless wearable sensor system and assistive robot.The wireless wearable sensor system can simultaneously collect the motion information of human body and the muscle strength change of the wearing part.Using one sensor node of the wireless wearable sensor system,this thesis collects the motion and muscle information of thigh,and realizes the accurate recognition of 11 human daily activities.At the same time,three different classification experiments are carried out to illustrate the improvement to the activity detection.The assistive robot has two functions: STS assistance and walking assistance.The related kinematics models of STS assistance and walking assistance are built based on the overall structure and the human-machine interaction module of the assistive robot.In order to achieve the compliance control of robot assistance,this thesis takes the human intention into the control of assistive robot,which presents a proxy-based STS assistance algorithm and reproduces a walking assistance algorithm based on admittance control.Based on these two assistance algorithms,the corresponding experiments of STS assistance and walking assistance are completed.The experimental results verify the assistance effect of the assistive robot,and basically explain the security,stability,and compliance of the motion assistance algorithms.The proposed elderly assistance system has made performance improvement or function perfection in the aspect of activity detection and behavior assistance.It is simple to use and low cost,which lays a foundation for the early practicality of the elderly assistance system.