Multi-mode Perception And Interaction Design Of Neurocognitive Gait Rehabilitation Robot

At present,China is entering an aging society,and the incidence of central brain system diseases(stroke,Parkinson’s disease,dementia,etc.)is increasing year by year.The demands for social medical care,rehabilitation and nursing care brought about by them are facing huge challenges,and the development of advanced cognition-motor function rehabilitation robot system has become a research hotspot in the field of rehabilitation engineering and robotics at home and abroad.However,there is still a gap in the "post-rehabilitation period" or "rehabilitation platform period" between rehabilitation robots and robots for the elderly,especially in families or communities.In the medical institution scene,especially the elderly with cognitive impairment and brain nerve injury,there is a lack of cognitive sports rehabilitation robots that can perform "perception-cognition-motor" integrated training.This article focuses on the design of a new type of "perception-cognition-motion" multi-modal intelligent perception system for neurocognitive sports rehabilitation robots,and focuses on constructing a multi-modal intelligent perception interaction method for "perception-cognition-motion" cooperative training feedback.To realize the leap from "functional assistance" to "instructor" for rehabilitation robots,to realize human body posture estimation on the robot,and to explore methods for judging human body stability by using video images.The main research contents are as follows:(1)Research on human body stability,using the Vicon motion capture system to collect detailed human body motion data,analyzing the stability of the human body when performing different movements,using a single pole and foot inverted pendulum model to calculate out of the dynamic stability domain of the center of mass of the human body.A calculation method for judging human instability has been found;(2)For the elderly who need cognitive-motor rehabilitation with cognitive impairment and brain nerve damage,based on the humanoid robot Pepper developed a neurocognitive gait rehabilitation robot,using the robot for multi-modal perception and human-computer interaction,in-depth research on the robot’s voice dialogue system,real-time map construction and precise positioning technology,real-time human tracking and obstacle avoidance functions.It can be adapted to sports rehabilitation and home care for this group of people;(3)The human body posture estimation technology in the field of artificial intelligence is integrated with the robot technology to realize the human posture estimation of the rehabilitation robot,and the human body joint point position detection is carried out through the robot’s vision.So that the robot can quickly acquire the current posture and movement of the target,calculate the dynamic stability domain of the human body according to the human posture information,and explore the feasible solution for judging the stability of the human body through the video image of the human body movement;(4)The above-mentioned research content and neurocognitive gait rehabilitation robot will be clinically applied and tested and certified by a thirdparty authority.They verified that neurocognitive gait rehabilitation robots play a positive role in cognitive-motor cooperative training.It provides a new rehabilitation training method for people with brain nerve injury and cognitive impairment,and points out a new development direction for the development of medical rehabilitation robots.