Study On HRI Technology Of Social-assisted Robot For Rehabilitation Of Celebral Palsy

Cerebral palsy cannot be cured with the current medical level,and it needs to be targeted exercise to gradually approach the normal development level.Most children with cerebral palsy are unable to communicate normally.Without timely and targeted social rehabilitation training,children with cerebral palsy will gradually lose touch with the social group as they grow older.This thesis aims to design a rehabilitation training system for children with cerebral palsy who have physical movement impairments and cannot master social body movement language,with the advice and guidance of rehabilitation experts and special school teachers.Through the design of social interaction scenarios,children with cerebral palsy can communicate and interact with a social assistance robot under the robot’s guidance,thus realizing social body movement rehabilitation training for children with cerebral palsy.The main content of this paper is described as follows:First,a brief overview of the research background is given,the current state of research in human-computer interaction techniques for cerebral palsy rehabilitation is analyzed,and finally,the article’s structure is briefly explained.Secondly,a survey was conducted on the mastery of social body movements of children with cerebral palsy.Through the analysis of shoulder shrugs in film and television dramas and real-life social interactions,standard shoulder shrugs were selected and guidelines for determining shoulder shrugs were developed.Thirdly,a social action interactive cerebral palsy rehabilitation training system is designed,the Kinect device is used to collect the position information of human bones,and the body movement recognition is realized by processing the collected bone information.Give feedback to realize the interactive function between the patient and the expression robot.Fourthly,through the comparison of various action recognition algorithms,the support vector machine algorithm is selected to optimize the action recognition system,improve the action recognition rate of the system,and effectively help children with cerebral palsy carry out rehabilitation training.Design the training result analysis module,automatically analyze the training results,generate rehabilitation training reports,use the KNN algorithm to classify the training results,and automatically generate the next stage of rehabilitation training programs.Finally,a social action-guided heuristic cerebral palsy rehabilitation training system is designed.By building a simulation robot,the human-computer interaction system is improved,and the dynamic simulation robot is used to guide the actions of children with cerebral palsy,guiding and inspiring the children in specific social action directions,and the children can imitate and learn according to the different social actions made by the robot.