Anthropomorphic Motion Generation Principle And Human-robot Interaction Control For Human-like Upper Limb Robot

The human-like upper-limb robot refers to a robot with a structure and function similar to that of the human upper extremity.As a highly complex bone and muscular system,the upper limb of the human body has flexible and variable movement capability,stable and efficient load capability,and adaptability to complex unknown environments.Almost all daily activities require the participation of the upper limbs.Research on the upper limb and its functional movement has always been the focus of researchers in physiology,anatomy,biomechanics,and neurosciences.Moreover,it is also the source of inspiration for designing human-like upper-limb robots.In view of the common fundamental scientific problems of human-like upper limb robots in reconstruction/repairing natural movement of human limbs,this paper carried out systematic and in-depth research on human-like upper limb robots from several perspectives,such as extraction of the natural movement characteristics,anthropomorphic motion planning of humanoid upper limb robot,human-robot interaction control,and clinical application.This paper carried out the following innovation research:The anatomical structure and motion characteristics were studied in the context of the upper limb’s natural functional movements in daily life.Then a seven-degree-of-freedom kinematic chain of the upper limb is established,and an experimental model for collecting the upper limb movement data irrelevant to the structural parameters of the volunteers’ bodies is proposed.Furthermore,a comparative study is conducted between healthy volunteers and volunteers with hemiplegia after stroke for the movement characteristics of the upper limb.The results revealed the similarity of movement characteristics of healthy volunteers’ upper limbs and the variability of the synergy characteristics of both limbs of volunteers with hemiplegia.Inspired by the human motor control and learning mechanism,the method for forming the proposed human movement of upper-limb robots is presented.The natural movement characteristics of the upper limbs are extracted using the Gaussian mixture model with the human-like arm robot as the research carrier.The anthropomorphic motion targets in joint space are calculated based on the human sensorimotor model.The Gaussian mixture regression method establishes a continuous task parameterization model of matching characteristic motion types.Finally,a collision-free motion trajectory planning method in an unstructured environment that satisfies the requirements of motion smoothness and anthropomorphism is proposed by using the functional analysis method.Experimental results show that the anthropomorphic degree of trajectories generated by the proposed method can reach more than 90.8%.Based on the clinical needs of upper limb rehabilitation treatment for patients with stroke,the method of human-like upper-limb rehabilitation exoskeleton robot design and its human-robot interaction control strategy design are proposed.And with the exoskeleton robot as the carrier,the posture synergy-based mechanism design principle is presented.This principle embeds the natural movement characteristics of upper limbs in the mechanical structure of the exoskeleton.So,the exoskeleton robot has the inherent property of reproducing the natural movement characteristics of the upper limb.The effectiveness and reliability of the proposed control method are validated by the recognition of the participant’s motion intention.Due to the lack of clinical evidence on the effectiveness of rehabilitation robots worldwide,a clinical study consisting of a preliminary clinical trial and a two-center randomized controlled clinical trial for the upper-limb rehabilitation exoskeleton robot is designed and carried out.This clinical study is the first clinical trial research for the exoskeleton robot with independent intellectual property rights in China.The clinical effectiveness of the rehabilitative exoskeleton robot was verified using various quantitative evaluation methods.The clinical trial showed that the rehabilitation exoskeleton robot could meet clinical needs.