Research On Upper Limb Motion Intention Recognition Based On Pressure And Attitude Signal Analysis

With the progress and development of science and technology,more and more intelligent devices have entered people's lives.Exoskeleton robot is one of them.Exoskeleton robot is an intelligent device that is worn on the user and assists people to move according to their intention.In the research of the exoskeleton robot,how to accurately identify the human motion intention is the key point of the robot control system.At present,intention recognition technology is mainly divided into electrophysiological signal recognition,such as EEG,EMG,and general physical signal recognition,such as posture signal,planar pressure signal,human-computer interaction force signal,etc.Electrophysiological signals are weak for motor intention,easy to be affected by the environment,and difficult to obtain,so intention classification and recognition are generally studied.Physical signals are easy to obtain and good antiinterference,which can specifically reflect the kinematic parameters of the motion intention,and is more conducive to the control research of the robot.By analyzing the functions of different exoskeletons and the application advantages of different intention recognition methods,this paper proposes an upper limb motion intention recognition method based on human-computer interaction pressure information and attitude information for the assisted upper limb exoskeleton.The pressure sensor and data acquisition system for interactive force information acquisition are designed,and the validity of the method is verified by experiments.Firstly,this paper introduces the research status of intention recognition methods of different exoskeleton devices,proposes a method to identify the motion direction of each joint of the upper limb by using human-computer interaction force information through analysis,and establishes the corresponding relationship between the magnitude and direction of interaction force and the motion intention of the joint.Secondly,according to the identification scheme,a pressure sensor is designed to collect the information of human-computer interaction force.Considering the function and comfort of the sensor,a strain pressure sensor is designed and manufactured using the air bag as an elastomer.The attitude sensor is introduced,the hardware circuit of the data acquisition system is designed according to the sensor,and the platform is built.Then,Keil is used to develop the software design and development of the data acquisition system to realize data acquisition,filtering,calculation and transmission.By analyzing the output of pressure sensor,pressure data was filtered.By comparing and analyzing the characteristics of three attitude information solutions,namely,complementary filtering,Kalman filtering and MDP,MDP was selected to solve attitude signal data.Finally,the testing experiment of the pressure sensor on the interaction force is designed to analyze the effect of the pressure sensor on the detection of the magnitude and direction of the interaction force.The prediction ability of the main vector of interaction force to motion intention is verified by designing the experiment of forearm buckling motion.The prediction ability of the principal moment of interaction force to the motion intention is verified by designing the experiments of forearm rotation motion.