Research On The Acquisition Device Of Contact Motion Information For Joint Assistance

Wearable exoskeleton robots have the abilities of protecting, supporting and assisting the body, so they can help or enhance the joint motion. The robots have broad application prospects in industrial, medical and military fields.The accurate and rapid acquisition of human motion information directly affect on the control of exoskeleton robots. This study oriented the methods of contact motion information acquisition. The surface electromyography (SEMG) acquisition device and muscle tension sensor were designed for exoskeleton robots. Muscle tension was taken as the input information of robot control, and SEMG was used to evaluate the assist effect.The SEMG device used MSP430to sample SEMG with multiple channels and deliver the state of muscle motion. Wireless data transmission based on Bluetooth module HC-05helped us to make a portable device. Input buffer circuit achieved impedance match. Floating circuit improved the common-mode rejection ratio of SEMG and prevented the circuit self-excited. Two different solutions, the hardware filter and the integer coefficients digital filter were implemented to cancel polarization voltage, high frequency noise, power interference and increase signal to noise ratio.The muscle tension sensor was designed based on the elastic performance of muscles. Two force components in the sensor were standardized on the mechanical calibration stage. A flexible rubber tyre was chosen to simulate upper arm muscles and a hardness tester was used to calibrate the muscle tension sensor. The difficulty of quantitative analysis was partially solved. A non-linear formula was invented to correct the sensor output and estimate the muscle tension.Based on MFC dialog class in Visual C++, two software was developed: interactive interface for SEMG collection; control software for exoskeleton robot. The former has the abilities of receiving, filtering, displaying and storing SEMG signal. The latter realized the muscle tension display and the exoskeleton control strategy.The muscle tension was adopted as the input indicator to control the upper limb exoskeleton. The Fuzzy-PID control strategy was designed to achieve the elbow moving and rotating. The SEMG of the biceps and triceps was collected before and after robot assist and analyzed in time-frequency domains. The SEMG indexes as MAV, RMS and Welch spectrum estimation were obtained with physiological signal processing methods. The SEMG amplitude decreased when robot assist was implemented, which illustrated the assist effect. Additionally, the principle, structure and correction method of the muscle tension sensor is credible.The muscle tension sensor and SEMG acquisition device were effective. This study lays the foundation for their wide application in exoskeleton robots in the future.