Study On Lower Limb Exoskeleton Perception And Control System Based On Multi-source Information Fusion

In recent years,the development of military affairs in various countries has been accelerating,and military strategic principles of "future soldiers" have been introduced one after another.This approach mainly focuses on intelligent digital combat platform,including operations,reconnaissance,communications,protection and other aspects,in which the transformation of individual combat equipment is particularly important.Under the requirements of future battlefield variability and complexity,individual soldier can carry all kinds of equipment,adapt to various environments,complete diverse tasks,and inevitably increase the burden of individual soldier,which has a great impact on individual soldier's physical fitness and mobility.Exoskeleton is a new type of man-machine interactive robot,and its wearability makes it possible to increase the soldier's load.As a highly coupled human-computer interaction robot,how to accurately and quickly identify human motion intention and provide assistance to the wearer before harming the wearer are two major problems that need to be solved urgently.To solve these two problems,this paper studies the two core issues of human motion intent recognition and compliance control methods,and designs related experiments for application and verification.Finally,the prototype of the exoskeleton robot was built and the algorithm was verified on the prototype.Then,the plantar pressure sensor and inertia sensor are used to collect the ground reaction information and the motion information of knee joint and hip joint,and machine learning method is used to classify the normal walking,running,ascending and descending stairs,ascending and descending slopes of human body.At the same time,the departure of human body under each gait action is established.The line database provides the desired trajectory for the exoskeleton.Considering the strong coupling characteristics in the process of man-machine cooperative motion and the safety and comfort caused by the rigidity of the motor actuator,this paper uses the model-based impedance control algorithm to control the force of the motor to ensure the safety of the wearer.At the same time,the EMG signal of lower limbs is collected by EMG sensor,and the activation degree of lower limbs muscle is calculated by activation degree model,so as to estimate the muscle strength,and introduce the lower limbs muscle strength into the impedance control algorithm,so as to improve the comfort of the wearer.Finally,the perception and control algorithms are simulated and validated on the V-rep simulation platform,and the experiments are carried out on the exoskeleton physical prototype platform.