Research On Structure Design And Control Algorithms Of Bionic Upper Arm Driven By Pneumatic Muscle Actuators

Pneumatic muscle actuator(PMA)is a kind of pneumatic actuators which can stretch or contract,similar to biological muscles.It has many advantages,such as high power / quality ratio,fast response,small noise,and has broad prospects of application.But PMAs can only provide axial force,and most of them are used to produce a rotational joint through two PMAs pulling with each other.Since mechanical characteristics of PMAs are similar to human muscles,they can be applied to design the structure of the bionic shoulder joint,which has larger rotation range and higher flexibility.Then the structure was optimized to obtain better mechanism performance,and different control algorithms were used to control the bionic upper arm.The concrete works are as follows:Firstly,this article analyzed bones and muscles of human upper arm and the mechanization of their motion.The serial-parallel structure was selected as the basic structure of bionic upper arm,and the topology analysis method was used to determine the rationality of mechanism.Then,the structure of bionic upper arm was designed.Secondly,the kinematics and dynamics of the mechanism were analyzed.Because of the particularity of serial-parallel mechanism,this paper built a coupling equation to calculate mechanism position when solving inverse kinematics.Kane method was used to establish the dynamic equation of the machine.Because of the redundancy of the bionic machine,dynamic equations have multiple solutions,and this paper used linear programming to calculate the output force of PMAs.At the same time,the singularity of the machine was analyzed,and the workspace was calculated by using inverse kinematics and forward kinematics.Besides,particle swarm optimization was applied to optimize the structure of bionic upper arm,and the multi-objective optimization of the humanoid upper arm was transformed into a single objective optimization by multiplying by weight coefficient.In order to improve the efficiency of optimization,adaptive algorithm was used to adjust weights in real time.Compared with the optimization results between two methods,it shows that adaptive optimization has higher efficiency and better optimization results.At last,the concept of nominal force of the PMA was redefined.According to experimental results of a single muscle,the relation curve between nominal force and shrinkage ratio were drawn,and then the mathematical model can be achieved by fitting these curves.After that,test platform was built and control program was written in the computer.The closed-loop PID control and fuzzy-PID control were designed,and two experimental movements were studied by two control methods respectively.Compared with experimental results of two control methods,the fuzzy-PID control algorithm has better control precision.