Research On Bionic Design And Human-like Motion Control Of Robotic Arm Driven By Pneumatic Artificial Muscles

Humanoid robots have similar body structure,motion pattern and tactile perception ability to human beings,and can meet the requirements of cooperative robots in security,motion flexibility and human-like motion mode.Therefore,humanoid robots have high academic research value and engineering application value.The design and control of bionic humanoid arm is an important part of the research of humanoid robots.It is of great significance to improve the interaction security and efficiency of cooperative robots.This paper focuses on the bionic design and human-like motion control of humanoid robotic arm.The main work of this paper includes:(1)By imitating the configuration and driving mode of the human arm,the bionic structure of the 7-DOF humanoid robotic arm driven by antagonistic pneumatic muscles is designed and the mechanical and electrical control system is built.The designed robotic arm conforms to the standard MMM(Master Motor Map)human motion model.Its shoulder,elbow and wrist joints have the same degree of freedom and range of motion as the human arm.Shoulder joint has three degrees of freedom,elbow and wrist joints have two degrees of freedom and the axes of shoulder,elbow and wrist are intersected at one point respectively.By imitating the antagonistic driving mode of human muscles,each degree of freedom is driven by a pair of antagonistic pneumatic muscles.Pneumatic muscles are attached to the joint axis through the Bowden line.Because each degree of freedom is driven independently,the degree of freedom of the joint does not affect each other and the control of different antagonistic muscles is decoupled.Therefore,the bionic robotic arm designed in this paper has the characteristics of high anthropomorphic degree,good intrinsic safety and easy to realize its driving control.(2)Simulating the tactile perception ability of human fingers,a flexible tactile perception system of robot finger belly is designed based on air pressure sensor.The tactile perception system includes a flexible tactile sensor and a signal processing module.It can well sense the contact force of the robotic finger,roughness,softness and relative sliding of the object.The flexible tactile sensor has a similar shape with finger belly,which is installed at the fingertip of the robotic arm to facilitate the flexible operation of environmental objects.The internal pressure of the flexible airbag has a good linear relationship with the contact force(0-1.6 kg)and the linearity is 1.42%,so the sensor can well detect the force between the finger and the environment.Through Fourier transform and Kalman filter,the roughness of the object surface can be judged by analyzing the frequency of the sensing signal when the finger actively slides on the surface of the object.The softness of the object can be judged by combining the joint motion information and contact force when the finger presses on the object.The sliding motion of the object between the fingers can be judged by combining the tangential force of the finger and the vibration information of the sensing signal.Therefore,the designed flexible perception system can perceive abundant hand tactile information and is suitable for human-robot cooperative applications.(3)Based on the performance index of human-like motion control,the optimal control algorithm of the robotic joint driven by antagonistic pneumatic muscles is designed.In this paper,the human-like motion control of the robotic joint driven by a pair of antagonistic pneumatic muscles is considered as a non-linear optimal control problem.Firstly,the mathematical model of the robotic joint is established.Then,the extended state observer is designed to realize the state feedback linearization of the robotic joint model and the standard integrator series model is obtained.Finally,the optimal control law is designed based on the linear model with the minimum jerk as the performance index.The experimental results show that the proposed algorithm can realize the human-like motion of the robotic joint in a wide range(0-120 degrees).The joint trajectory is insensitive to the change of load(1-5 kg)and the joint motion has good anti-disturbance ability.The algorithm proposed in this paper has only two parameters to be tuned.The relationship between the parameters and the control effect of the system is studied through experiments and the tuning rules of the parameters are given.The proposed algorithm is suitable for human-like motion control of the robotic joint driven by antagonistic pneumatic muscles.It is simple in form and has few parameters to be tuned so it has high engineering practical value.