Optimal Design And Swinging Control Of Bionic Joint Driven By Pneumatic Muscles

With the application of bionic and service robot expanding, the demandfor safety, light weight, and compliance actuators is increasing. As a flexible novelactuator, pneumatic muscle use compressed air as working medium, with theadvantages of high force-weight, simple structure, and non-pollution. Those featuresmake pneumatic muscle possess vast potential development in the field of industrialcontrol, biomimetic robots, rehabilitation and physiotherapy. However, the abroadstudy of the pneumatic muscle only the recent decade, and many problems not only ofbasic theory but also of practical application to be studied and resolved. Therefore, thepurpose of this paper is lay the foundation for further used of pneumatic muscle, bythe study of characteristics and control of bionic joint.It is difficult to obtain a precise mathematical model of pneumatic musclethrough theoretical derivation as it is complex. Therefore, take McKibben pneumaticmuscle for example, obtain the kinetic parameters such as modulus of elasticity,damping coefficient and output force by experiments.Bionic joint driven by antagonistic pneumatic muscle is constrained bycontraction of both pneumatic muscles while swinging. The maximum availablecontraction cannot alters for a certain pneumatic muscle. Thus, it should optimize themechanical parameters to improve the swing angle of bionic joint. Establish themathematical model of bionic joint driven by antagonistic pneumatic muscle formechanical parameters optimizing. Simulation was implemented by SimMechanics,and effects of asymmetry mechanism to the total contraction, and pentagonalmechanism to joint compliance were given in curves. According to human arm jointcharacteristics, parameters of bionic shoulder and elbow joints were optimal designed.In order to ensures that bi-pneumatic-muscle achieve maximum contraction in thesame time, based on the given right pneumatic muscle, left one’s parameters wasdesigned reverse. Compared to symmetrical mechanism, angular range of shoulder isenhanced on7%, and that of elbow enhanced on42%; the contraction amount ofbi-pneumatic-muscle can be reduced arrive7%(most) under the same rotation scope. Optimal mechanism is more closely to human joint.Study traditional PID closed-loop control of bionic joint, and verify the theeffectiveness and feasibility of the algorithm. However, the overshoot and oscillationamplitude is large, and adjust time is long which cannot meet the requirements oftrajectory tracking. Thus, design adaptive, self-parameter-adjust neuron PID controlto solve the problem above. Experiment result shows that positional accuracy ishigher, response curve oscillation can be reduced, and reduce adjust time by neuronPID control. The steady-state error of step response is less than0.001rad (0.06°), themaximum error of sine curve trajectory tracking is less than0.025rad.(1.5°).