Series Elastic Actuator Cascade PID Controller Design Using Genetic Algorithm Method

Abstract

Series Elastic Actuator (SEA) has many advantages over traditional stiff actuator, such as intrinsic safety and reduction of energy consumption. SEA has been successfully applied in the robotic field as exploration vehicles or actuation mechanisms on bipedal robots and intelligent prostheses. However, a good performance of this electromechanical system depends a lot on the design of its control system. This paper deals with the design of a cascade PID controller based on genetic algorithm (GA) method for SEA position control. The purpose of this controller makes the system settle faster, the steady-state error tends to zero, the overshoot lower, and makes the system less sensitive to disturbances. After built of mathematical model and compute of the system transfer functions, GA method is used to tune, simultaneously, the inner (torque control) and outer (impedance control) loops of cascade PID controller. The sum of integral absolute error (IAE) values of two controller inputs is used as the objective function. The performance of the designed controller is evaluated by simulation under the MATLAB/Simulink software. The same controller is re-designed using the Ziegler-Nichols method in order to compare both methods in terms of response performance and robustness. The comparison shows that the GA method is more effective than conventional method.