Pareto design of an adaptive robust hybrid of PID and sliding control for a biped robot via genetic algorithm optimization

• Published in: Nonlinear dynamics 2015-01, Vol.79 (1), p.251-263
• Main Authors: Taherkhorsandi, M., Mahmoodabadi, M. J., Talebipour, M., Castillo-Villar, K. K.
• Format: Article
• Language: English
• Subjects: Adaptive control; Adaptive control systems; Automotive Engineering; Classical Mechanics; Control; Control stability; Control systems design; Controllers; Derivatives; Design optimization; Dynamic stability; Dynamical Systems; Engineering; Genetic algorithms; Mathematical analysis; Mechanical Engineering; Motion planning; Motion stability; Multiple objective analysis; Nonlinear control; Nonlinear dynamics; Optimal control; Optimization; Original Paper; Pareto optimization; Proportional integral derivative; Robot control; Robots; Robust control; Sliding mode control; Tracking control; Tracking errors; Vibration; Walking
• ISSN: 0924-090X; 1573-269X
• DOI: 10.1007/s11071-014-1661-1

In case of helping elder people and doing complicated jobs, biped robots have greater capabilities in comparison to other mobile robots, such as wheeled robots. However, owing to their extremely nonlinear dynamics and natural instability, their motion planning and control become a more challenging and crucial task. Thus, this paper tends to present an adaptive robust hybrid of PID and sliding control optimized by multi-objective genetic algorithm optimization to control a biped robot walking in the lateral plane on slope. More precisely, proportional, integral, and derivative (PID) control is a reliable and stable controller and sliding mode control (SMC) is a robust controller having an appropriate tracking function. To utilize these unique features of each controller, optimal SMC is employed as a supervisory controller to enhance the performance of optimal adaptive PID control and provide sufficient control input. An adaptation mechanism is used to update the proportional, integral, and derivative gains of PID control online. To eliminate the tedious trial-and-error process of determining the control coefficients, multi-objective genetic algorithm optimization is utilized to design and choose the control parameters by an optimal approach. Three points of the Pareto front of the genetic algorithm optimization are selected to design the controller. While the dynamic equations of the biped robot walking in the lateral plane are immensely nonlinear, the control method can operate effectively and the results demonstrate the proper performance of the controller in two aspects of low tracking error and optimal control input.