Dual Beetle Antennae Search system for optimal planning and robust control of 5-link biped robots

Abstract

In this paper, we presented an optimization-based control framework for the trajectory planning and the stable control of multiple non-linear systems, i.e., 5-link Biped Robot, inverted pendulum, and acrobot using the BAS (Beetle Antennae Search) algorithm with PID (Proportional–Integral–Derivative) controller. Traditional approaches treat path planning and stable control of the system as a separate problem, or multiple controls are employed to achieve both goals. Therefore, these approaches are complex, computationally, and time-wise expensive. The addition of disturbance to the system makes the control further intricate. In this paper, we proposed an optimization-based approach for the path planning and stable control of the non-linear systems by employing a hybrid controller, which includes BAS and PID. We devised a switching mechanism, where BAS will first perform the trajectory optimization through feedforward control. In the second phase, a disturbance is introduced in the system, and BAS is used to optimize the K values of PID for the control stability. In the simulation, we applied this technique on three non-linear systems, i.e., 5-link Biped Robot, inverted pendulum, and acrobot, and promising results are obtained. In addition we performed a comparison of dual Beetle Antennae Search algorithm with known swarm algorithms, i.e. PSO (Particle Swarm Optimization) and GA (Genetic Algorithm). The results showed that BAS and PSO are comparable. For example, in inverted pendulum the objective function converges to 7.566×10−4 and 0.029 by BAS and PSO respectively. Likewise, in 5-link Biped Robot the objective function converges to 0.0250 and 0.0351 by BAS and PSO respectively. In addition, the number of function evaluations in case of BAS increases linearly, i.e., 3 times of the iterations. However, the function evaluations increase exponentially in case of PSO and GA.