An adaptive PID control algorithm for the two-legged robot walking on a slope

The crux in designing the PID controller lies in determining its gain values, which play a major role in deciding its performance. The gains are fed as inputs to the controller and are to be decided before its run. On the other hand, the effectiveness of the biped walk purely depends on the performance of the PID controller. Initially, the upper and lower body gaits of the two-legged robot are determined using the concept of inverse kinematics. Further, the dynamics of the biped robot is derived by using Lagrange–Euler formulation. The main objective of the present research is to decide the gains of the torque-based PID controller with the help of a neural network trained by using nature-inspired optimization algorithms, namely MCIWO and PSO. The adaptiveness of the algorithm lies in modifying the gains of the controller based on the magnitude of the error in the angular displacement received at the input to the NN. Once the controller is developed, its effectiveness is tested in computer simulations. Finally, the optimum controlled gait angles obtained by the best approach are tested on a real biped robot.