PID Based Optimal Neural Control of Single Wheel Robot (SWR)

This study investigates the application of Proportional-integral-derivative (PID) based Artificial neural network (ANN) controller for the stabilisation of highly nonlinear and multivariable Single wheel robot (SWR). The nonlinear governing equations of motion for the proposed system were derived using Newton’s second law. The gains of PID controller were optimised using auto-tuning function. The results of PID were used for training of ANN controller designed using optimal number of neurons in the hidden layer. The performance of the proposed controller was measured in terms of settling time, overshoot ranges and steady state error responses. The results indicate superior performance of ANN controller designed using 30 neurons in the hidden layer.