A generalised method for improving the performance of controllers by combining PID and DELTA rule

In this paper, the proposed method aims to implement an adaptive loop by combining the PID and DELTA rule. Here the DELTA rule is implemented by using an Adaptive Linear Neural Neuron (ADALINE). ADALINE´s weights are adjusted by using the output of PID controller. An inverted pendulum which consists of a cart hinged to a pole is used here as one of the plants. An uncertain model of inverted pendulum is used here. In order to show the robustness of proposed controller, it is also tested using an uncertain first order plant and an uncertain model of robot arm. Also a novel technique is proposed here which aims to make a controller robust and adaptive to changes in the system by applying PID and Delta rule to the controller´s output. For this backstepping control of a plant is taken here as an example. The simulation results prove that the proposed scheme is robust and adaptive and well suited for systems with uncertain variations in the plant. Simulink is used here as an interactive tool for modeling both the controller and the plants.