A self-tuning neuromorphic controller to minimize swing angle for overhead cranes

Overhead cranes in manufacturing industries are generally operated manually or by some orthodox control methods. The crane operator focusses on reducing the undesired oscillations developed during the crane movement and make the trolley location converge to desired position precisely. In this study, a self-tuning neuromorphic controller technique is used for online adaptive control of a non-linear crane-mass system by applying a linear quadratic regulator for controlling the trolley position and swing motion of the crane system with an ability to adapt itself with the varying parameters and external disturbances. To achieve this, a Generalized Adaptive Linear Element (GADALINE) Artificial Neural Network (ANN) is proposed that updates weight and bias states which subsequently minimizes the error function. Additional control can be achieved with the application of momentum term with the ADALINE model to diminish the zigzag effect in weight adjustment and to accelerate the convergence of the network. This added functionality provides robustness to deal with variation in parameters.