Neural-Assisted Synthesis of a Linear Quadratic Controller for Applications in Active Suspension Systems of Wheeled Vehicles

This article presents a neural algorithm based on Reinforcement Learning for optimising Linear Quadratic Regulator (LQR) creation. The proposed method allows designing such target function that automatically leads to changes in the quality and resource matrix so LQR regulator achieves desired performance. solution’s stability optimality are controller’s responsibility. However, mechanism obtaining, without expert knowledge, appropriate Q R matrices, which will lead gain realise control quality. presented was tested derived quadrant model of suspension system. Its application improved user comfort by 67% compared passive solution 14% non-optimised LQR.