CONTROL OF VEHICLE VIBRATION PROTECTION SYSTEMS BASED ON REINFORCEMENT LEARNING UNDER UNCERTAINTY

The article presents the results of studies of the control system of vibration protection mechanisms of a car based on reinforcement learning under uncertainty. The problem of occurrence of uncertainties when using a trained control system in a new environment is considered. It is shown that as a solution to the problem it is proposed to take into account possible of them in advance and introduce certain functions that allow the system to maintain operability. A solution to the problem of control of an inverted pendulum is presented. The problem was solved on the basis of Matlab and its software part Simulink.