Synthesis of Neural Network Compensation of an Autonomous Electro-Hydraulic Actuator Based on a Sliding Mode Control

Autonomous electro-hydraulic actuators (AEHA) are widely used in industrial applications such as mobile manipulators, robotic exoskeleton, aerospace applications. AEHA has a number of significant advantages in relation to the electric drive: a high ratio of the developed force to the weight of the drive, compact dimensions of the autonomous drive, which are provided by the integrated layout of all elements of the electro-hydraulic system. However, the control of such systems becomes much more complicated due to the presence of significant nonlinearities in the control object. These include the non-linear characteristic of the switchgear, dry friction, and hard mechanical stops. All this makes it difficult to implement a control system based on classical controllers. The paper proposes a technique for synthesizing a controller that provides a sliding mode with compensatory neural network control for objects with an indefinite static characteristic. Using the direct Lyapunov method, we prove the stability of a feedback system. Numerical simulation results are presented to demonstrate the effectiveness of the proposed approaches. The study was supported by the Russian Science Foundation grant No. 23-29-00609, https://rscf.ru/project/23-29-00609