Project Title: Earth Moving Equipment Electro-Hydraulic Actuator Control Using Learning CMAC Neural Network Control Algorithm

Research Team: Dr. Sabri Cetinkunt, Unnat Pinsopon, Chenyao Chen

Sponsor: Caterpillar Inc.

Project Description:

Hydraulic power based actuation has been the only practical actuator technology in mobile equipment applications. The hydraulic actuators are the main actuation technology used in almost all of the earth moving equipment. The increasing competition in the world market requires that the performance of the hydraulic control systems be improved. The current control systems utilize the industry standard proportional-integral-derivative (PID) control algorithm plus model based feed forward compensation algorithms. The PID control algorithm can not effectively deal with the dead-band and the valve gain variations as function of load and engine speed. Furthermore, the model based compensation requires extensive off line modeling and system identification for each line of product. For instance, one major manufacturer has 1700 different types of hydraulic servo valves in their product line. The goal of this project is to improve the dynamic performance of the hydraulic control system by using advanced control algorithms. It is further desired that the off-line modeling and system identification work be reduced by the use of intelligent, learning control paradigms.

To this end, we have been developing adaptive control systems based on neural network concepts for the earth moving equipment hydraulic system control. These control systems are used in the implement system, the steering system, the continuously variable transmission, and the brake systems.