EMBEDDED, DYNAMIC MODEL IDENTIFICATION FOR DIFFERENTIAL MACHINE MOTION CONTROL

A real-time dynamic, machine motion model identification method and application are described for controlling multiple-link, hydraulic machines navigated by an operator. The steady-state machine control applies motion kinematics transformation. A dynamic control correction model is added to improve performance. The model parameters represent link masses, inertia, hydraulic actuator coupling, actuator and external forces, all variables during operation. The model parameters are identified and updated dynamically, real-time. A new adaptive controller is introduced in which the reference model uses transformed kinematics but no dynamics. The dynamic model parameters are directly used in the adaptive controller. The new control system is applied to a small hydraulic excavator. The test results show that the dynamic model can be effectively identified real-time and its constants directly used in the feedback control loop for reducing dynamic machine motion errors due to machine inertia, delays and other distortions in the hydraulic system.