Execution Time Oriented Design of an Adaptive Controller for Mobile Machines

Mobile and stationary mechatronic systems are often driven by hydraulic actuators which are controlled using valves. Hydraulic systems have a high power density, are robust and cost-effective in the application and can work reliably under rough environmental conditions. For many years, PID controllers build the state-of-the-technology for valve control. Although PID controllers are easy to synthesize and can be executed on low-performing devices, the demand for more adaptive control approaches to deal with non-linearities in hydraulic circuits is increasing. Adaptive controllers tend to be more computationally expensive than PID controllers, so the capability of common hardware devices to execute the algorithms in real-time must also be considered. This paper contributes a feasibility study examining whether adaptive control approaches can be used for valve control in hydraulic-driven mobile machines. Thus, a literature review on adaptive valve control approaches is conducted and a model-based linear adaptive control approach consisting of system identification and control logic execution is extracted. The adaptive controller is applied to a hydraulic testbed and the controller performance is measured in dependence on the cycle time of the controllers to estimate the worst-case execution time. It is investigated whether common control units in mobile machines can reach the execution times and how many valves can be controlled by one control unit in terms of the execution time.