Drivetrain clunk control via a reference governor

Presence of backlash in an automotive drivetrain leads to an undesirable phenomenon during torque delivery called clunk that adversely affects the drivability of a vehicle. Torque shaping controllers can help in reducing clunk, but may cause the vehicle to feel sluggish due to their conservative torque delivery during lash crossing. In this work, we develop a novel, model-based torque shaping clunk controller using a landing reference governor approach, which improves the drivability of a vehicle by reducing clunk while also quickly delivering the driver requested torque. The competing design requirements of the controller are framed as an optimal controls problem with hard constraints, and this optimization is performed offline. The performance of the designed controller is assessed using experimentally-validated simulation models, and the controller’s robustness is analyzed for variations in driver requested torque. Our controller is easily calibratable, and implemented as look-up tables (LUTs) for making it possible to run on embedded processors