Linear Quadratic Regulator based torque profiling during gear shift on multi-speed heavy-duty electrified transmission for optimal actuator durability, shift quality and performance

The extensive adoption of fully electric personal passenger vehicles has spurred the development and advancement of electrification on commercial vehicles. Multispeed transmissions are an attractive option for powertrain electrification due to their ability to optimize performance and cost. This paper describes an optimal control method utilizing a Linear Quadratic Regulator (LQR) to control the torque during the gear shift on a multispeed electrified transmission to optimize for clutch actuator durability and shift performance. The dynamic state-space model of the system has been obtained using System-Identification. An LQR controller is formulated to minimize driveline oscillations and transmission-input-torque using the aforementioned model by manipulating the electrical torque applied by the traction motor at the transmission input. The LQR controller is implemented in a simulation framework wherein the impact of vehicle parameters on the shift quality metrics is also assessed. Subjective and objective requirements are considered in the tuning process for the LQR controller. The LQR controller is utilized to generate a high resolution profiled torque table which is then used by a downstream breakpoint optimization algorithm to generate torque profile calibrations with significantly reduced breakpoints. These calibrations are then deployed onto a production ready Transmission Control Unit and experimentally validated on a Class-8 Heavy Duty vehicle retrofitted with an electrified multispeed transmission. Experimental validation is performed at three different Gross-Combination-Weight (GCW) configurations. Significant reduction of clutch actuation effort, shift time while maintaining adherence to subjective and objective shift quality metrics has been demonstrated compared to traditional S-Shaped and Linear torque profiling. The impact of GCW on the LQR controller formulation and tuning and consequent torque profile generation is also shown.