Extended Stated Model Predictive Control Algorithm for Gearbox Gearshifting of Heavy-Duty Trucks with Operation Condition Prediction

The optimization of gearbox ratios is crucial for achieving a balance between fuel consumption, power performance, and comfort in vehicles by regulating engine operating conditions. However, challenges exist due to the complexity of powertrain efficiency and uncertainty of future operating conditions. In this paper, a gear model prediction control algorithm based on future operating condition prediction and model accuracy compensation is proposed for a heavy-duty truck equipped with an automated manual transmission (AMT). First, a power demand and expected vehicle speed prediction algorithm is proposed using historical throttle pedal action data and vehicle dynamics model, aiming to compensate for driver intention uncertainty. Second, for the predicted operating conditions, a nonlinear model predictive control (NMPC) is used to optimize the transmission ratio with multiple objectives such as fuel consumption, power performance, and smoothness, and then discretized to the required gear ratio. Finally, the deviation between the NMPC model and the actual powertrain system is actively observed and compensated using an extended state observer (ESO) to improve the robustness to uncertainty factors such as vehicle weight and slope. Simulation results show that: 1) Under the condition of deep throttle acceleration, the future operating condition prediction MPC algorithm can save about 5.67% fuel. 2) With the use of ESO, the vehicle speed tracking error is reduced by 27.6%, and energy consumption is improved by 3.51% when the vehicle cargo weight changes by 50%. 3) In a typical 65-second acceleration segment, the fuel consumption of this algorithm is reduced by 0.64% compared with the traditional offline MAP algorithm, and the vehicle speed tracking error is reduced by 2.52%.