Integrated Torque and Inertia Phase Model Predictive Controls of Automatic Transmission Upshifts

Advanced step-ratio automatic transmission shift control that can provide optimal shift performance and reduce calibration effort is highly sought after in today's product development environment. This article explores Control Allocation (CA) and Model Predictive Control (MPC) concepts to meet the requirements for both conventional and electrified powertrains such as parallel hybrid electric transmissions. The CA and MPC strategies, previously proposed for the upshift inertia phase, are extended in this article to torque phase to obtain integrated optimal shift control strategies. These strategies optimally coordinate oncoming and off-going clutch torque with engine torque to balance the conflicting shift quality criteria reflecting shift comfort, duration, and efficiency. The designer sets the shift quality requirements through torque and inertia phase durations, reference and constraint profiles, and cost function weights, which facilitates calibration of the control strategy performance. To further improve the performance, the CA strategy is extended with a virtual torque sensor system during the torque phase. Two variants of MPC strategy are proposed and compared: (i) separate MPCs for torque and inertia phase and (ii) integrated MPC whose prediction model in torque phase spreads to the inertia phase. The CA and MPC strategies are realized in the form of constrained quadratic programs and efficiently solved on-line by an interior-point solver. Their performance is demonstrated and compared via nonlinear powertrain model simulations for 1-3 shift, including a system robustness analysis with respect to actuation parameter uncertainties.