Cooperative optimization of energy recovery and braking feel based on vehicle speed prediction under downshifting conditions

Regenerative braking can effectively recover vehicle kinetic energy, but its energy conversion efficiency is low under low-speed conditions, and there is also the problem of premature exit from energy recovery due to insufficient reverse electromotive force. The cooperation of gearbox gears can increase the speed range for the motor to recover energy, but unreasonable shifting will cause fluctuations in braking force and affect the consistency of the braking feel. Therefore, this paper aims to collaboratively optimize energy recovery and braking force fluctuations during gear shifting. Firstly, based on the model of braking system and transmission, the influence of shift strategy on braking impact and energy recovery is studied. In view of the challenge of determining a shift strategy with uncertain target braking speeds, a speed prediction model reconstructed by the support vector regression (SVR) model and the hybrid nonlinear autoregressive neural network (NAR) is proposed. On the basis of NAR-SVR speed prediction, the coupling effect of braking impact force and energy recovery efficiency is considered, and the collaborative optimization of regenerative braking torque and shift time is solved through a multi-objective cuckoo search algorithm. The hardware-in-the-loop test results verified that under high-speed conditions, the braking energy recovery rate of the proposed strategy was increased by 47.06%, and the peak braking impact was reduced by 61.4%. This research can provide a reference for the brake downshift optimization strategy and regenerative braking research of vehicles with non-decoupled electro-hydraulic composite braking systems.