A swarm intelligence-based predictive regenerative braking control strategy for hybrid electric vehicle

Braking energy recovery is one of the main technologies affecting the economic performance of an electric vehicle. To improve economy from as much recovered braking energy as possible on the premise of ensuring vehicle security is the goal of regenerative braking control strategy. However, due to the non-linear and multi-objective characteristics of hybrid braking system, finding the optimal regenerative braking control strategy, considering safety, economy, and comfort, remains a challenge. Considering the efficient characteristics of regenerative braking system and battery aging, a swarm intelligence-based predictive regenerative braking control strategy is proposed. Particle swarm optimisation is used as the main part of the strategy, the ant colony algorithm is used to modify the iterative process of particle swarm optimisation to avoid convergence to a locally optimal solution, and model predictive control theory is applied in the control strategy to realise the optimal control. Then, under emergency braking conditions and urban cycling conditions, the stability and economy of proposed strategy are test by the simulation experiments. Finally, to reduce the computational complexity of the control strategy, an equivalent control strategy is proposed based on the nearest point method, and its effectiveness is verified by hardware-in-loop experiment.

Automated summary

Braking energy recovery is a crucial technology for electric vehicles' economic performance. However, finding the optimal regenerative braking control strategy considering safety, economy, and comfort remains challenging due to the non-linear and multi-objective characteristics of hybrid braking systems. In this study, a swarm intelligence-based predictive regenerative braking control strategy is proposed, which is tested for stability and economy through simulation experiments. Additionally, an equivalent control strategy is suggested to reduce computational complexity.