Particle swarm optimization-based optimal power management of plug-in hybrid electric vehicles considering uncertain driving conditions

• The particle swarm optimization was employed to determine an optimal power management.
• Relationship between the driving condition and optimization result was investigated.
• The driving condition recognition method was proposed via a fuzzy logic controller.
• Optimal power management against uncertain driving conditions was proposed.
• The proposed strategy can reduce the energy loss by 1.76% for unknown driving cycles.

This paper proposes a novel optimal power management approach for plug-in hybrid electric vehicles against uncertain driving conditions. To optimize the threshold parameters of the rule-based power management strategy under a certain driving cycle, the particle swarm optimization algorithm was employed, and the optimization results were used to determine the optimal control actions. To better implement the power management strategy in real time, a driving condition recognition algorithm was proposed to identify real-time driving conditions through a fuzzy logic algorithm. To adjust the thresholds of the rule-based strategy adaptively under uncertain driving cycles, a dynamic optimal parameters algorithm has been further established accordingly, and it is helpful for avoiding the problem that the thresholds of the rule-based strategy are very sensitive to the driving cycles. Finally, in combination with the above efforts, a detailed operational flowchart of the particle swarm optimization algorithm-based optimal power management through driving cycle recognition has been proposed. The results illustrate that the proposed strategy could greatly improve the control performance for different driving conditions. Especially for the uncertain driving cycles, the reduction in energy loss can be up to 1.76%.