重庆理工大学学报2026，Vol.40Issue(15)：64-73,10.DOI:10.3969/j.issn.1674-8425(z).2026.08.008
混合动力汽车工况识别与强化学习能量管理策略
Hybrid electric vehicle condition recognition and reinforcement learning energy management strategy
摘要
Abstract
With the rapid development of hybrid electric vehicles,optimizing energy management strategies has become crucial for enhancing vehicles' performance.To improve the adaptability of energy management strategies for plug-in hybrid electric vehicles,this paper employed clustering algorithms under standard driving conditions to extract four typical scenarios,building a scenario recognition model.It was then integrated with reinforcement learning to design an adaptive energy management strategy based on driving conditions.The weight factors of the reward function were adjusted according to different conditions.Next,a comprehensive energy management strategy that integrates scenario recognition and deep reinforcement learning was proposed.Through simulations and comparisons with dynamic programming and rule-based strategies,this approach achieved 94.29％in effectiveness compared to dynamic programming strategy while maintaining good adaptability under typical operating conditions for vehicles.This paper may provide some insights into real-time energy management for hybrid electric vehicles under multiple operational conditions.关键词
插电式混合动力汽车/工况识别/能量管理策略/深度强化学习Key words
plug-in hybrid electric vehicles/driving condition recognition/energy management strategies/deep reinforcement learning分类
交通工程引用本文复制引用
杜常清,黄家鑫,赵奕凡,满兴家,公冶祥玉,郑超..混合动力汽车工况识别与强化学习能量管理策略[J].重庆理工大学学报,2026,40(15):64-73,10.基金项目
广西科技重大专项项目(桂科AA23062058) （桂科AA23062058）
武汉市科技重大专项项目(2024010702020025) （2024010702020025）
