The invention provides a multi-subject joint reinforcement learning method for bidirectional model migration, which is used for resisting multipoint poisoning attacks in the field of intelligent transportation and optimizing regional traffic efficiency. The method uses a reinforcement learning method to reinforce intersections of a traffic road network to serve as complex individuals, uses a multi-agent reinforcement learning method and a model transfer learning method to carry out regional combination on the individuals to resist poisoning attacks for an intelligent traffic signal (ISIG) system, and uses a bidirectional model transfer mechanism to reversely remove the combination, so as to improve the reliability of the system. And the security of the traffic signal planning algorithm and the regional traffic efficiency under the condition of no attack are improved.
本发明提供了一种双向模型迁移的多主体联合强化学习方法,用于在智能交通领域抵御多点投毒攻击,优化区域通行效率。该方法使用强化学习方法加固交通路网的各交叉口作为联合体个体,使用多智能体强化学习方法、模型迁移学习方法,将个体进行区域联合以抵御针对智能交通信号(ISIG)系统的投毒攻击,并利用双向模型迁移机制逆向解除联合,提高无攻击时交通信号规划算法安全性及区域通行效率。
Multi-subject joint reinforcement learning method for bidirectional model migration
一种双向模型迁移的多主体联合强化学习方法
2024-04-09
Patent
Electronic Resource
Chinese
IPC: | G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / H04L TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION , Übertragung digitaler Information, z.B. Telegrafieverkehr |
Traffic signal lamp control method based on migration reinforcement learning
European Patent Office | 2023
|Strategy migration-based fuel cell vehicle deep reinforcement learning energy management method
European Patent Office | 2022
|