• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>IEEE Symposium Series on Computational Intelligence >A Game Theory Approach to Attack-Defense Strategy for Perception of Connected Vehicles
【24h】

A Game Theory Approach to Attack-Defense Strategy for Perception of Connected Vehicles

机译：攻击型汽车攻击战略的博弈论方法

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The malicious attacks to vehicle communication facilities can expose connected vehicles to security threats, making them lose control and causing traffic accidents. It is usually hard to predict the attacked targets, which makes it difficult to design a proactive defensive strategy. This study tackles the issue of connected vehicles under bounded data injection attacks. We propose an active attack-defense model based on game theory and solve the model using a reinforcement learning-based method. The vehicle-attacker system is modeled as a game-theoretic framework based on the generalized weakened fictitious play, in which each player uses the best actions to react to the opponents’ empirical actions. A novel mixed adversarial reinforcement learning is employed to learn the attack-defense model which makes the model self-evolutionary. Moreover, our model is validated in a car-following scenario through numerical simulations and comparative studies under different forms of attacks. Results show that our model can recognize the attacked sensors and the defensive effectiveness of connected vehicles increases by nearly 30% under bounded attacks.
机译：对车辆通信设施的恶意攻击可以将连通的车辆暴露在安全威胁中，使其失去控制并造成交通事故。通常很难预测攻击的目标，这使得难以设计主动的防御策略。本研究解决了有界数据注入攻击下连通车辆的问题。我们提出了一种基于博弈论的积极攻击防御模型，并使用基于强化学习的方法解决模型。车辆攻击者系统是基于广义弱化的虚拟游戏的游戏理论框架，其中每个玩家使用最佳行动来对对手的经验行动作出反应。采用一种新颖的混合抗逆性加强学习来学习使模型自进化的攻击防御模型。此外，我们的模型通过不同形式的攻击形式的数值模拟和比较研究验证了汽车之后的情景。结果表明，我们的模型可以识别受到攻击的传感器，并且连接车辆的防御性效果在有界攻击下的近30％增加。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号