The invention discloses a game theory-based surrounding vehicle interaction behavior prediction method. The method comprises: collecting state information of an automatic driving vehicle and surrounding vehicles in real time; predicting future motion trails of the automatic driving vehicle and surrounding vehicles; establishing a driving profit evaluation function to perform profit calculation onthe movement of the automatic driving vehicle and the surrounding vehicles under different driving behaviors; through a game theory method, carrying out modeling analysis on interaction behaviors of surrounding vehicles and automatic driving vehicles, and carrying out probability prediction on the future interaction behaviors of the surrounding vehicles by comprehensively considering future possible movement conditions of the surrounding vehicles and expected revenue under different interaction behaviors of the surrounding vehicles and the automatic driving vehicles. According to the method, the uncertainty of the motion trail caused by the interaction behavior between the automatic driving vehicle and the surrounding vehicles is fully considered, and the possible interaction behavior of the surrounding vehicles in the future is predicted, so that the automatic driving vehicle is safer and more reliable during decision planning.
本发明公开了一种基于博弈论的周围车辆交互行为预测方法,该方法包括实时采集自动驾驶车辆与周围车辆的状态信息;对自动驾驶车辆与周围车辆未来运动轨迹进行预测;建立行驶收益评估函数对自动驾驶车辆与周围车辆采取不同驾驶行为下的运动进行收益计算;通过博弈论的方法,将周围车辆与自动驾驶车辆的交互行为进行建模分析,并且综合考虑周围车辆未来可能的运动情况以及其与自动驾驶车辆不同交互行为下的期望收益,对周围车辆未来的交互行为进行概率预测。本发明充分考虑到了自动驾驶车辆与周围车辆之间的交互行为所带来的运动轨迹的不确定性,通过对周围车辆未来可能的交互行为进行预测,使得自动驾驶车辆在做决策规划时,更加安全可靠。
Game theory-based surrounding vehicle interaction behavior prediction method
一种基于博弈论的周围车辆交互行为预测方法
2020-06-12
Patent
Elektronische Ressource
Chinesisch
Europäisches Patentamt | 2021
|