混合交通环境中自动驾驶汽车安全可达性分析与优化控制研究

项目名称： 混合交通环境中自动驾驶汽车安全可达性分析与优化控制研究

项目编号： No.61503048

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 自动化技术、计算机技术

项目作者： 高凯

作者单位： 长沙理工大学

项目金额： 21万元

中文摘要： 可达性分析是一种在约束条件下求解个体不发生碰撞轨迹可达集的方法，已成为自动驾驶汽车领域的研究热点。本项目围绕自动驾驶汽车在混合交通环境中的安全可达性分析与优化控制问题开展研究。首先，研究自动驾驶汽车多工作模式切换下的动力学特性，建立车辆的非线性混合模型；其次，分析复杂交通环境下危险碰撞条件，探究与周围其他车辆的协作驾驶行为，提出一种自动驾驶汽车安全可达性分析方法，获得最大可控安全域；然后，考虑车辆综合性能指标，设计基于牛顿法的极值搜索控制器；最后，基于自动驾驶汽车混合动力学模型和安全可达集条件，寻找工作模式切换与系统稳定性之间的关联条件，构建多向量Lyapunov函数，得到有限时间稳定的充分条件。本项目为自动驾驶汽车安全控制提供了一种新的理论分析方法和快速优化的控制方法，对于提高自动驾驶汽车在混合交通环境中的安全性，加快自动驾驶汽车技术的应用和产业化步伐具有重要的理论意义和应用价值。

中文关键词： 自动驾驶；多目标优化；城市交通；协调控制；运行控制

英文摘要： Reachability analysis is a method that is used to get trajectory reachability set without collision under some restriction, which is becoming a hot topic about autonomous driving.This project will study the safety set calculation and optimal control problems for autonomous driving vehicles in a hybrid complex traffic environment. Firstly, we investigate the dynamic characteristics of autonomous driving vehicle working in different modes. Then a nonlinear stochastic hybrid model is established for vehicles. Secondly, the collision condition for vehicles in different modes is analyzed in a complex traffic environment. The cooperation and interaction between the autonomous vehicle and neighboring vehicles is studied. Then a new safety reachability analysis method is proposed to calculate the largest controllable domain. Third, a comprehensive performance indicator is designed, achieving a tradeoff between the safety and comfort. Then a newton based extremum seeking controller is designed. Fourth, the connection between mode switching and system stability is established based on the dynamic model and the safety reachability condition. A multi-vector Lyapunov function is constructed and a sufficient condition for finite-time stability is obtained. This project provide a new analysis method and a rapid optimal control method, which is of theoretical significance and application value to both the commercialization of autonomous driving vehicles, and the safety and efficiency of modern traffic systems.

英文关键词： Autonomous driving;Multi-objective optimization;Urban traffic;Coordinated control;Operational control