基于预测模型的无人赛车路径规划算法研究  

Research on Path Planning Algorithm of Driverless Race Car Based on Predictive Model

作　　者：龚国铮 郑少武 钟思祺 叶鸣 李巍华[1] Gong Guozheng;Zheng Shaowu;Zhong Siqi;Ye Ming;Li Weihua(South China University of Technology,Guangzhou 510641;Guo Dong Intelligent Drive Technologies Co.,Ltd.,Shenzhen 518054;Guangzhou Huagong Automobile Inspection Technology Co.,Ltd.,Guangzhou 510640)

机构地区：[1]华南理工大学,广州510641 [2]深圳裹动智驾科技有限公司,深圳518054 [3]广州华工机动车检测技术有限公司,广州510640

出　　处：《汽车技术》2022年第7期32-41,共10页Automobile Technology

基　　金：广东省重点领域研发计划资助项目(2019B090912001)。

摘　　要：针对无人驾驶电动方程式赛车极限工况下的路径规划问题,提出一种基于预测模型的无人赛车路径规划算法。通过目标函数设计将路径规划问题转化为最优化问题,并将车辆动力学预测模型作为等式约束,依据赛车的稳态行驶特性,构建赛车行驶稳定性组合约束,并综合考虑边界约束,生成满足赛车行驶要求的路径。在不同附着条件下进行了仿真测试,结果表明,在满足实时性要求的前提下,所提出的算法能够有效减小车辆前、后轴侧滑的风险,且在多项指标上优于RRT^(*)算法以及基于三次自然样条曲线的最优二次轨迹规划方法。For the path planning of the driverless electric formula car under extreme conditions,this paper proposes a path planning algorithm based on prediction model.By designing the objective function,the path planning problem is translated into an optimization problem and the vehicle dynamics predictive model is used as an equality constraint.According to the steady-state driving characteristics of the race car,the combination constraint of the driving stability is constructed,combined with boundary constraints to generate the desired path of the race car.Simulation test is carried out under different adhesion conditions,which show that on the premise of meeting the real-time requirements,the proposed algorithm can effectively reduce the risk of vehicle front and rear axle sideslip and is superior to the RRT^(*)algorithm and the optimal quadratic trajectory planning method based on cubic natural spline curve on multi metrics.

关 键 词：无人驾驶赛车 路径规划 模型预测控制 操纵稳定性 最优路径 

分 类 号：U461.6[机械工程—车辆工程]