摘要
基于车辆二自由度单轨模型进行航位推算,并在此基础上,用扩展卡尔曼滤波方法融合环视相机检测结果设计了一种自主泊车定位及库位跟踪算法,可利用低成本传感器在地下车库环境下实时估计车辆位姿,同时优化环视相机库位角点的检测结果。仿真及试验结果表明,提出的定位及库位跟踪算法可以有效减小车辆运动数据噪声,提供更准确的车辆库位相对位置,为后续的规划和控制提供精度更高的定位结果。
In this research, cruising location was estimated based on the 2 DOF(Degree Of Freedom) vehicle single-track model. On this basis, an autonomous parking positioning and location tracking algorithm was designed with extended Kalman filter method and the detection results of the closed-loop camera, which can be used in the underground garage environment with the help of low cost sensors, to estimate the vehicle pose in real time and optimize the detection results of the corner points of the closed-loop camera. Simulation and experimental results show that the proposed positioning and garage location tracking algorithm can effectively reduce vehicle motion data noise, provide more accurate vehicle garage location relative position, and provide higher accuracy positioning results for subsequent planning and control.
引文
[1]左培文,孟庆阔,李育贤.自动泊车系统发展现状及前景分析[J].上海汽车,2017(2):44-46.
[2]熊璐,严森炜,余卓平,等.基于库位跟踪的自动泊车决策规划系统[J].汽车技术,2018(8):18-26.
[3]李文新,张正祥.一种基于环视相机的自动泊车方法[J].工业C,2016(11):270.
[4]王旱霖,赵翔,王春香,等.一种基于环视图像的车辆开门碰撞预警方法[J].华中科技大学学报(自然科学版),2015(增刊1):113-116.
[5]秦永元,张洪钺,王叔华.卡尔曼滤波与组合导航原理:第2版[M].西安:西北工业大学出版社,2012.
[6]高博麟,陈慧,陈威,等.汽车质心侧偏角融合估计方法[J].汽车工程,2013,35(8):716-722.
[7]Jiang Y,Chen H,Xiong G,et al.Kinematic Constraints in Visual Odometry of Intelligent Vehicles[C]//Intelligent Vehicles Symposium.IEEE,2012:1126-1131.
[8]Li L,Zhang L,Li X,et al.Vision-Based Parking-Slot Detection:A Benchmark and a Learning-Based Approach[C]//IEEE International Conference on Multimedia and Expo.IEEE,2017:649-654.
[9]Thrun S.Probabilistic Robotics[J].Communications of the ACM,2005,45(3):52-57.