基于支持向量机的局部路径规划算法
|
摘要
针对在非结构化道路环境下地面自主车辆的路径规划问题,本文提出了一种基于64激光雷达以及支持向量机的局部路径规划算法。利用非线性支持向量机分类器在栅格地图上提取出安全的路径;并在多帧投影数据上使用RANSAC算法提取出路径,并使用三次多项式进行描述,从而计算出道路曲率;结合地面自主车辆自身状态在RANSAC路径上选取控制点,并使用贝塞尔曲线拟合出最终路径。该算法能够有效地从局部栅格地图中提取道路,以弥补基于视觉的道路检测算法在受到恶劣光照、天气影响时的不足。通过实车试验验证了所提出的方法的有效性和正确性。
To solve the path-planning problem of unmanned ground vehicles( UGV) in an unstructured road environment,a local path-planning method based on multiple Lidar and support vector machine( SVM) is proposed in this paper. First,a safe path is extracted from the grid map by nonlinear SVM. A path described with a cubic polynomial can be extracted by using RANSAC algorithm on the multi-frame projection data and the path can be used to calculate the road curvature. Then,the control points are selected by considering the state of the UGV and the final path is generated by using Bezier curve fitting. The proposed method can effectively extract the path from the grid map when the performance of the vision-based road detection algorithms is limited by poor lighting and weather. Finally,the validity and correctness of the proposed algorithm is verified on a real vehicle.
To solve the path-planning problem of unmanned ground vehicles( UGV) in an unstructured road environment,a local path-planning method based on multiple Lidar and support vector machine( SVM) is proposed in this paper. First,a safe path is extracted from the grid map by nonlinear SVM. A path described with a cubic polynomial can be extracted by using RANSAC algorithm on the multi-frame projection data and the path can be used to calculate the road curvature. Then,the control points are selected by considering the state of the UGV and the final path is generated by using Bezier curve fitting. The proposed method can effectively extract the path from the grid map when the performance of the vision-based road detection algorithms is limited by poor lighting and weather. Finally,the validity and correctness of the proposed algorithm is verified on a real vehicle.
引文
[1]李青,郑南宁,马琳,等.基于主元神经网络的非结构化道路跟踪[J].机器人,2005,27(3):247-251.LI Qing,ZHENG Nanning,MA Lin,et al. Tracking of unstructured road based on principal component analysis neural networks[J]. Robot,2005,27(3):247-251.
[2]朱株.基于三维数据面向无人车导航的非结构化场景理解[D].杭州:浙江大学,2014.ZHU Zhu. 3D data based understanding of unstructured scenes for vehicle navigation[D]. Hangzhou:Zhejiang University,2014.
[3]GERAERTS R. Planning short paths with clearance using explicit corridors[C]//Proceedings of 2010 IEEE International Conference on Robotics and Automation. Anchorage,USA,2010:1997-2004.
[4]BHATTACHARYA P,GAVRILOVA M L. Roadmapbased path planning-using the voronoi diagram for a clearance-based shortest path[J]. IEEE robotics&automation magazine,2008,15(2):58-66.
[5]RICHTER C,WARE J,ROY N. High-speed autonomous navigation of unknown environments using learned probabilities of collision[C]//Proceedings of 2014 IEEE International Conference on Robotics and Automation. Hong Kong,China,2014:6114-6121.
[6]YU Chong,QIU Qiwen,CHEN Xiong. A hybrid two-dimensional path planning model based on frothing construction algorithm and local fast marching method[J]. Computers&electrical engineering,2013,39(2):475-487.
[7]GARRIDO S,MORENO L,GMEZ J V. Motion planning using fast marching squared method[M]//CARBONE G,GOMEZ-BRAVO F. Motion and Operation Planning of Robotic Systems:Background and Practical Approaches.Cham:Springer,2015:223-248.
[8]CHEN Chao,RICKERT M,KNOLL A. Kinodynamic motion planning with space-time exploration guided heuristic search for car-like robots in dynamic environments[C]//Proceedings of 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems. Hamburg, Germany,2015:2666-2671.
[9]刘梓,唐振民,任明武.基于3D激光雷达的实时道路边界检测算法[J].华中科技大学学报(自然科学版),2011,39(S2):351-354.LIU Zi,TANG Zhenmin,REN Mingwu. Algorithm of real-time road boundary detection based on 3D lidar[J].Journal of Huazhong University of Science&Technology(natural science edition),2011,39(S2):351-354.
[10]BAGCHI S. Rapid close surrounding evaluation for autonomous commercial vehicles[D]. G9teborg:Chalmers University of Technology,2016.
[11]MARATHE A S,VYAS V,CHAVHAN M. Petrographic image classification using optimized radial basis function support vector machine&validation of its asymptotic behavior[C]//Proceedings of 2016 IEEE International Conference on Signal Processing,Communications and Computing. Hong Kong,China,2016:1-6.
[12]FONTANELLI D, CAPPELLETTI M, MACII D. A RANSAC-based fast road line detection algorithm for high-speed wheeled vehicles[C]//Proceedings of 2011IEEE International Instrumentation and Measurement Technology Conference. Binjiang,China,2011:1-6.
[13]SMADJA L,NINOT J,GAVRILOVIE T. Road extraction and environment interpretation from LiDAR sensors[J].International archives of photogrammetry and remote sensing,2010,38:281-286.
[14]LI Yingmao,GANS N R. Predictive-RANSAC:an effective data fitting and tracking method with application in curve tracking in video[C]//Proceedings of 2016 American Control Conference. Boston,USA,2016:3620-3625.
[15]LI Wen,LI Yan,MA Yide. An efficient contour tracking and coding algorithm based on vertex chain code[C]//Proceedings of the 2012 8th International Conference on Intelligent Information Hiding and Multimedia Signal Processing. Piraeus,Greece,2012:514-517.
[2]朱株.基于三维数据面向无人车导航的非结构化场景理解[D].杭州:浙江大学,2014.ZHU Zhu. 3D data based understanding of unstructured scenes for vehicle navigation[D]. Hangzhou:Zhejiang University,2014.
[3]GERAERTS R. Planning short paths with clearance using explicit corridors[C]//Proceedings of 2010 IEEE International Conference on Robotics and Automation. Anchorage,USA,2010:1997-2004.
[4]BHATTACHARYA P,GAVRILOVA M L. Roadmapbased path planning-using the voronoi diagram for a clearance-based shortest path[J]. IEEE robotics&automation magazine,2008,15(2):58-66.
[5]RICHTER C,WARE J,ROY N. High-speed autonomous navigation of unknown environments using learned probabilities of collision[C]//Proceedings of 2014 IEEE International Conference on Robotics and Automation. Hong Kong,China,2014:6114-6121.
[6]YU Chong,QIU Qiwen,CHEN Xiong. A hybrid two-dimensional path planning model based on frothing construction algorithm and local fast marching method[J]. Computers&electrical engineering,2013,39(2):475-487.
[7]GARRIDO S,MORENO L,GMEZ J V. Motion planning using fast marching squared method[M]//CARBONE G,GOMEZ-BRAVO F. Motion and Operation Planning of Robotic Systems:Background and Practical Approaches.Cham:Springer,2015:223-248.
[8]CHEN Chao,RICKERT M,KNOLL A. Kinodynamic motion planning with space-time exploration guided heuristic search for car-like robots in dynamic environments[C]//Proceedings of 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems. Hamburg, Germany,2015:2666-2671.
[9]刘梓,唐振民,任明武.基于3D激光雷达的实时道路边界检测算法[J].华中科技大学学报(自然科学版),2011,39(S2):351-354.LIU Zi,TANG Zhenmin,REN Mingwu. Algorithm of real-time road boundary detection based on 3D lidar[J].Journal of Huazhong University of Science&Technology(natural science edition),2011,39(S2):351-354.
[10]BAGCHI S. Rapid close surrounding evaluation for autonomous commercial vehicles[D]. G9teborg:Chalmers University of Technology,2016.
[11]MARATHE A S,VYAS V,CHAVHAN M. Petrographic image classification using optimized radial basis function support vector machine&validation of its asymptotic behavior[C]//Proceedings of 2016 IEEE International Conference on Signal Processing,Communications and Computing. Hong Kong,China,2016:1-6.
[12]FONTANELLI D, CAPPELLETTI M, MACII D. A RANSAC-based fast road line detection algorithm for high-speed wheeled vehicles[C]//Proceedings of 2011IEEE International Instrumentation and Measurement Technology Conference. Binjiang,China,2011:1-6.
[13]SMADJA L,NINOT J,GAVRILOVIE T. Road extraction and environment interpretation from LiDAR sensors[J].International archives of photogrammetry and remote sensing,2010,38:281-286.
[14]LI Yingmao,GANS N R. Predictive-RANSAC:an effective data fitting and tracking method with application in curve tracking in video[C]//Proceedings of 2016 American Control Conference. Boston,USA,2016:3620-3625.
[15]LI Wen,LI Yan,MA Yide. An efficient contour tracking and coding algorithm based on vertex chain code[C]//Proceedings of the 2012 8th International Conference on Intelligent Information Hiding and Multimedia Signal Processing. Piraeus,Greece,2012:514-517.