Assistant driving decision method of vehicle lane change based on dynamic probability grid and Bayesian decision network
-
摘要: 提出了一种车辆变道辅助决策方法, 向驾驶人提供变道行为决策; 构建了由车载GPS、相机传感器和雷达组成的行车环境信息传感装置, 利用非采样B样条曲线模型对车道线建模, 通过控制点位置求解与搜索策略实现车道线的检测、跟踪与类型识别; 根据车道线信息确立有效行车区域, 并建立了一种动态概率网格的行车环境几何模型, 对有效行车区域进行紧凑型表征; 考虑了车辆对行车环境表征结果可靠性的影响, 根据高斯分布将车辆位置信息映射到动态概率网格中, 计算了每个行车单元的占用概率; 将车道线信息与网格单元占用概率作为初始节点状态参数, 输入贝叶斯决策网络, 估计概率网格单元的占用状态, 量化输出当前行车环境表征结果以及不变道、向左变道、向右变道3种变道决策的期望效用值, 通过计算各决策的期望效用值比率确定最优变道决策。试验结果表明: 在场景1中“向左变道”决策的期望效用值最大, 为0.70, 视为最优决策, 在其动态概率网格中, 右侧车道线“实线”状态参数为100.00%, 因此, “向右变道”决策效用期望值最小, 决策系统输出的最优决策“不变道”符合中国交通法规, 也表明检测车道线类型的必要性; 场景2的“不变道”和“向右变道”决策期望效用值分别为0.43和0.44, 比率接近1, 无法判断最优决策, 驾驶人可根据经验决定是否变道。Abstract: An assistant decision method of vehicle lane change was proposed to provide lane changing behavior decision for drivers.A driving environment information perception device consisting of on-board GPS, camera sensors, and radars was constructed.A non-uniform B-spline (NUBS) curve model was applied to modeling lane line.The lane line was detected, tracked and classified by coordinate calculation and searching strategy of control points.The effective driving region was determined based on lane line information.A driving environment geometric model of dynamic probability grid was established and considered as a compact representation of effective driving region.The influence of vehicle on the reliability of driving environment representation result was considered.Vehicle coordination information was mapped into the dynamic probabilitygrid based on Gaussian distribution, and the occupancy probability of each cell unit was calculated.Lane line information and occupied probability of cell unit were used as initial node state parameters and input into Bayesian decision network to estimate the occupancy state of probability grid unit, output the quantified representation result of current driving environment, and compute the expect utility values of three lane changing decisions including "no lane changing", "left lane changing"and "right lane changing".The optimal lane changing decision was determined by computing the ratio of each decision-making expect utility value.Experimental result shows that, in situation 1, the maximum expect utility value is 0.70 of "left lane changing", which is the optimal decision.In the corresponding dynamic probability grid, the"solid"state parameter of right lane line is 100.00%, so the expect utility of "right lane changing"is the lowest.The output optimal decision of decision-making system is "no lane changing", which is consistent with Chinese traffic laws and verifies the necessity of lane line classification.In situation 2, the decision expect utility values of"no lane changing"and "right lane changing"are 0.43 and 0.44, respectively, the ratio is close to 1, so it is unable to judge the optimal decision, and drivers can make the decision whether or not to change lane based on experience.
-
图 2 无遮挡车道线检测分类结果
Figure 2. Detecting and sorting result of lane lines without occlusions
图 4 直车道动态概率网格几何模型
Figure 4. Geometric model of dynamic probability grid of straight lane
图 9 采用Bresenham算法确定本车安全距离内的单元
Figure 9. Cells in safe distance of ego-vehicle determined by Bresenham algorithm
图 11 场景1行车环境动态概率网格表征结果
Figure 11. Dynamic probability grid description result of driving environment in scene 1
图 14 场景2行车环境动态概率网格表征结果
Figure 14. Dynamic probability grid description result of driving environment in scene 2
-
[1] TAWARI A, SIVARAMAN S, TRIVEDI M M, et al. Looking-in and looking-out vision for urban intelligent assistance: estimation of driver attentive state and dynamic surround for safe merging and braking[C]//IEEE. 25th IEEE Intelligent Vehicles Symposium. New York: IEEE, 2014: 115-120. [2] CONCEIOL, ROSSETTI R J F. Multivariate modelling for autonomous vehicles: research trends in perspective[C]//IEEE. 19th International Conference on Intelligent Transportation Systems. New York: IEEE, 2016: 83-87. [3] MUNIR A. Safety assessment and design of dependable cybercars: for today and the future[J]. IEEE Consumer Electronics Magazine, 2017, 6 (2): 69-77. doi: 10.1109/MCE.2016.2640738 [4] MUNIR A, KOUSHANFAR F. Design and performance analysis of secure and dependable cybercars: a steer-by-wire case study[C]//IEEE. 13th IEEE Annual Consumer Communications and Networking Conference. New York: IEEE, 2016: 1066-1073. [5] ZHU Gang, YANG Ming, LI Hao, et al. Curvature mapbased magnetic guidance for automated vehicles in an urban environment[J]. IEEE Transactions on Intelligent Transportation Systems, 2016, 17 (12): 3541-3552. doi: 10.1109/TITS.2016.2557066 [6] WANG Fei-yue, YANG Liu-qing, YANG Jian, et al. Urban intelligent parking system based on the parallel theory[C]//IEEE. 2006International Conference on Computing, Networking and Communications. New York: IEEE, 2016: 1-5. [7] 穆柯楠. 基于车-路视觉协同的行车环境感知方法研究[D]. 西安: 长安大学, 2016.MU Ke-nan. Research on the driving environment perception method based on visual cooperative vehicle-infrastructure system[D]. Xi'an: Chang'an University, 2016. (in Chinese). [8] 康俊民, 赵祥模, 徐志刚. 无人车行驶环境特征分类方法[J]. 交通运输工程学报, 2016, 16 (6): 140-148. http://transport.chd.edu.cn/article/id/201606017KANG Jun-min, ZHAO Xiang-mo, XU Zhi-gang. Classification method of running environment features for unmanned vehicle[J]. Journal of Traffic and Transportation Engineering, 2016, 16 (6): 140-148. (in Chinese). http://transport.chd.edu.cn/article/id/201606017 [9] SUHR J K, JUNG H G. Sensor fusion-based vacant parking slot detection and tracking[J]. IEEE Transactions on Intelligent Transportation Systems, 2014, 15 (1): 21-36. doi: 10.1109/TITS.2013.2272100 [10] MOSQUET X, ANDERSEN M, ARORA A. A roadmap to safer driving through advanced driver assistance systems[R]. Washington DC: MEMA, 2015. [11] VELEZ G, OTAEGUI O. Embedding vision-based advanced driver assistance systems: a survey[J]. IET Intelligent Transport Systems, 2017, 11 (3): 103-112. doi: 10.1049/iet-its.2016.0026 [12] SIVARAMAN S, TRIVEDI M M. Dynamic probabilistic drivability maps for lane change and merge driver assistance[J]. IEEE Transactions on Intelligent Transportation Systems, 2014, 15 (5): 2063-2073. doi: 10.1109/TITS.2014.2309055 [13] SCHUBERT R, SCHULZE K, WANIELIK G. Situation assessment for automatic lane-change maneuvers[J]. IEEE Transactions onIntelligent Transportation Systems, 2010, 11 (3): 607-616. doi: 10.1109/TITS.2010.2049353 [14] NGUYEN T N, MICHAELIS B, AL-HAMADI A, et al. Stereo-camera-based urban environment perception using occupancy grid and object tracking[J]. IEEE Transactions on Intelligent Transportation Systems, 2012, 13 (1): 154-165. doi: 10.1109/TITS.2011.2165705 [15] PABLO M, JORGE B, AHMED H, et al. Stereo visionbased local occupancy grid map for autonomous navigation in ROS[C]//VISAPP. 11th International Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. Setúbal: VISAPP, 2016: 703-708. [16] OH S I, KANG Hang-bong. Fast occupancy grid filtering using grid cell clusters from LIDAR and stereo vision sensor data[J]. IEEE Sensors Journal, 2016, 16 (19): 7258-7266. doi: 10.1109/JSEN.2016.2598600 [17] NADARAJAN P, BOTSCH M, SARDINA S. Predictedoccupancy grids for vehicle safety applications based on autoencoders and the Random Forest algorithm[C]//IEEE. 2017International Joint Conference on Neural Networks. New York: IEEE, 2017: 1244-1251. [18] NADARAJAN P, BOTSCH M. Probability estimation for predicted-occupancy grids in vehicle safety applications based on machine learning[C]//IEEE. 2016Intelligent Vehicles Symposium. New York: IEEE, 2016: 1285-1292. [19] DANESCU R, ONIGA F, NEDEVSCHI S. Modeling and tracking the driving environment with a particle-based occupancy grid[J]. IEEE Transactions on Intelligent Transportation Systems, 2011, 12 (4): 1331-1342. doi: 10.1109/TITS.2011.2158097 [20] ONIGA F, NEDEVSCHI S. Processing dense stereo data using elevation maps: road surface, traffic isle, and obstacle detection[J]. IEEE Transportations on Vehicle Technology, 2010, 59 (3): 1172-1182. doi: 10.1109/TVT.2009.2039718 [21] OTTO C, LEON F P. Long-term trajectory classification and prediction of commercial vehicles for the application in advanced driver assistance systems[C]//IEEE. 2012 American Control Conference. New York: IEEE, 2012: 2904-2909. [22] GE Y E, XU C F, SZETO W Y, et al. Investigating freeway traffic hypercongestion between an on-ramp and its immediate upstream off-ramp[J]. Transportmetrica A: Transport Science, 2015, 11 (3): 187-209. doi: 10.1080/23249935.2014.945509 [23] SCHREIER M, WILLERT V, ADAMY J. Bayesian, maneuverbased, long-term trajectory prediction and criticality assessment for driver assistance systems[C]//IEEE. 17th IEEE International Conference on Intelligent Transportation Systems. New York: IEEE, 2014: 334-341. [24] TRUONG Q B, LEE B R, HEO N G, et al. Lane boundaries detection algorithm using vector lane concept[C]//IEEE. 10th International Conference on Control, Automation, Robotics and Vision. New York: IEEE, 2009: 2319-2325. [25] SHANMUGASUNDAR G, SIVARAMAKRISHNAN R, BALASUBRAMANI S. Method of trajectory generation of a generic robot using Bresenham's circle algorithm[J]. Indian Journal of Science and Technology, 2016, 9 (48): 1-6. -
下载:
点击查看大图
图(15) / 表(3)
计量
- 文章访问数: 627
- HTML全文浏览量: 102
- PDF下载量: 933
- 被引次数: 0
