To address the safety hazards in driving blind spot scenarios, a collaborative obstacle avoidance planning algorithm for vehicles based on a collision risk field model was proposed. Considering the impact of vehicle-to-vehicle communication continuity, a communication continuity risk field model is constructed. This risk field was incorporated into the dynamic path planning algorithm, and an adaptive obstacle avoidance safety cost function based on the communication continuity impact factor was designed. Simulation results demonstrate that the proposed risk model-based obstacle avoidance planning algorithm can accurately reflect dynamic risks in driving blind spots, enhancing the safety of autonomous vehicles in extreme situations.
随着路边云技术的发展,智能交通系统的潜力得到了显著提升[1]。然而,在实际应用中,尤其是在偏远地区或车联万物(Vehicle to everything,V2X)基础设施尚未普及的地方,自动驾驶车辆在障碍物盲区场景下进行安全避障仍面临着挑战[2]。这种挑战体现在自动驾驶系统难以量化评估由偶然不连续的信息造成的碰撞潜在风险并进行安全有效的轨迹规划。
现有的避障规划算法是基于实时传感器数据,从避障起点到终点之间规划出一条最优路径。如基于A*算法的路径优化算法[9]、虚拟力场与模型预测理论的结合算法[10],以及结合快速行进树算法(Fast marching tree algorithm,FMT)的动态规划算法[11]和区域采样随机树算法[12](Region sampling rapidly-exploring random tree,RS-RRT),均在提高规划效率和稳定性方面取得了进展。然而,这些方法虽然在动态环境中可以实现避障和路径调整,以确保车辆可以应对简单突发情况,但大多采用简单的避障模型,同时对数据完整度的依赖过高,没有考虑V2V信息来源不稳定的情况,忽略由此造成的盲区潜在碰撞风险,在驾驶盲区场景下的表现较差。
综上,在驾驶盲区场景下,V2V协同中的非连续时间段信息对碰撞风险场模型及规划模块的影响尚均未得到充分考虑,亟需建立三者间关系。针对上述问题,本文提出一种基于V2V通信持续性风险模型的避障规划算法(Risk model dynamic programming,RMDP),首先,将V2V探测传递的障碍物坐标信息转变为全局坐标,以便用于碰撞风险场模型的构建和路径规划的精准执行。然后,基于等效力思想[13]构建考虑车辆运动状态和道路条件的碰撞风险模型以定量描述驾驶盲区场景下的碰撞风险。考虑到V2V传递信息的非连续性,设计通信持续性影响因子表示障碍物时空信息缺失的潜在风险,并将其以代价函数形式整合纳入避障规划算法中,从而提升自动驾驶系统的危险态势感知敏感度和规划效率。最后,通过联合仿真验证了本文方法的安全性和可靠性。
LiH M, DongX, RasheedI, et al. A software-defined networking roadside unit cloud resource management framework for vehicle ad hoc networks[J]. Journal of Advanced Transportation, 2022, 2022: 1-13.
[2]
LuJ Y, PengZ X, ShiR W, et al. A quantitative blind area risks assessment method for safe driving assistance[J]. Journal of Systems Architecture, 2024, 150: 1-15.
[3]
WangL H, ZhangF Q, CuiY H, et al. Stochastic velocity prediction for connected vehicles considering V2V communication interruption[J]. IEEE Transactions on Intelligent Transportation Systems, 2023, 24(11): 11654-11667.
[4]
JoJ H, ShimJ N, KimB, et al. AoA-based position and orientation estimation using lens MIMO in cooperative vehicle-to-vehicle systems[J]. IEEE Journal on Selected Areas in Communications, 2023, 41(12): 3719-3735.
ZhaoRui, LiYun, HuHong-yu, et al. Vehicle collision warning method at intersection based on V2I communication[J]. Journal of Jilin University(Engineering and Technology Edition), 2023, 53(4): 1019-1029.
ZhengXun-jia, JiangJun-hao, HuangHe-ye, et al. Novel quantitative approach for assessing driving risks and simulation study of its prevention and control strategies[J]. Journal of Mechanical Engineering, 2024, 60(10): 207-221.
[11]
LeeM, JoK, SunwooM. Collision risk assessment for possible collision vehicle in occluded area based on precise map[C]∥IEEE International Conference on Intelligent Transportation Systems, Yokohama, Japan, 2017: 1-6.
FengShuo, JiXian-you, ChengBo, et al. Path planning algorithm based on dynamic obstacle movement information[J]. Computer Engineering and Applications, 2022, 58(21): 279-285.
RenYue, ZhengLing, ZhangWei, et al. A study on active collision avoidance control of autonomous vehicles based on model predictive control[J]. Automotive Engineering, 2019, 41(4): 404-410.
[16]
SzczepanskiR. Safe artificial potential field-novel local path planning algorithm maintaining safe distance from obstacles[J]. IEEE Robotics and Automation Letters, 2023, 8(8): 4823-4830.
HanXiao-jian, ZhaoWei-qiang, ChenLi-jun, et al. Local path planning of bus based on RS-RRT algorithm[J]. Journal of Jilin University(Engineering and Technology Edition), 2019, 49(5): 1428-1440.
[19]
ZhengXun-jia, ZhangDi, GaoHong-bo, et al. A novel framework for road traffic risk assessment with HMM-based prediction model[J]. Sensors, 2018, 18(12): 4313-4327.
[20]
LiuFang-ming, ChenJian, ZhangQi-xia, et al. Online MEC offloading for V2V networks[J]. IEEE Transactions on Mobile Computing, 2023, 22(10): 6097-6109.
[21]
MengZong-lin, XiaXin, XuRun-sheng, et al. HYDRO-3D: hybrid object detection and tracking for cooperative perception using 3D LiDAR[J]. IEEE Transactions on Intelligent Vehicles, 2023, 8(8): 4069-4080.
[22]
KeX C, HoK C. Localization by doppler derivatives and doppler-shifted frequencies[J].EEE Transactions on Signal Processing, 2024, 72: 2890-2904.
[23]
FomeniF D. A lifted-space dynamic programming algorithm for the quadratic knapsack problem[J]. Discrete Applied Mathematics, 2023, 335: 52-68.
[24]
ZhangLin, WuDa-de, ZhaoMing-jing, et al. Uncertainty relation and the constrained quadratic programming[J]. Physica Scripta, 2024, 99(6): 1-35.
[25]
ReiterR, NurkanovicA, FreyJ, et al. Frenet-Cartesian model representations for automotive obstacle avoidance within nonlinear MPC[J]. European Journal of Control, 2023, 74: 1-8.
[26]
YoonS, KwonY, RyuJ, et al. Reinforcement-learning-based trajectory learning in frenet frame for autonomous driving[J]. Applied Sciences-Basel, 2024, 14(16):1-14.
[27]
ZhangYi-xiao, HaoRui, ZhangTing-ting, et al. A trajectory optimization-based intersection coordination framework for cooperative autonomous vehicles[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(9): 14674-14688.