技术领域Technical Field
本发明公开基于车辆风险场和均衡博弈的道路通行控制系统和方法，属于交通控制技术领域。The invention discloses a road traffic control system and method based on a vehicle risk field and equilibrium game, belonging to the technical field of traffic control.
背景技术Background Art
道路交汇区的通行控制一直以来都是交通控制技术的重中之重，现如今，随着自动驾驶技术的发展，网联自主车辆（Connected Autonomous Vehicle，CAV）车辆迅速普及，道路的车辆组成逐渐由单一的人工驾驶车辆组成的传统模式，向着更加复杂更多样的混合车流模式转变，形成了一种新型的交通模式。然而，相比于智能网联交通的发展迅速，先进交通控制技术的更迭却较为落后，绝大多数的信号调控系统仍采用以信号灯为路侧主体，人工调节通行相位为调控方案的传统模式，这种传统模式在过去的交通系统中发挥着重要的作用，然随着网联自主车辆的发展，这种传统的交通控制模式已逐步落后于现有的交通模式，乃至对新型交通模式的发展有所制约。基于此，本发明提出了基于车辆风险场和均衡博弈的道路通行控制系统和方法，路侧主体采用可升降的调控信号设备，在充分利用地下空间的同时，可以保证驾驶员视野的连续性，保障道路安全。控制方法采用基于车辆风险场及均衡博弈的综合理论体系，将风险势函数及博弈收益函数有机融入设备激活函数中，做到因时制宜，因地制宜，灵活控制通行方案，提高道路通行效率。Traffic control at road intersections has always been a top priority in traffic control technology. Nowadays, with the development of autonomous driving technology, connected autonomous vehicles (CAV) are rapidly becoming popular, and the composition of vehicles on the road has gradually changed from the traditional mode of a single manually driven vehicle to a more complex and diverse mixed traffic mode, forming a new traffic mode. However, compared with the rapid development of intelligent connected transportation, the replacement of advanced traffic control technology is relatively backward. Most signal control systems still use the traditional mode of using signal lights as the main body of the roadside and manually adjusting the traffic phase as the control scheme. This traditional mode played an important role in the past traffic system. However, with the development of connected autonomous vehicles, this traditional traffic control mode has gradually fallen behind the existing traffic mode, and even restricted the development of new traffic modes. Based on this, the present invention proposes a road traffic control system and method based on vehicle risk field and equilibrium game. The main body of the roadside adopts a control signal device that can be raised and lowered. While making full use of the underground space, it can ensure the continuity of the driver's field of vision and ensure road safety. The control method adopts a comprehensive theoretical system based on vehicle risk field and equilibrium game, organically integrating the risk potential function and game profit function into the equipment activation function, so as to adapt to the time and local conditions, flexibly control the traffic plan and improve road traffic efficiency.
发明内容Summary of the invention
本发明的目的在于提供基于车辆风险场和均衡博弈的道路通行控制系统和方法，以解决现有技术中，合流路段的通行权冲突问题的问题。The purpose of the present invention is to provide a road traffic control system and method based on vehicle risk field and equilibrium game to solve the problem of right of way conflict in merging sections in the prior art.
基于车辆风险场和均衡博弈的道路通行控制系统，包括视云融合检测模块、系统主体硬件模块、系统控制算法模块、系统交互通信模块；The road traffic control system based on vehicle risk field and equilibrium game includes a visual cloud fusion detection module, a system main hardware module, a system control algorithm module, and a system interactive communication module;
视云融合检测模块采用视云融合检测方法获取交通状况信息，采用激光扫描仪获取车辆行人的点云信息，采用雷视一体机获取周遭交通环境的视频信息与雷达数据；The visual-cloud fusion detection module uses the visual-cloud fusion detection method to obtain traffic condition information, uses a laser scanner to obtain point cloud information of vehicles and pedestrians, and uses a radar-visual integrated device to obtain video information and radar data of the surrounding traffic environment;
系统主体硬件模块采用内镶嵌式构造，系统外壳采用碳素钢，内衬PVC层与橡胶层作为缓冲层，内置螺旋电机作为主体骨架，采用stm32单片机作为控制单元，设定上下限位开关控制上升幅度，系统内部空间搭载两架雷视一体机以及一个激光扫描仪；The main hardware module of the system adopts an embedded structure. The system shell is made of carbon steel, lined with PVC layer and rubber layer as a buffer layer, with a built-in spiral motor as the main frame, and an STM32 single-chip microcomputer as the control unit. The upper and lower limit switches are set to control the rise range. The internal space of the system is equipped with two integrated laser vision machines and a laser scanner.
系统控制算法模块采用风险势博弈方法，基于分子力场理论方法建立交通系统行为风险场，非线性加权组合风险势场，对组合风险场沿道路方向进行纵横二维划分，确定道路风险势，在风险势的势井、势峰位置进行通行权变换达到优化道路通行的目的；The system control algorithm module adopts the risk potential game method, establishes the traffic system behavior risk field based on the molecular force field theory method, nonlinearly weighted combined risk potential field, divides the combined risk field into two dimensions along the road direction, determines the road risk potential, and transforms the right of way at the potential well and potential peak of the risk potential to achieve the purpose of optimizing road traffic;
系统交互通信模块利用GNSS-IMU融合定位方法对周围车辆准确定位，获取车辆自身的运动状态信息，计算对位车辆的实时状态，分析车辆运行状态，构建实时的虚拟路况地图，搭建车车、车人、车与环境之间的信息交互桥梁，达到优化道路通行权的效果。The system interactive communication module uses the GNSS-IMU fusion positioning method to accurately locate surrounding vehicles, obtain the vehicle's own motion status information, calculate the real-time status of the corresponding vehicles, analyze the vehicle's operating status, construct a real-time virtual road condition map, and build an information interaction bridge between vehicles, vehicles and people, and vehicles and the environment, so as to achieve the effect of optimizing road access rights.
基于车辆风险场和均衡博弈的道路通行控制方法，使用所述的基于车辆风险场和均衡博弈的道路通行控制系统，包括：A road traffic control method based on a vehicle risk field and equilibrium game, using the road traffic control system based on a vehicle risk field and equilibrium game, comprises:
S1建立由车辆风险势场、道路边界风险势场、道路车道线势场、非对称行人势场和道路障碍物风险势场组成的叠加风险势场，车辆在场中运行的结果是各类风险势场的相互作用力叠加的结果，建立笛卡尔坐标系，对势场进行纵横二维分析；S1 establishes a risk potential field composed of vehicles , Road boundary risk potential field 、Road lane line potential field , asymmetric pedestrian potential field and road obstacle risk potential field The result of the vehicle running in the field is the result of the superposition of the interaction forces of various risk potential fields. A Cartesian coordinate system is established to conduct a two-dimensional analysis of the potential field.
由、、、、组成叠加风险势场：Depend on , , , , Composition of superimposed risk potential field :
； ;
S2在叠加风险势场和速度风险势场基础上，对多目标风险势进行纵横二维量化；S2 quantifies the multi-target risk potential in two dimensions, vertically and horizontally, based on the superposition risk potential field and the velocity risk potential field;
S3基于S1和S2的结果，获得基于道路风险场非线性通行权激活函数，进行道路通行控制。Based on the results of S1 and S2, S3 obtains the nonlinear right-of-way activation function based on the road risk field and performs road traffic control.
S1包括，车辆风险势场作用势来自于车车交互作用力，基于兰纳-琼斯势建立一个统一的车辆作用势场函数模型，引入车辆需求风险距离以及跟驰状态，构建车辆风险势场为：S1 includes that the vehicle risk potential field potential comes from the vehicle-to-vehicle interaction force. A unified vehicle potential field function model is established based on the Lanner-Jones potential. The vehicle required risk distance and following state are introduced to construct the vehicle risk potential field as follows:
； ;
式中，为势场能量尺度，为斥力项幂次，为引力项幂次，为车辆需求的风险距离，是的导数，为车辆的实际间距，为围绕车辆任意质点与车辆质心点之间的顺时针夹角，为修正短程斥力势场分布范围。In the formula, is the energy scale of the potential field, is the power of the repulsive force, is the gravitational power, is the risk distance required by the vehicle, yes The derivative of is the actual distance between vehicles, is the clockwise angle between any mass point around the vehicle and the center of mass of the vehicle, To correct the distribution range of the short-range repulsive potential field.
S1包括，将道路下侧道路边界线作为轴原点，采用高斯类函数对道路车道线势场进行表示：S1 includes taking the road boundary line at the lower side of the road as Axis origin, using Gaussian function to represent the road lane line potential field:
； ;
式中，为道路横截面车道线数量，为任意一点纵坐标，为不同车道线势场的势场强度系数，为道路线类型，为1时为白色虚线，为2时为黄色双黄线，为第条道路线的纵坐标位置，与道路宽度成正比，表示道路边界风险势场的场强增减速度。In the formula, is the number of lane lines in the road cross section, is the ordinate of any point, is the potential field intensity coefficient of different lane line potential fields, is the road line type, When it is 1, it is a white dotted line. When it is 2, it is a double yellow line. For the The vertical coordinate position of the road line, It is proportional to the road width and indicates the rate of increase or decrease of the field strength of the road boundary risk potential field.
S1包括，道路边界风险势场为：S1 includes the road boundary risk potential field:
；； ; ;
式中，为道路两边道路边界线，设定为道路右边界线，为道路左边界线，为点处车辆与点处道路边界线处纵坐标差值，为车辆质心所处位置，为道路边界线所处位置，为位置增益函数。In the formula, Set the road boundary lines on both sides of the road The right boundary line of the road. The left boundary line of the road. for Vehicles at point The vertical coordinate difference of the road boundary line at the point, is the position of the vehicle's center of mass, is the location of the road boundary line, is the position gain function.
S1包括，构建道路障碍物风险势场为：S1 includes constructing the road obstacle risk potential field as follows:
； ;
式中，为道路影响因子风险势场系数，、为车辆质心的横纵坐标，、为道路影响因子质心的横纵坐标，、、、为风险势场的形状系数，、由影响因子的形状尺寸决定。In the formula, is the risk potential field coefficient of the road influencing factor, , is the horizontal and vertical coordinates of the vehicle's center of mass, , is the horizontal and vertical coordinates of the centroid of the road influencing factor, , , , is the shape coefficient of the risk potential field, , Determined by the shape and size of the influencing factor.
S1包括，非对称行人势场包括非对称横向行人势场和非对称纵向行人势场；S1 includes, the asymmetric pedestrian potential field includes an asymmetric transverse pedestrian potential field and an asymmetric longitudinal pedestrian potential field;
非对称横向行人势场为：The asymmetric lateral pedestrian potential field is:
； ;
； ;
式中，为非对称横向行人势场，、为风险势场比例调节系数，、为车辆横纵坐标与行人横纵坐标的差值，为车辆横向判断参数，差值为负时，车辆在行人右边，，差值为正时，车辆在行人左边，，为车辆需要最短纵向避障距离，为包含驾驶员反应时间的避障最短时间，表示车辆需要最短横向避障距离，等于车辆宽度、行人所需风险空间半径及风险阈度之和；In the formula, is the asymmetric lateral pedestrian potential field, , is the risk potential field proportional adjustment coefficient, , is the difference between the horizontal and vertical coordinates of the vehicle and the horizontal and vertical coordinates of the pedestrian, is the vehicle lateral judgment parameter. When the difference is negative, the vehicle is on the right side of the pedestrian. , when the difference is positive, the vehicle is on the left side of the pedestrian, , The vehicle needs the shortest longitudinal obstacle avoidance distance. is the minimum obstacle avoidance time including the driver’s reaction time, Indicates the shortest lateral obstacle avoidance distance required by the vehicle, which is equal to the sum of the vehicle width, the radius of the risk space required by pedestrians, and the risk threshold;
构建非对称纵向行人势场：Construct an asymmetric longitudinal pedestrian potential field:
； ;
； ;
式中，为非对称纵向行人势场，为风险势场比例调节系数，表示车辆需要最短纵向避障距离，为车辆纵向判断参数，为车辆正常行驶至行人处时行人所行进的距离，为行人步行速度。In the formula, is the asymmetric longitudinal pedestrian potential field, is the risk potential field proportional adjustment coefficient, Indicates that the vehicle requires the shortest longitudinal obstacle avoidance distance. is the vehicle longitudinal judgment parameter, The distance that the pedestrian travels when the vehicle normally drives to the pedestrian. is the walking speed of pedestrians.
S1包括，构建速度风险势场，标定车辆时刻的速度势能为：S1 includes constructing the speed risk potential field , calibrate the vehicle The velocity potential energy at the moment is:
； ;
式中，为车辆时刻的速度势能，为时刻的侧向分布系数，为时刻的车辆运行速度，单位为；In the formula, For vehicles The velocity potential energy at time, for The lateral distribution coefficient at time, for The vehicle speed at the time, in units of ;
速度风险势场的强度为：The intensity of the velocity risk potential field is:
；； ; ;
； ;
； ;
式中，为车辆势能差；为速度风险势场强度系数，为纵向风险影响因子，为侧向风险影响因子，为道路路段平均纵坡，取值为，、分别为纵向和侧向的距离，方向由势能高处指向势能低处，单位为，单位为，为纵向速度风险势场强度，为侧向速度风险势场强度，为速度风险势场矢量强度，为风险方向的调整系数。In the formula, is the vehicle potential energy difference; is the velocity risk potential field intensity coefficient, is the vertical risk influencing factor, is the lateral risk influencing factor, is the average longitudinal slope of the road section, The value is , , are the longitudinal and lateral distances respectively, and the direction is from the high potential energy to the low potential energy. The unit is , The unit is , is the longitudinal velocity risk potential field strength, is the lateral velocity risk potential field strength, is the vector strength of the velocity risk potential field, is the adjustment factor for risk direction.
S2包括，将叠加风险势场和速度风险势场作用势根据方向矢量拆分，车道线势场与道路边界线势场视为矢量势场，方向沿轴方向指向道路内侧，将相互作用势按坐标系方向拆分：S2 includes splitting the superposition risk potential field and the speed risk potential field according to the direction vector, and considering the lane line potential field and the road boundary line potential field as vector potential fields. The axis direction points to the inside of the road, and the interaction potential is split according to the direction of the coordinate system:
； ;
； ;
式中，为轴方向的分作用势叠加，为轴方向的分作用势叠加，为车辆行驶方向与道路中心线方向的顺时针夹角，为航向角。In the formula, for The partial potential in the axial direction is superimposed. for The partial potential in the axial direction is superimposed. is the clockwise angle between the vehicle's driving direction and the road centerline direction, is the heading angle.
S3包括，设立设备激活函数为：S3 includes setting up device activation functions for:
； ;
式中，、为合并车道车辆风险势，合并车道车辆风险势按照S2的过程计算，，为交汇点处与道路方向夹角；In the formula, , is the risk potential of vehicles in the merging lanes. The risk potential of vehicles in the merging lanes is calculated according to the process of S2. , is the angle between the intersection point and the road direction;
标定启动阈值，对应激活函数值为，设定时间区间，当激活函数值在时间区间内均大于时，表明此时道路处于拥堵风险高区间，系统主体硬件模块启动；Calibration start threshold , the corresponding activation function value is , set the time interval , when the activation function value is in the time interval The internal When , it indicates that the road is in a high congestion risk range, and the main hardware module of the system is started;
启动后系统主体硬件模块通过视云融合获取合并道路的道路方向风险势，量化道路风险势场，选取风险势相对高路段放行，放行后再次判断双方当前风险势，选取风险势高路段给予通行权；After startup, the main hardware module of the system obtains the road direction risk potential of the merged road through visual cloud fusion, quantifies the road risk potential field, selects the road section with relatively high risk potential for release, and after release, judges the current risk potential of both parties again, and selects the road section with high risk potential for granting the right of way;
设立通行权激活函数：Set up the right-of-way activation function:
； ;
式中，为排队车辆数，为第一合并车道的沿车道方向风险势，为第二合并车道的沿车道方向风险势，为第一合并车道排队车辆沿车道方向风险势，为第二合并车道排队车辆沿车道方向风险势；In the formula, is the number of vehicles in the queue, is the risk potential along the lane direction of the first merging lane, is the risk potential along the lane direction of the second merging lane, is the risk potential of vehicles queuing in the first merging lane along the lane direction, The risk potential of vehicles queuing in the second merging lane along the lane direction;
比较合并车道的通行权激活函数，选取函数数值高的路段通行：Compare the right-of-way activation functions of the merged lanes and select the road section with the higher function value:
； ;
选取通行权切换决策后，判断道路纵向风险收益：After selecting the right-of-way switching decision, determine the longitudinal risk-benefit of the road :
；；； ; ; ;
式中，为车辆当前决策后所受作用纵向作用势，为作用势对距离的高阶参数，为车辆当前所受作用纵向作用势，为道路纵向作用势，是指数参数；In the formula, is the longitudinal potential of the vehicle after the current decision, is a high-order parameter of the potential versus distance, is the longitudinal potential currently acting on the vehicle, is the longitudinal potential of the road, is the exponential parameter;
若大于0，则合并双方此时不存在冲突风险，即调整放行。like If it is greater than 0, there is no conflict risk between the merging parties at this time, that is, the adjustment is released.
相对比现有技术，本发明具有以下有益效果：相较于传统的自由流通行方案或者固定相位的通行方案，本方法更具有动态性和全局性，在保证通行效率的同时，合流车辆的加速度稳定性和航向角稳定性具有较大提升，保证车辆的合流安全；设备采用“埋地式”安装，在道路不拥堵，没有较高合流风险的情况下，设备不激活，位于地下，允许合流车辆采用自由流方式高效率合流，不占用道路面积，不遮挡驾驶员视线，同时也降低设备损耗。当检测到存在合流风险时，设备从地下探出，给出道路通行权。Compared with the prior art, the present invention has the following beneficial effects: compared with the traditional free-flow traffic scheme or the fixed-phase traffic scheme, the present method is more dynamic and global. While ensuring the traffic efficiency, the acceleration stability and heading angle stability of the merging vehicles are greatly improved, ensuring the merging safety of the vehicles; the equipment is installed in an "underground" manner. When the road is not congested and there is no high merging risk, the equipment is not activated and is located underground, allowing merging vehicles to merge efficiently in a free-flow manner without occupying the road area or blocking the driver's line of sight, while also reducing equipment loss. When a merging risk is detected, the equipment emerges from underground and gives the right of way to the road.
附图说明BRIEF DESCRIPTION OF THE DRAWINGS
图1是道路交通场景示意图；Figure 1 is a schematic diagram of a road traffic scene;
图2是车道线风险场强图；Figure 2 is a lane line risk field strength map;
图3是道路边界线风险场强图；Figure 3 is a risk field intensity map of the road boundary line;
图4是叠加风险场强图；Figure 4 is a superimposed risk field strength map;
图5是道路障碍风险势场分布图；FIG5 is a distribution diagram of the road obstacle risk potential field;
图6是道路障碍风险势示意图；FIG6 is a schematic diagram of the risk potential of road obstacles;
图7是人车非对称风险势场分布图；Figure 7 is a distribution diagram of the asymmetric risk potential field of people and vehicles;
图8是人车非对称风险势示意图；Figure 8 is a schematic diagram of the asymmetric risk potential of people and vehicles;
图9是道路场景纵横风险势场矢量分解图；Figure 9 is a vector decomposition diagram of the longitudinal and transverse risk potential field of a road scene;
图10是道路横向叠加作用势图；Figure 10 is a road lateral superposition potential diagram;
图11是道路纵向叠加作用势图。Figure 11 is a diagram of the longitudinal superposition potential of the road.
具体实施方式DETAILED DESCRIPTION
为使本发明的目的、技术方案和优点更加清楚，下面对本发明中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本发明的一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention is described clearly and completely below. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
基于车辆风险场和均衡博弈的道路通行控制系统，包括视云融合检测模块、系统主体硬件模块、系统控制算法模块、系统交互通信模块；The road traffic control system based on vehicle risk field and equilibrium game includes a visual cloud fusion detection module, a system main hardware module, a system control algorithm module, and a system interactive communication module;
视云融合检测模块采用视云融合检测方法获取交通状况信息，采用激光扫描仪获取车辆行人的点云信息，采用雷视一体机获取周遭交通环境的视频信息与雷达数据；The visual-cloud fusion detection module uses the visual-cloud fusion detection method to obtain traffic condition information, uses a laser scanner to obtain point cloud information of vehicles and pedestrians, and uses a radar-visual integrated device to obtain video information and radar data of the surrounding traffic environment;
系统主体硬件模块采用内镶嵌式构造，系统外壳采用碳素钢，内衬PVC层与橡胶层作为缓冲层，内置螺旋电机作为主体骨架，采用stm32单片机作为控制单元，设定上下限位开关控制上升幅度，系统内部空间搭载两架雷视一体机以及一个激光扫描仪；The main hardware module of the system adopts an embedded structure. The system shell is made of carbon steel, lined with PVC layer and rubber layer as a buffer layer, with a built-in spiral motor as the main frame, and an STM32 single-chip microcomputer as the control unit. The upper and lower limit switches are set to control the rise range. The internal space of the system is equipped with two integrated laser vision machines and a laser scanner.
系统控制算法模块采用风险势博弈方法，基于分子力场理论方法建立交通系统行为风险场，非线性加权组合风险势场，对组合风险场沿道路方向进行纵横二维划分，确定道路风险势，在风险势的势井、势峰位置进行通行权变换达到优化道路通行的目的；The system control algorithm module adopts the risk potential game method, establishes the traffic system behavior risk field based on the molecular force field theory method, nonlinearly weighted combined risk potential field, divides the combined risk field into two dimensions along the road direction, determines the road risk potential, and transforms the right of way at the potential well and potential peak of the risk potential to achieve the purpose of optimizing road traffic;
系统交互通信模块利用GNSS-IMU融合定位方法对周围车辆准确定位，获取车辆自身的运动状态信息，计算对位车辆的实时状态，分析车辆运行状态，构建实时的虚拟路况地图，搭建车车、车人、车与环境之间的信息交互桥梁，达到优化道路通行权的效果。The system interactive communication module uses the GNSS-IMU fusion positioning method to accurately locate surrounding vehicles, obtain the vehicle's own motion status information, calculate the real-time status of the corresponding vehicles, analyze the vehicle's operating status, construct a real-time virtual road condition map, and build an information interaction bridge between vehicles, vehicles and people, and vehicles and the environment, so as to achieve the effect of optimizing road access rights.
基于车辆风险场和均衡博弈的道路通行控制方法，使用所述的基于车辆风险场和均衡博弈的道路通行控制系统，包括：A road traffic control method based on a vehicle risk field and equilibrium game, using the road traffic control system based on a vehicle risk field and equilibrium game, comprises:
S1建立由车辆风险势场、道路边界风险势场、道路车道线势场、非对称行人势场和道路障碍物风险势场组成的叠加风险势场，车辆在场中运行的结果是各类风险势场的相互作用力叠加的结果，建立笛卡尔坐标系，对势场进行纵横二维分析；S1 establishes a risk potential field composed of vehicles , Road boundary risk potential field 、Road lane line potential field , asymmetric pedestrian potential field and road obstacle risk potential field The result of the vehicle running in the field is the result of the superposition of the interaction forces of various risk potential fields. A Cartesian coordinate system is established to conduct a two-dimensional analysis of the potential field.
由、、、、组成叠加风险势场：Depend on , , , , Composition of superimposed risk potential field :
； ;
S2在叠加风险势场和速度风险势场基础上，对多目标风险势进行纵横二维量化；S2 quantifies the multi-target risk potential in two dimensions, vertically and horizontally, based on the superposition risk potential field and the velocity risk potential field;
S3基于S1和S2的结果，获得基于道路风险场非线性通行权激活函数，进行道路通行控制。Based on the results of S1 and S2, S3 obtains the nonlinear right-of-way activation function based on the road risk field and performs road traffic control.
S1包括，车辆风险势场作用势来自于车车交互作用力，基于兰纳-琼斯势建立一个统一的车辆作用势场函数模型，引入车辆需求风险距离以及跟驰状态，构建车辆风险势场为：S1 includes that the vehicle risk potential field potential comes from the vehicle-to-vehicle interaction force. A unified vehicle potential field function model is established based on the Lanner-Jones potential. The vehicle required risk distance and following state are introduced to construct the vehicle risk potential field as follows:
； ;
式中，为势场能量尺度，为斥力项幂次，为引力项幂次，为车辆需求的风险距离，是的导数，为车辆的实际间距，为围绕车辆任意质点与车辆质心点之间的顺时针夹角，为修正短程斥力势场分布范围。In the formula, is the energy scale of the potential field, is the power of the repulsive force, is the gravitational power, is the risk distance required by the vehicle, yes The derivative of is the actual distance between vehicles, is the clockwise angle between any mass point around the vehicle and the center of mass of the vehicle, To correct the distribution range of the short-range repulsive potential field.
S1包括，将道路下侧道路边界线作为轴原点，采用高斯类函数对道路车道线势场进行表示：S1 includes taking the road boundary line at the lower side of the road as Axis origin, using Gaussian function to represent the road lane line potential field:
； ;
式中，为道路横截面车道线数量，为任意一点纵坐标，为不同车道线势场的势场强度系数，为道路线类型，为1时为白色虚线，为2时为黄色双黄线，为第条道路线的纵坐标位置，与道路宽度成正比，表示道路边界风险势场的场强增减速度。In the formula, is the number of lane lines in the road cross section, is the ordinate of any point, is the potential field intensity coefficient of different lane line potential fields, is the road line type, When it is 1, it is a white dotted line. When it is 2, it is a double yellow line. For the The vertical coordinate position of the road line, It is proportional to the road width and indicates the rate of increase or decrease of the field strength of the road boundary risk potential field.
S1包括，道路边界风险势场为：S1 includes the road boundary risk potential field:
；； ; ;
式中，为道路两边道路边界线，设定为道路右边界线，为道路左边界线，为点处车辆与点处道路边界线处纵坐标差值，为车辆质心所处位置，为道路边界线所处位置，为位置增益函数。In the formula, Set the road boundary lines on both sides of the road The right boundary line of the road. The left boundary line of the road. for Vehicles at point The vertical coordinate difference of the road boundary line at the point, is the position of the vehicle's center of mass, is the location of the road boundary line, is the position gain function.
S1包括，构建道路障碍物风险势场为：S1 includes constructing the road obstacle risk potential field as follows:
； ;
式中，为道路影响因子风险势场系数，、为车辆质心的横纵坐标，、为道路影响因子质心的横纵坐标，、、、为风险势场的形状系数，、由影响因子的形状尺寸决定。In the formula, is the risk potential field coefficient of the road influencing factor, , is the horizontal and vertical coordinates of the vehicle's center of mass, , is the horizontal and vertical coordinates of the centroid of the road influencing factor, , , , is the shape coefficient of the risk potential field, , Determined by the shape and size of the influencing factor.
S1包括，非对称行人势场包括非对称横向行人势场和非对称纵向行人势场；S1 includes, the asymmetric pedestrian potential field includes an asymmetric transverse pedestrian potential field and an asymmetric longitudinal pedestrian potential field;
非对称横向行人势场为：The asymmetric lateral pedestrian potential field is:
； ;
； ;
式中，为非对称横向行人势场，、为风险势场比例调节系数，、为车辆横纵坐标与行人横纵坐标的差值，为车辆横向判断参数，差值为负时，车辆在行人右边，，差值为正时，车辆在行人左边，，为车辆需要最短纵向避障距离，为包含驾驶员反应时间的避障最短时间，表示车辆需要最短横向避障距离，等于车辆宽度、行人所需风险空间半径及风险阈度之和；In the formula, is the asymmetric lateral pedestrian potential field, , is the risk potential field proportional adjustment coefficient, , is the difference between the horizontal and vertical coordinates of the vehicle and the horizontal and vertical coordinates of the pedestrian, is the vehicle lateral judgment parameter. When the difference is negative, the vehicle is on the right side of the pedestrian. , when the difference is positive, the vehicle is on the left side of the pedestrian, , The vehicle needs the shortest longitudinal obstacle avoidance distance. is the minimum obstacle avoidance time including the driver’s reaction time, Indicates the shortest lateral obstacle avoidance distance required by the vehicle, which is equal to the sum of the vehicle width, the radius of the risk space required by pedestrians, and the risk threshold;
构建非对称纵向行人势场：Construct an asymmetric longitudinal pedestrian potential field:
； ;
； ;
式中，为非对称纵向行人势场，为风险势场比例调节系数，表示车辆需要最短纵向避障距离，为车辆纵向判断参数，为车辆正常行驶至行人处时行人所行进的距离，为行人步行速度。In the formula, is the asymmetric longitudinal pedestrian potential field, is the risk potential field proportional adjustment coefficient, Indicates that the vehicle requires the shortest longitudinal obstacle avoidance distance. is the vehicle longitudinal judgment parameter, The distance that the pedestrian travels when the vehicle normally drives to the pedestrian. is the walking speed of pedestrians.
S1包括，构建速度风险势场，标定车辆时刻的速度势能为：S1 includes constructing the speed risk potential field , calibrate the vehicle The velocity potential energy at the moment is:
； ;
式中，为车辆时刻的速度势能，为时刻的侧向分布系数，为时刻的车辆运行速度，单位为；In the formula, For vehicles The velocity potential energy at time, for The lateral distribution coefficient at time, for The vehicle speed at the time, in units of ;
速度风险势场的强度为：The intensity of the velocity risk potential field is:
；； ; ;
； ;
； ;
式中，为车辆势能差；为速度风险势场强度系数，为纵向风险影响因子，为侧向风险影响因子，为道路路段平均纵坡，取值为，、分别为纵向和侧向的距离，方向由势能高处指向势能低处，单位为，单位为，为纵向速度风险势场强度，为侧向速度风险势场强度，为速度风险势场矢量强度，为风险方向的调整系数。In the formula, is the vehicle potential energy difference; is the velocity risk potential field intensity coefficient, is the vertical risk influencing factor, is the lateral risk influencing factor, is the average longitudinal slope of the road section, The value is , , are the longitudinal and lateral distances respectively, and the direction is from the high potential energy to the low potential energy. The unit is , The unit is , is the longitudinal velocity risk potential field strength, is the lateral velocity risk potential field strength, is the vector strength of the velocity risk potential field, is the adjustment factor for risk direction.
S2包括，将叠加风险势场和速度风险势场作用势根据方向矢量拆分，车道线势场与道路边界线势场视为矢量势场，方向沿轴方向指向道路内侧，将相互作用势按坐标系方向拆分：S2 includes splitting the superposition risk potential field and the speed risk potential field according to the direction vector, and considering the lane line potential field and the road boundary line potential field as vector potential fields. The axis direction points to the inside of the road, and the interaction potential is split according to the direction of the coordinate system:
； ;
； ;
式中，为轴方向的分作用势叠加，为轴方向的分作用势叠加，为车辆行驶方向与道路中心线方向的顺时针夹角，为航向角。In the formula, for The partial potential in the axial direction is superimposed. for The partial potential in the axial direction is superimposed. is the clockwise angle between the vehicle's driving direction and the road centerline direction, is the heading angle.
S3包括，设立设备激活函数为：S3 includes setting up device activation functions for:
； ;
式中，、为合并车道车辆风险势，合并车道车辆风险势按照S2的过程计算，，为交汇点处与道路方向夹角；In the formula, , is the risk potential of vehicles in the merging lanes. The risk potential of vehicles in the merging lanes is calculated according to the process of S2. , is the angle between the intersection point and the road direction;
标定启动阈值，对应激活函数值为，设定时间区间，当激活函数值在时间区间内均大于时，表明此时道路处于拥堵风险高区间，系统主体硬件模块启动；Calibration start threshold , the corresponding activation function value is , set the time interval , when the activation function value is in the time interval The internal When , it indicates that the road is in a high congestion risk range, and the main hardware module of the system is started;
启动后系统主体硬件模块通过视云融合获取合并道路的道路方向风险势，量化道路风险势场，选取风险势相对高路段放行，放行后再次判断双方当前风险势，选取风险势高路段给予通行权；After startup, the main hardware module of the system obtains the road direction risk potential of the merged road through visual cloud fusion, quantifies the road risk potential field, selects the road section with relatively high risk potential for release, and after release, judges the current risk potential of both parties again, and selects the road section with high risk potential for granting the right of way;
设立通行权激活函数：Set up the right-of-way activation function:
； ;
式中，为排队车辆数，为第一合并车道的沿车道方向风险势，为第二合并车道的沿车道方向风险势，为第一合并车道排队车辆沿车道方向风险势，为第二合并车道排队车辆沿车道方向风险势；In the formula, is the number of vehicles in the queue, is the risk potential along the lane direction of the first merging lane, is the risk potential along the lane direction of the second merging lane, is the risk potential of vehicles queuing in the first merging lane along the lane direction, The risk potential of vehicles queuing in the second merging lane along the lane direction;
比较合并车道的通行权激活函数，选取函数数值高的路段通行：Compare the right-of-way activation functions of the merged lanes and select the road section with the higher function value:
； ;
选取通行权切换决策后，判断道路纵向风险收益：After selecting the right-of-way switching decision, determine the longitudinal risk-benefit of the road :
；；； ; ; ;
式中，为车辆当前决策后所受作用纵向作用势，为作用势对距离的高阶参数，为车辆当前所受作用纵向作用势，为道路纵向作用势，是指数参数；In the formula, is the longitudinal potential of the vehicle after the current decision, is a high-order parameter of the potential versus distance, is the longitudinal potential currently acting on the vehicle, is the longitudinal potential of the road, is the exponential parameter;
若大于0，则合并双方此时不存在冲突风险，即调整放行。like If it is greater than 0, there is no conflict risk between the merging parties at this time, that is, the adjustment is released.
针对当前合流路段的通行权冲突问题，本发明提出了一种基于车辆风险场和车车均衡博弈的道路通行权控制方法。首先，在笛卡尔坐标系下，基于分子力场理论方法，建立网联自主车辆行为风险势场，对风险势场进行纵横二维划分，运用高次多项式生成车辆行驶轨迹簇；其次采用风险收益函数，判断是否产生冲突风险场，若出现冲突风险场，则启动设备；最后考虑均衡博弈收益函数及稳定性收益函数，选取放行收益较高路段给予通行权。为验证方案可行性，采用Perscan联合Matlab/Simulink仿真，设置多个仿真环境，验证分析目标交汇区在复杂交互博弈交通场景中的方案优化效果。研究结果表明：优化后通行方案相较于优化前自由流通行方案，满足车辆安全要求的同时，车辆加速度稳定性提高约16.3%，车辆航向角稳定性提高约17.1%。基于车辆风险场和车车均衡博弈的道路通行权控制方法，选取风险收益较高路段放行，考虑了车辆行为均衡博弈的风险收益，能够满足优化通行权方案的安全性、平滑性、舒适性。Aiming at the current right-of-way conflict problem in the merging section, the present invention proposes a road right-of-way control method based on vehicle risk field and vehicle-to-vehicle equilibrium game. Firstly, in the Cartesian coordinate system, based on the molecular force field theory method, the risk potential field of the networked autonomous vehicle behavior is established, the risk potential field is divided into two dimensions in the vertical and horizontal directions, and the vehicle driving trajectory cluster is generated using high-order polynomials; secondly, the risk-return function is used to determine whether a conflict risk field is generated. If a conflict risk field occurs, the device is started; finally, considering the equilibrium game benefit function and the stability benefit function, the road section with higher release benefit is selected to grant the right of way. In order to verify the feasibility of the scheme, Perscan is used in conjunction with Matlab/Simulink simulation, and multiple simulation environments are set up to verify and analyze the optimization effect of the scheme in the target intersection area in the complex interactive game traffic scene. The research results show that compared with the free flow traffic scheme before optimization, the optimized traffic scheme meets the vehicle safety requirements, while the vehicle acceleration stability is improved by about 16.3%, and the vehicle heading angle stability is improved by about 17.1%. The road right-of-way control method based on vehicle risk field and vehicle-to-vehicle equilibrium game selects road sections with higher risk-benefit for release, taking into account the risk-benefit of vehicle behavior equilibrium game, and can meet the safety, smoothness and comfort of the optimized right-of-way scheme.
场是描述物体间相互作用关系，传递能量的方式，可以看作是由无数个质点组成的连续分布，每一个质点的运动都收到场力的影响。场分为标量场和矢量场，标量场包含一个物理量而矢量场包含多个物理量如力、速度、方向、场强等。势能是储存在一个系统即场中的能量，势能并不为单一物体所有，而是相互作用之间的物体所共有。场与场中势能的密切结合即形成了势场，势场表示的是由目标点Uattr(q)所引起的吸引场和由阻碍引起的斥力场Urep(q)之和，因此可以将势场作为物体周围空间内相互作用力的一种描述。在交通系统中同样存在着上述势场，目标车辆可以看作单一分子质点，行驶车辆受周围车辆影响既不过度靠近又不过分远离的现象就类比于物理场中质子受引力斥力影响寻求最优平衡点的过程。智能网联环境下驾驶辅助决策系统起到了至关重要的作用。Field is a way of describing the interaction between objects and transferring energy. It can be regarded as a continuous distribution composed of countless particles. The movement of each particle is affected by the field force. Fields are divided into scalar fields and vector fields. Scalar fields contain one physical quantity while vector fields contain multiple physical quantities such as force, speed, direction, field strength, etc. Potential energy is the energy stored in a system, namely the field. Potential energy is not owned by a single object, but is shared by the objects that interact with each other. The close combination of the field and the potential energy in the field forms a potential field. The potential field represents the sum of the attraction field caused by the target point Uattr(q) and the repulsion field Urep(q) caused by the obstacle. Therefore, the potential field can be used as a description of the interaction force in the space around the object. The above potential field also exists in the traffic system. The target vehicle can be regarded as a single molecular particle. The phenomenon that the moving vehicle is neither too close nor too far away due to the influence of the surrounding vehicles is analogous to the process of protons in the physical field seeking the optimal balance point under the influence of gravitational repulsion. The driving assistance decision system plays a vital role in the intelligent network environment.
道路交通环境中，车道线一般分为两种，一种是将同向车辆分隔开的车道分界线，一种是将不同方向车辆分隔开的车向分界线，如图1所示。In a road traffic environment, lane lines are generally divided into two types: one is the lane dividing line that separates vehicles in the same direction, and the other is the vehicle-direction dividing line that separates vehicles in different directions, as shown in Figure 1.
道路边界线风险势场与道理边界处产生，强度随着车辆靠近向无限远处延伸。道路边界线风险势场相对道路线势场具有更高层次的限制作用，相比于车辆风险势场，道路边界线风险势场位置相对固定，场强的增幅速度更加迅速，随着车辆与边界线距离的缩短而程非线性的增加速度。并且道路边界线风险势场对于车辆的斥力强度远大于对车辆的引力强度。The road boundary line risk potential field is generated at the boundary of the road, and its intensity extends to infinity as the vehicle approaches. The road boundary line risk potential field has a higher level of restriction than the road line potential field. Compared with the vehicle risk potential field, the road boundary line risk potential field has a relatively fixed position, and the field strength increases more rapidly. As the distance between the vehicle and the boundary line decreases, the process increases nonlinearly. In addition, the repulsive force of the road boundary line risk potential field on the vehicle is much greater than the gravitational force on the vehicle.
对道路车道线风险势场及道路边界风险势场进行叠加，车道线风险势场如图2所示，道路边界线风险势场如图3所示，叠加后风险势场如图4所示。如图4所示，势场叠加后，道路边界线处的势场大于双黄线处的势场并且远大于白虚线车道线的势场，并随着距离增大而减弱。三者的叠加势场场强在道路边界处达到峰值，并于道路中点处降至最低值。The risk potential field of the road lane line and the risk potential field of the road boundary are superimposed. The risk potential field of the lane line is shown in Figure 2, the risk potential field of the road boundary line is shown in Figure 3, and the superimposed risk potential field is shown in Figure 4. As shown in Figure 4, after the potential field is superimposed, the potential field at the road boundary line is greater than the potential field at the double yellow line and much greater than the potential field of the white dashed lane line, and weakens with increasing distance. The superimposed potential field strength of the three reaches a peak at the road boundary and drops to the minimum value at the midpoint of the road.
道路障碍物风险势场由道路中的静止障碍，道路自身的环境因素等组成。道路障碍物产生的势场以自身质点为原点，向周边扩散。影响因子风险势场的边界较为固定，当车辆处于边界之外时，势场相互作用力下降较快。风险势场示意图如图5，图6所示。The road obstacle risk potential field is composed of static obstacles on the road and environmental factors of the road itself. The potential field generated by the road obstacle takes its own particle as the origin and spreads to the surrounding area. The boundary of the risk potential field of the influencing factor is relatively fixed. When the vehicle is outside the boundary, the interaction force of the potential field decreases rapidly. The schematic diagram of the risk potential field is shown in Figure 5 and Figure 6.
分析行人过街及车辆运行特性，行人风险势场为非对称势场，行人风险势场的侧向风险远远大于纵向风险，侧向风险中，来车向风险远远大于去车向风险，同时在纵向风险中，前进方向风险远远大于后方风险。Analyzing the characteristics of pedestrian crossing and vehicle operation, the pedestrian risk potential field is an asymmetric potential field. The lateral risk of the pedestrian risk potential field is much greater than the longitudinal risk. Among the lateral risks, the risk from oncoming vehicles is much greater than the risk from outgoing vehicles. At the same time, among the longitudinal risks, the risk in the forward direction is much greater than the risk in the rear direction.
由图7和图8所示，非对称行人过街风险势场中，来车方向场强明显大于去车方向场强，行人行进方向场强明显大于行人后方场强。在横向行人风险势场中，主要考虑车辆的行驶速度，需保证车辆拥有足够的时间空间进行减速避让，纵向行人风险势场中，主要确保行人，车辆相互未觉察的情形下按原速度过街仍能避免发生事故，车速越高，相碰撞所需时间越短，行人行走距离越小，车辆所需的横向避障距离越小，因此，纵向风险势场主要与行人的过街速度，车辆的运行速度以及双方相对位置相关。As shown in Figures 7 and 8, in the asymmetric pedestrian crossing risk potential field, the field strength in the direction of the oncoming vehicle is significantly greater than the field strength in the direction of the outgoing vehicle, and the field strength in the direction of the pedestrian's movement is significantly greater than the field strength behind the pedestrian. In the lateral pedestrian risk potential field, the vehicle's speed is mainly considered, and it is necessary to ensure that the vehicle has enough time and space to slow down and avoid. In the longitudinal pedestrian risk potential field, the main thing is to ensure that pedestrians and vehicles can avoid accidents by crossing the street at the original speed without noticing each other. The higher the speed, the shorter the time required for collision, the shorter the pedestrian's walking distance, and the smaller the lateral obstacle avoidance distance required by the vehicle. Therefore, the longitudinal risk potential field is mainly related to the pedestrian's crossing speed, the vehicle's running speed, and the relative position of both parties.
速度风险势场可展现不同作用势下驾驶员的速度倾向性。车辆速度离散性越小，车辆速度越趋于平均，收到的相互作用势越大，反之离散型越大，速度风险势场的势场强度越小。The speed risk potential field can show the speed tendency of drivers under different potentials. The smaller the vehicle speed discreteness, the more average the vehicle speed is, and the greater the interaction potential received. Conversely, the greater the discreteness, the smaller the potential field strength of the speed risk potential field.
基于上述势场理论，系统考虑多目标因素，建立多目标车辆风险势场量化模型。Based on the above potential field theory, the system considers multi-objective factors and establishes a multi-objective vehicle risk potential field quantitative model.
实际交通场景中，多数时间车辆并未沿道路线方向行驶，如图9所示。In actual traffic scenarios, vehicles do not travel along the road line most of the time, as shown in Figure 9.
行驶方向出现偏斜，则车辆斜方向也与势场之间存在相互作用势，因此根据所建立笛卡尔坐标系，将整个势场作用势根据方向矢量拆分，车道线势场与道路边界线势场视为矢量势场。沿坐标轴方向势场风险强度增长幅度如图10、图11所示。If the driving direction is deviated, there is an interaction potential between the vehicle's oblique direction and the potential field. Therefore, according to the established Cartesian coordinate system, the entire potential field is split according to the direction vector, and the lane line potential field and the road boundary line potential field are regarded as vector potential fields. The growth range of the potential field risk intensity along the coordinate axis direction is shown in Figures 10 and 11.
基于Prescan和Matlab/Simulink联合搭建仿真环境，针对不同道路势场环境进行通行权优化效果验证。A simulation environment is jointly built based on Prescan and Matlab/Simulink to verify the right-of-way optimization effect in different road potential field environments.
表1不同场景仿真参数设置数据Table 1 Simulation parameter setting data for different scenarios
； ;
采用点云传感器与雷达传感器对环境共同感知，根据仿真数据，定义梯度下降算法规则，选取函数沿梯度下降最小方向迭代，采用深度回归模型，对模型效果参数进行敛散性分析，判断建立优化通行权模型各参数敛散性。对通行权模型进行训练，迭代通行风险效益最高的放行方案，后根据筛选出合理安全放行方案。Point cloud sensors and radar sensors are used to perceive the environment together. Based on simulation data, the gradient descent algorithm rules are defined, and the function is selected to iterate along the minimum direction of gradient descent. A deep regression model is used to analyze the convergence and divergence of the model effect parameters, and the convergence and divergence of each parameter in the optimized right-of-way model is determined. The right-of-way model is trained, and the release plan with the highest risk benefit is iterated, and then a reasonable and safe release plan is screened out.
在车辆行驶的轨迹优化中，对于一个n车博弈的情况，设立各参与车的策略集依次为，收益函数分别为，其中为各参与车选择策略组合时参与车的收益，将该博弈表示如下：In the optimization of vehicle driving trajectories, for a game of n vehicles, the strategy sets of each participating vehicle are as follows: The profit functions are ,in Select a strategy combination for each participating car Time participation car The profit of the game is expressed as follows:
； ;
当其他参与车辆均采取纯策略时，即均采取保守策略或均采取激进策略。均采取保守策略时，换道车辆获得最高效率收益，其余参与车辆均停车避让，其余参与车辆的效率收益为，博弈系统的总效率收益为。均采取激进策略时，即车辆在保证安全的前提下均不选择避让，换道车辆获得最低效率收益，其余参与车辆的效率收益为，，式中G为换道车辆继续跟驰的收益，系统总效率收益为；When other participating vehicles adopt pure strategies, that is, they all adopt conservative strategies or aggressive strategies. When all adopt conservative strategies, the lane-changing vehicle obtains the highest efficiency benefit. , the other participating vehicles all stop to avoid, and the efficiency gain of the other participating vehicles is , the total efficiency benefit of the game system is When both vehicles adopt aggressive strategies, that is, the vehicles do not choose to avoid the lane change while ensuring safety, the lane change vehicle obtains the lowest efficiency benefit. , the efficiency gains of the other participating vehicles are, , where G is the benefit of the lane-changing vehicle continuing to follow, and the total efficiency benefit of the system is ;
设定博弈规则1，完全信息博弈，即换道车辆展现换道意图，每位参与者均采取对自身最有利的决策，同时博弈双方即换道车辆和博弈车辆得失互补，博弈收益如下表所示：Set the game rule 1, complete information game, that is, the lane-changing vehicle shows the intention to change lanes, and each participant takes the most favorable decision for himself. At the same time, the gains and losses of the two parties in the game, that is, the lane-changing vehicle and the game vehicle, complement each other. The game benefits are shown in the following table:
表2规则1博弈过程Table 2 Rule 1 Game Process
； ;
在上述矩阵博弈中，存在策略组合满足，则为博弈的策略均衡点，此时换道车辆的均衡收益就是博弈的收益值。In the above matrix game, there is a strategy combination satisfy , then it is the strategic equilibrium point of the game. At this time, the equilibrium payoff of the lane-changing vehicle is the payoff value of the game.
设定博弈规则2，完全信息博弈，换道车辆展现换道意图，每位参与者均采取自身最有利决策，若决策存在冲突则系统收益高一方获胜，双方参与者所获取收益存在折现系数，，，即双方的收益函数修正为：Set game rules 2, complete information game, lane-changing vehicles show their intention to change lanes, each participant takes the most favorable decision for themselves, if there is a conflict in the decision, the party with higher system benefits wins, and the benefits obtained by both participants have a discount coefficient , , , that is, the profit function of both parties is modified to:
表3 规则2博弈过程Table 3 Game process of rule 2
； ;
设定博弈规则3，在规则2的基础上，加入不完全信息博弈，博弈双方不展现自身意图，不再具有博弈先后顺序，判断是否换道，换道方式采取激进策略和保守策略概率均取决于决策所带来的收益，即，博弈收益如下表所示：Game rule 3 is set. On the basis of rule 2, an incomplete information game is added. The two parties do not show their intentions and there is no longer a game order. The probability of deciding whether to change lanes and taking an aggressive or conservative strategy depends on the benefits brought by the decision, that is, , the game benefits are shown in the following table:
表4规则3博弈过程Table 4 Rule 3 Game Process
； ;
为量化博弈收益，考虑智能网联车辆前瞻性及优势所在，引入道路场景车辆风险势场模型，建立道路风险场，采用车辆安全收益及效率收益作为轨迹优化的主要评判标准。结合车辆博弈行为的决策效率收益与决策安全收益，构建博弈收益函数为：In order to quantify the game benefits, considering the foresight and advantages of intelligent connected vehicles, the road scene vehicle risk potential field model is introduced to establish the road risk field, and the vehicle safety benefit and efficiency benefit are used as the main evaluation criteria for trajectory optimization. Combining the decision efficiency benefit and decision safety benefit of vehicle game behavior, the game benefit function is constructed as follows:
； ;
式中，为效益权重系数；In the formula, is the benefit weight coefficient;
建立博弈规则1博弈收益模型，即已知换道车辆产生换道意图，博弈车辆均以自身获得最高效益优先，博弈车辆双方收益可分别表示为：The game profit model of game rule 1 is established. That is, it is known that the lane-changing vehicle has the intention to change lanes. The game vehicles give priority to obtaining the highest benefits. The benefits of both game vehicles can be expressed as:
； ;
式中，为换道车辆与博弈车辆的道路水平方向位置差，换道结束后沿车方向行进距离，，分别为车辆换道后运行速率。为切入角。博弈规则1中，博弈双方均追求效率，取0.3-0.4，提高效率收益权重，切入角取30°。目标车道和换道车辆所在车道行车速率差越大，激进决策正收益所需两车相距距离越大，同时目标车道博弈车辆相距距离为以上激进策略收益才大于保守策略。由上所得博弈模型解如下表所示：In the formula, is the horizontal position difference between the lane-changing vehicle and the game vehicle on the road, The distance traveled in the direction of the vehicle after the lane change is completed. , are the running speeds of vehicles after changing lanes. is the entry angle. In game rule 1, both parties pursue efficiency. Take 0.3-0.4 to improve the efficiency benefit weight and cut-in angle Take 30°. The greater the difference in speed between the target lane and the lane where the lane-changing vehicle is located, the greater the distance between the two vehicles required for a positive benefit of the aggressive decision. At the same time, the distance between the target lane game vehicles is The above aggressive strategy benefits more than the conservative strategy. The game model solution obtained from the above is shown in the following table:
表5 完全信息动态演化博弈模型决策矩阵Table 5 Decision matrix of the complete information dynamic evolutionary game model
； ;
演化博弈理论思想为不断演化策略，使得参与者都能选取高收益策略，演化博弈通过演化动态方程描述选择不同策略后的演化趋势。针对双方参与者所获取收益存在折现系数，，，的演化博弈动态方程设立如下：The idea of evolutionary game theory is to continuously evolve strategies so that participants can choose high-yield strategies. Evolutionary games describe the evolutionary trends after choosing different strategies through evolutionary dynamic equations. There is a discount coefficient for the benefits obtained by both participants. , , , the evolutionary game dynamic equation is established as follows:
； ;
式中，n为变道车辆换道过程中让行的车辆数。Where n is the number of vehicles that give way during the lane-changing process.
此时博弈情况分为三种，方案A，变道车辆采取保守策略即取消变道，这类情况变道车辆跟驰收益为，博弈车辆跟驰收益为，方案B，变道车辆采取激进策略，博弈车辆均采取保守策略，此时系统收益为，事中n为受影响采取保守策略车辆，为第i辆车博弈收益，方案C，当变道车辆采取激进策略，博弈车辆也采取激进策略时，系统收益分析如下，当时，博弈双方沿道路方向的差值为，假定车身长度为3m，寻求博弈系统效益最大化，换道车辆采激进策略后收益如下：At this time, there are three game situations. In plan A, the lane-changing vehicle adopts a conservative strategy, that is, canceling the lane change. In this case, the lane-changing vehicle's following profit is , the game vehicle following the car benefits are , Scheme B, the lane-changing vehicle adopts an aggressive strategy, and the game vehicles all adopt a conservative strategy. At this time, the system profit is , n is the affected vehicle that adopts conservative strategy, is the game profit of the i-th vehicle, scheme C, when the lane-changing vehicle adopts an aggressive strategy and the game vehicle also adopts an aggressive strategy, the system profit analysis is as follows: When , the difference between the two sides of the game along the road direction is Assuming the vehicle body length is 3m, the game system benefits are maximized. The benefits of the lane-changing vehicle after adopting an aggressive strategy are as follows:
； ;
博弈效益如下表所示：The gaming benefits are shown in the following table:
表6决策博弈效益Table 6 Decision-making game benefits
； ;
取得系统收益最高，根据矩阵纳什均衡点对博弈方案采用逆向归纳法总结分析，，用常数，代替,博弈方案如下表所示：To obtain the highest system benefits, the reverse induction method is used to summarize and analyze the game plan based on the matrix Nash equilibrium point. , Use constant , Instead, the game plan is shown in the following table:
表7 折现系数修正动态演化博弈模型方案选取Table 7 Selection of discount coefficient modified dynamic evolution game model scheme
； ;
博弈双方未展现自身意图，不具有博弈先后顺序，换道车辆采取激进决策的概率为p，博弈车辆采取紧急决策的概率为q，p、q正比与决策收益，采用Harsanyi转换，引入一个率先行动者，Nature，从而将博弈车辆关于换道车辆的不完全信息转换为自然行动的不完美信息，则博弈系统最大收益G可表示为：The two parties in the game do not show their intentions and have no game order. The probability of the lane-changing vehicle taking a radical decision is p, and the probability of the game vehicle taking an emergency decision is q. p and q are proportional to the decision benefits. Harsanyi transformation is used to introduce a first mover, Nature, so as to convert the incomplete information of the game vehicle about the lane-changing vehicle into the imperfect information of the natural action. Then the maximum benefit G of the game system can be expressed as:
； ;
式中：为目标车道正常跟驰收益；为换道车辆所在车道跟驰收益。Where: The normal following benefit of the target lane; It is the lane following benefit of the lane where the lane-changing vehicle is located.
对博弈双方的严格列策略进行分析，当换道车辆的保守策略严格劣于激进策略时，即，当博弈车辆的激进策略严格优于保守策略时，即无论换道车辆如何抉择，博弈车辆均采取激进策略需符合条件,随着保守策略所获得收益的增加。The strict strategy of both parties in the game is analyzed. When the conservative strategy of the lane-changing vehicle is strictly inferior to the aggressive strategy, that is, , when the aggressive strategy of the game vehicle is strictly better than the conservative strategy, that is, no matter how the lane-changing vehicle chooses, the game vehicle adopts an aggressive strategy and meets the condition , as the returns obtained from conservative strategies increase.
对于换道车辆而言，无论激进策略与保守策略的效益相差多少，随着车辆保守决策收益的增加，选择保守策略的可能性逐渐增加，对于同等的保守决策收益而言，激进策略与保守策略效益相差约大，车辆采取激进策略可能性越高，对于博弈车辆而言，与均衡点前，博弈车辆采取激进策略的可能性随车辆保守决策收益的增加而增大，当处于均衡点处，即换道车辆激进决策概率的增加，博弈车辆采取激进策略概率的可能性也增加且随着激进策略和保守策略中间效益的差值越大，采取激进策略的可能性越高。随着折现系数的增加，折现系数对于决策影响越小，即激进决策的概率与折现系数成反比关系。由上得出，系统博弈模型的解如下表所示：For lane-changing vehicles, no matter how much the benefits of the aggressive strategy differ from those of the conservative strategy, as the benefits of the vehicle's conservative decision increase, the possibility of choosing a conservative strategy gradually increases. For the same conservative decision benefits, the greater the difference in benefits between the aggressive strategy and the conservative strategy, the higher the possibility of the vehicle adopting an aggressive strategy. For gaming vehicles, before the equilibrium point, the possibility of gaming vehicles adopting an aggressive strategy increases with the increase in the benefits of the vehicle's conservative decision. When at the equilibrium point, that is, the increase in the probability of aggressive decisions by lane-changing vehicles, the probability of gaming vehicles adopting an aggressive strategy also increases, and as the difference in benefits between the aggressive strategy and the conservative strategy increases, the possibility of adopting an aggressive strategy increases. As the discount factor increases, the discount factor has a smaller impact on decision-making, that is, the probability of an aggressive decision is inversely proportional to the discount factor. From the above, the solution to the system game model is shown in the following table:
表8 不完全信息动态演化博弈模型方案选取Table 8 Scheme selection of incomplete information dynamic evolutionary game model
； ;
针对道路场景车车交互的动态博弈特性，建立了系统博弈总效益最优的行为决策，根据不同博弈规则，通过Matlab仿真量化博弈结果，分析了不同博弈规则下的最优决策，为车辆博弈决策与轨迹优化控制提供了理论依据，并得出以下结论：通过建立得失互补的车车博弈模型，该博弈模型为收益的完全转化模型，已知换道车辆存在换道意图，该博弈系统博弈双方的决策选择主要取决于双方所距距离以及双方车辆运行速率差值。引入博弈双方的折现系数，已知换道车辆存在换道意图，博弈双方进行决策选定，博弈规则为双方若有一方选择激进策略，则获取总收益，另一方承担损失，若双方均采取保守策略，则博弈车辆存在损失，若双方均采取激进策略，则以系统最高收益进行辨别，选择系统收益更高一方胜出，但双方收益均存在折现。该博弈模型系统博弈双方的决策主要取决于折现系数及博弈车辆保守收益与激进收益的比值。建立不完全信息的车车博弈模型，即换道车辆未展现任何意图，博弈双方需先判断对方决策的发生概率，后根据系统最高收益选择最优决策，通过量化模型决策概率并对决策收益分析得出，双方策略的选择取决于激进策略与保守策略收益比值及博弈车辆折现系数。探究车车交互博弈的复杂特性，对车车之间动态交互行为决策提供参考，但模型也存在一定局限性，如对车辆驾驶员的记忆性以及驾驶理性要求较高，后续研究还需考虑驾驶员本身驾驶风格的局限性来完善车辆决策演化博弈模型。According to the dynamic game characteristics of vehicle-to-vehicle interaction in road scenes, the behavior decision with the optimal total benefit of the system game is established. According to different game rules, the game results are quantified through Matlab simulation, and the optimal decision under different game rules is analyzed, which provides a theoretical basis for vehicle game decision-making and trajectory optimization control, and draws the following conclusions: By establishing a vehicle-to-vehicle game model with complementary gains and losses, the game model is a complete conversion model of benefits. It is known that the lane-changing vehicle has the intention to change lanes. The decision-making choices of the two parties in the game system mainly depend on the distance between the two parties and the difference in the running speed of the two vehicles. The discount coefficient of the two parties in the game is introduced. It is known that the lane-changing vehicle has the intention to change lanes. The two parties in the game make decisions. The game rules are that if one party chooses an aggressive strategy, it will obtain the total benefit and the other party will bear the loss. If both parties adopt a conservative strategy, the game vehicle will suffer a loss. If both parties adopt an aggressive strategy, the highest system benefit will be used to distinguish, and the party with higher system benefit will be selected as the winner, but the benefits of both parties are discounted. The decision-making of the two parties in the game model system mainly depends on the discount coefficient and the ratio of the conservative benefit to the aggressive benefit of the game vehicle. Establish a car-to-car game model with incomplete information, that is, the lane-changing vehicle does not show any intention. The two parties in the game need to first judge the probability of the other party's decision, and then choose the optimal decision based on the highest system benefit. By quantifying the model's decision probability and analyzing the decision benefits, it is concluded that the choice of both parties' strategies depends on the ratio of aggressive strategy to conservative strategy benefits and the game vehicle discount coefficient. Explore the complex characteristics of car-to-car interactive games and provide a reference for dynamic interactive behavior decisions between cars. However, the model also has certain limitations, such as high requirements for the driver's memory and driving rationality. Subsequent research needs to consider the limitations of the driver's own driving style to improve the vehicle decision evolution game model.
以上实施例仅用于说明本发明的技术方案，而非对其限制，尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分或者全部技术特征进行等同替换，而这些修改或者替换，并不使相应技术方案的本质脱离本发明各实施例技术方案的范围。The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some or all of the technical features may be replaced by equivalents, and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410843225.7ACN118379886B (en) | 2024-06-27 | 2024-06-27 | Road traffic control system and method based on vehicle risk field and equilibrium game |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410843225.7ACN118379886B (en) | 2024-06-27 | 2024-06-27 | Road traffic control system and method based on vehicle risk field and equilibrium game |
| Publication Number | Publication Date |
|---|---|
| CN118379886A CN118379886A (en) | 2024-07-23 |
| CN118379886Btrue CN118379886B (en) | 2024-09-03 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410843225.7AActiveCN118379886B (en) | 2024-06-27 | 2024-06-27 | Road traffic control system and method based on vehicle risk field and equilibrium game |
| Country | Link |
|---|---|
| CN (1) | CN118379886B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118692244B (en)* | 2024-08-28 | 2024-12-31 | 中国第一汽车股份有限公司 | Vehicle driving safety assessment method and device |
| CN119169865B (en)* | 2024-09-04 | 2025-06-13 | 长沙理工大学 | A driving risk intelligent warning method, system and device considering vehicle interaction |
| CN119164416B (en)* | 2024-09-09 | 2025-04-25 | 长沙理工大学 | Path planning method, device, equipment and storage medium |
| CN119541271B (en)* | 2025-01-23 | 2025-05-13 | 四川国蓝中天环境科技集团有限公司 | A risk assessment method for human-vehicle interactive game based on risk perception |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113470388A (en)* | 2021-07-19 | 2021-10-01 | 南阳理工学院 | Traffic signal lamp control method and device based on quantum learning classification |
| CN114399911A (en)* | 2022-01-22 | 2022-04-26 | 菏泽栖凰软件开发有限公司 | Internet intelligent traffic control device for urban road and control system thereof |
| CN117116072A (en)* | 2023-05-17 | 2023-11-24 | 东风汽车集团股份有限公司 | Expressway collaborative decision safe driving method and system |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110363986B (en)* | 2019-06-28 | 2021-09-10 | 江苏大学 | Centralized confluence area vehicle speed optimization method |
| CN115214672A (en)* | 2022-08-16 | 2022-10-21 | 重庆清研理工智能控制技术研究院有限公司 | A human-like decision-making, planning and control method for autonomous driving considering workshop interaction |
| CN116363905B (en)* | 2023-05-19 | 2023-09-05 | 吉林大学 | A lane-changing timing discrimination and active safety control method in a heterogeneous traffic flow merge area |
| CN117002520A (en)* | 2023-07-04 | 2023-11-07 | 吉林大学 | Intelligent vehicle rolling optimization non-cooperative game decision-making method |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113470388A (en)* | 2021-07-19 | 2021-10-01 | 南阳理工学院 | Traffic signal lamp control method and device based on quantum learning classification |
| CN114399911A (en)* | 2022-01-22 | 2022-04-26 | 菏泽栖凰软件开发有限公司 | Internet intelligent traffic control device for urban road and control system thereof |
| CN117116072A (en)* | 2023-05-17 | 2023-11-24 | 东风汽车集团股份有限公司 | Expressway collaborative decision safe driving method and system |
| Title |
|---|
| 基于安全势场的网联自主车辆跟驰行为决策及模型;贾彦峰等;交通运输系统工程与信息;20220228;第22卷(第1期);第87-89页* |
| 基于速度风险势场的高速公路行车风险甄别方法;王博等;浙江大学学报(工学版);20230531;第57卷(第5期);第1002-1003页* |
| 基于非对称势场的人车协同博弈避撞;卢少波等;汽车工程;20221231;第44卷(第10期);第1486页* |
| Publication number | Publication date |
|---|---|
| CN118379886A (en) | 2024-07-23 |
| Publication | Publication Date | Title |
|---|---|---|
| CN118379886B (en) | Road traffic control system and method based on vehicle risk field and equilibrium game | |
| CN110362910B (en) | Game theory-based method for establishing lane-changing conflict coordination model for autonomous vehicles | |
| CN108595823B (en) | Autonomous main vehicle lane changing strategy calculation method combining driving style and game theory | |
| CN109035862B (en) | A multi-vehicle cooperative lane changing control method based on vehicle-vehicle communication | |
| Schörner et al. | Predictive trajectory planning in situations with hidden road users using partially observable markov decision processes | |
| CN114265398A (en) | Trajectory planning method and device for automatic driving vehicle, electronic equipment and storage medium | |
| CN111081065A (en) | Intelligent vehicle cooperative lane change decision-making model under mixed traffic conditions | |
| CN115257746A (en) | Uncertainty-considered decision control method for lane change of automatic driving automobile | |
| CN106997689A (en) | V2P collision free methods based on crossing | |
| CN111445697A (en) | A lane-changing control method for cooperative vehicles on expressways under the condition of intelligent network connection | |
| CN114613142B (en) | A rule-based lane-changing control method for autonomous vehicles at intersections | |
| CN115862329B (en) | Vehicle collaborative lane change guiding method based on expressway traffic accidents | |
| CN118062052A (en) | Automatic driving automobile safety motion planning method considering driving behavior | |
| CN115071758B (en) | A Control Switching Method for Human-Machine Co-driving Based on Reinforcement Learning | |
| WO2023004698A1 (en) | Method for intelligent driving decision-making, vehicle movement control method, apparatus, and vehicle | |
| CN114516328A (en) | Rule-based motorcade following model method in intelligent network environment | |
| CN113011634A (en) | Intelligent network connection ramp merging method based on distributed optimal control | |
| CN113593225B (en) | Single-point intersection vehicle control method oriented to pure internet environment | |
| CN118135831A (en) | Interaction strategy optimization method for intelligent bicycle and manual driving bicycle | |
| Guo et al. | Toward human-like behavior generation in urban environment based on Markov decision process with hybrid potential maps | |
| CN116740945A (en) | Method and system for multi-vehicle collaborative grouping intersection of expressway confluence region in mixed running environment | |
| Zhu et al. | Defensive or competitive Autonomous Vehicles: Which one interacts safely and efficiently with pedestrians? | |
| CN108229078B (en) | A rule-based group motion simulation method in urban traffic scene | |
| Li et al. | A pedestrian trajectory prediction model for right-turn unsignalized intersections based on game theory | |
| CN116476861A (en) | Automatic driving decision system based on multi-mode sensing and layering actions |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |