技术领域technical field
本发明属于智能交通技术领域，涉及基于神经网络的城市交通密度预测方法及基于交通预测的车辆路径规划方法。The invention belongs to the technical field of intelligent transportation, and relates to a neural network-based urban traffic density prediction method and a traffic prediction-based vehicle path planning method.
背景技术Background technique
随着智能交通的发展，城市道路基础设施不断完善，城市获取各种交通信息的途径丰富多样，很轻松就能获得所需要的交通数据。交通数据具有数量巨大，种类繁多，规律难以表述等特征，人工智能和神经网络可以很好地帮助我们找寻出大量交通数据中隐藏的规律。利用神经网络找寻交通状况中的规律，可以预测未来交通状况的走向，制定更加有效，高效的交通管理措施。这因此利用神经网络进行交通预测是具有极大潜力的，并且它们的优点使其成为当前智能交通技术领域的研究热点之一。With the development of intelligent transportation and the continuous improvement of urban road infrastructure, cities can obtain various traffic information in a variety of ways, and it is easy to obtain the required traffic data. Traffic data has the characteristics of huge quantity, various types, and laws that are difficult to express. Artificial intelligence and neural networks can help us find out the hidden laws in a large amount of traffic data. Using the neural network to find the rules in traffic conditions can predict the direction of future traffic conditions and formulate more effective and efficient traffic management measures. Therefore, the use of neural networks for traffic prediction has great potential, and their advantages make them one of the current research hotspots in the field of intelligent transportation technology.
而目前传统的车辆路径规划算法是这样子实现的：对于车辆从起点到目的地，通过遍历路网的节点和道路，寻找到距离目的地最近的道路。传统的车辆路径规划算法能够找到起点和目的地之间的最短路径，但也存在着如下许多缺陷：不能灵活智能地适应复杂多变的城市道路交通；距离远时计算量巨大；容易指引所有车辆往少量道路上行驶造成交通拥堵；选择的道路不具备可调整性。在智能交通和人工智能发展的基础上，传统的车辆路径规划算法已逐渐被改进并且被其他路径规划算法所替换，以便在智能交通与人工智能技术的环境下进行有效的城市交通管理，并进一步适应智能交通与人工智能的发展与应用。The current traditional vehicle route planning algorithm is implemented in this way: for the vehicle from the starting point to the destination, by traversing the nodes and roads of the road network, find the road closest to the destination. The traditional vehicle route planning algorithm can find the shortest path between the starting point and the destination, but there are many defects as follows: it cannot flexibly and intelligently adapt to complex and changeable urban road traffic; the calculation amount is huge when the distance is long; it is easy to guide all vehicles to drive on a small number of roads and cause traffic congestion; the selected road is not adjustable. On the basis of the development of intelligent transportation and artificial intelligence, the traditional vehicle route planning algorithm has been gradually improved and replaced by other route planning algorithms in order to carry out effective urban traffic management in the environment of intelligent transportation and artificial intelligence technology, and further adapt to the development and application of intelligent transportation and artificial intelligence.
经过对现有文献的检索发现，Jinglin Li等人以及Meng Chen等人分别于2019年在《IEEE INTERNET OF THINGS JOURNAL(IEEE物联网杂志)》和2018年在《IEEETransactions on Intelligent Transportation Systems(IEEE智能交通系统学报)》上发表了题为“An End-to-End Load Balancer Based on Deep Learning for VehicularNetwork Traffic Control(基于深度学习的用于车载网络流量控制的端到端负载均衡器)”与“PCNN:Deep Convolutional Networks for Short-Term Traffic CongestionPrediction(PCNN：深度卷积网络短期交通拥堵预测)”的文章。这些文章基于神经网络，分别提出了区域级和道路级的城市交通流量预测方法。尽管区域级和道路级的城市交通流量预测方法能满足部分我们对城市交通状况预测的需求，对指定交通管理措施起到一定指导作用，但对于城市级数量的道路交通密度预测的问题而言，它们难以满足未来智能交通中路径规划的要求。After searching the existing literature, it was found that Jinglin Li et al. and Meng Chen et al. published a paper titled "An End-to-End Load Balancer Based on Deep Learning for VehicularNetwork Traffic Control (End-to-End Load Balancer for Vehicular Network Traffic Control Based on Deep Learning)" and "PCNN: Deep Convolutional Networks for Short-Term Traffic Congestion Prediction (PCNN: Deep Convolutional Network Short-Term Traffic Congestion Prediction)" articles. Based on neural networks, these papers proposed regional-level and road-level urban traffic flow prediction methods, respectively. Although regional-level and road-level urban traffic flow prediction methods can meet some of our needs for urban traffic condition prediction and play a certain guiding role in specifying traffic management measures, for the problem of city-level road traffic density prediction, they are difficult to meet the requirements of route planning in future intelligent transportation.
经检索还发现，为了进一步优化车辆路径规划算法，Chang Guo等人在2018年发表的题为“Real-Time Path Planning in Urban Area via VANET-Assisted TrafficInformation Sharing(通过VANET辅助的交通信息共享进行城市实时路径规划)”的文章中，提出了一种根据实时信息改变车辆行驶路径的方法。通过实时判断前方道路车流状况，判断前方是否是拥堵路段，是否需要更换道路行驶，该方法在一定程度上避免车辆在拥堵路段上行驶。此外，经检索还发现，Miao Wang等人在2015年《IEEE TRANSACTIONS ONVEHICULAR TECHNOLOGY(IEEE车辆技术杂志)》上发表了题为“Real-Time Path PlanningBased on Hybrid-VANET-Enhanced Transportation System(基于实时路径规划混合式VANET增强运输系统)”的文章，该文章在基于路网流量最优化的基础上，对于拥堵程度不同的道路和十字路口进行优先度排序，在拥堵路段和十字路口的车辆可以得到优先更换道路的权利，该方法在一定程度上可以促进路网流量最优化。After searching, it was also found that in order to further optimize the vehicle path planning algorithm, Chang Guo et al. published in 2018 entitled "Real-Time Path Planning in Urban Area via VANET-Assisted Traffic Information Sharing (Real-Time Path Planning in Urban Area via VANET-Assisted Traffic Information Sharing)" In the article, a method of changing the vehicle driving path based on real-time information is proposed. By judging the traffic flow of the road ahead in real time, judging whether the road ahead is congested and whether it is necessary to change the road, this method can avoid vehicles driving on congested roads to a certain extent. In addition, after searching, it was also found that Miao Wang et al published an article entitled "Real-Time Path Planning Based on Hybrid-VANET-Enhanced Transportation System (Real-Time Path Planning Based on Hybrid-VANET-Enhanced Transportation System)" in "IEEE TRANSACTIONS ONVEHICULAR TECHNOLOGY (IEEE Vehicle Technology Magazine)" in 2015. Carry out priority sorting with intersections, and vehicles in congested roads and intersections can get the right to change roads first. This method can promote the optimization of road network flow to a certain extent.
综上所述，现有技术存在的问题是：(1)现有的交通预测方法一般局限为区域级或道路级，难以预测城市数量级的道路车流量(2)大部分路径规划算法无法适应多变的城市交通状况(3)实时根据道路信息调整车辆路径，往往出现明知前方拥堵却没有可选道路的情况。解决上述技术问题的意义在于：基于目前人工智能技术的发展与智能交通技术的进步，更加高效可靠的城市交通预测方法可以提高交通预测的准确度和丰富交通预测的信息量，为未来城市交通管理措施提供新的思路并促进智能交通领域技术及人工智能在交通领域应用的发展。To sum up, the problems existing in the existing technology are: (1) The existing traffic prediction methods are generally limited to the regional or road level, and it is difficult to predict the road traffic flow of the city order; (2) Most route planning algorithms cannot adapt to the changing urban traffic conditions; The significance of solving the above technical problems lies in: Based on the current development of artificial intelligence technology and the progress of intelligent transportation technology, more efficient and reliable urban traffic forecasting methods can improve the accuracy of traffic forecasting and enrich the amount of traffic forecasting information, provide new ideas for future urban traffic management measures and promote the development of intelligent transportation technology and the application of artificial intelligence in the transportation field.
发明内容Contents of the invention
针对现有技术存在的问题，本发明目的在进行城市路网栅格化道路车辆密度预测的基础上提供了一种栅格化的精细路径规划方法。Aiming at the problems existing in the prior art, the purpose of the present invention is to provide a gridded fine path planning method on the basis of gridded urban road network vehicle density prediction.
本发明技术方案是这样实现的，基于路网栅格化道路车流预测的路径规划方法，所述城市路网栅格化道路车流密度预测方法及栅格化精细路径规划方法包括以下步骤：The technical solution of the present invention is realized in this way, based on the route planning method of gridded road traffic flow prediction of road network, the gridded road traffic density prediction method of urban road network and the gridded fine path planning method include the following steps:
步骤1：基于城市的路网图，将城市划分为M×N(M、N均为30-200)的网格区域，在城市的路网中，将某一时刻的所有车辆的GPS或北斗数据根据经纬度和方向分配到所属的网格内，得到城市路网图的向不同方向的密度图；Step 1: Based on the city's road network map, divide the city into M×N (both M and N are 30-200) grid areas. In the city's road network, assign the GPS or Beidou data of all vehicles at a certain time to the grid according to the longitude, latitude and direction, and obtain the density map of the urban road network map in different directions;
步骤2：将得到的路网的某一时刻的密度图作为神经网络的输入，利用前30±15分钟的路网密度数据预测未来10±5分钟的神经网络；Step 2: Use the obtained density map of the road network at a certain moment as the input of the neural network, and use the road network density data of the previous 30 ± 15 minutes to predict the neural network in the next 10 ± 5 minutes;
步骤3：根据预测前某一时刻的道路周围网格区域的车流密度以及道路车流密度和平均速度的关系，计算得到预测后各道路的车辆平均行驶速度。Step 3: According to the traffic density of the grid area around the road at a certain moment before the prediction and the relationship between the road traffic density and the average speed, calculate the average vehicle speed of each road after the prediction.
步骤4：根据某车辆当前所在位置划分一个栅格区域，并根据预测的未来10±5分钟的道路车辆密度和平均行驶速度信息，选择车辆在某一栅格区域内的道路。Step 4: Divide a grid area according to the current location of a vehicle, and select the road where the vehicle is in a certain grid area according to the predicted road vehicle density and average driving speed information in the next 10±5 minutes.
步骤5：若某车辆的行驶终点不在栅格区域内，在车辆即将完成该10±5分钟内规划的路程时重复步骤4，直到终点位于某栅格区域内。Step 5: If the driving end point of a certain vehicle is not within the grid area, repeat step 4 when the vehicle is about to complete the planned journey within 10±5 minutes until the end point is within a certain grid area.
进一步，所述利用卷积神经网络来对城市级别区域范围内的交通(车辆)密度预测，可以有效预测未来一定时间的交通变化(车辆道路流量)变化情况。Further, the use of the convolutional neural network to predict the traffic (vehicle) density within the city-level area can effectively predict the traffic change (vehicle road flow) change in a certain period of time in the future.
所述的城市路网图进行网格划分，且根据车辆的行驶方向划分东南西北四个方向区间，因此可以根据网格与具体道路进行匹配，将城市区域级的交通预测精细化到每一条道路上。The urban road network map is divided into grids, and four directions of southeast, northwest, and north are divided according to the driving direction of the vehicle. Therefore, the grid can be matched with specific roads, and the traffic forecast at the urban regional level can be refined to each road.
车辆进行路径规划同时考虑栅格区域内道路的密度预测信息计算车辆的最短路径以及栅格区域边缘点距离终点的估算距离，保证车辆在前往终点的有效道路上。本发明主要是利用预测信息来把握未来交通的变化趋势。The path planning of the vehicle takes into account the density prediction information of the roads in the grid area to calculate the shortest path of the vehicle and the estimated distance from the edge point of the grid area to the end point to ensure that the vehicle is on the effective road to the end point. The present invention mainly utilizes forecast information to grasp the changing trend of future traffic.
道路的车流密度与道路平均速度有密切关系，本方法利用道路的车流密度计算每个道路的车辆平均速度并用于路径规划。The traffic density of the road is closely related to the average speed of the road. This method uses the traffic density of the road to calculate the average speed of each road and use it for path planning.
车辆利用预测的交通信息来进行路径规划，可以避免选择未来将会变得拥堵的路段，长期还具备交通诱导、均衡城市道路车流量的作用。Vehicles use predicted traffic information for route planning, which can avoid choosing road sections that will become congested in the future, and can also play the role of traffic guidance and balancing urban road traffic flow in the long run.
尤其是车辆每次只对10分钟区域内路径进行规划，可以有效减少交通预测随着时间变长，预测准确度指数下降的问题，同时可以减少道路突发事故带来的拥堵。可以有效减少交通预测随着时间变长，准确度指数下降的问题，同时可以更加适应复杂多变的交通情况。In particular, the vehicle only plans the route within the 10-minute area at a time, which can effectively reduce the problem of the decline in the accuracy index of the traffic forecast as the time becomes longer, and at the same time reduce the congestion caused by sudden road accidents. It can effectively reduce the problem that the accuracy index of traffic prediction decreases with time, and can be more adaptable to complex and changeable traffic conditions.
所述的车辆进行路径规划时同时考虑栅格区域内利用车辆密度预测信息计算的最短路径以及栅格区域边缘点距离终点的估算距离，保证车辆在前往终点的有效道路上。When the vehicle is planning the path, the shortest path calculated using the vehicle density prediction information in the grid area and the estimated distance from the edge point of the grid area to the end point are considered at the same time, so as to ensure that the vehicle is on the effective road to the end point.
以交通大数据中心作为收集交通信息，进行交通预测并为车辆提供路径规划的中心控制管理机制实现步骤如下：当一辆车辆需要从起点到终点进行路径规划时，划分以起点为原点，半径为10分钟路程的扇形区域内的道路作为栅格区域，向交通管理器发送进入时刻以及起点终点；交通大数据中心不断根据城市交通收集的信息预测未来的交通信息，在收到车辆的路径需求信息后根据栅格内的预测信息以及栅格边缘点到终点的估算距离为车辆进行栅格内路径的规划；The traffic big data center is used as the central control management mechanism to collect traffic information, make traffic forecasts and provide vehicles with path planning. The implementation steps are as follows: When a vehicle needs to plan a path from the starting point to the destination, divide the road in the fan-shaped area with the starting point as the origin and a radius of 10 minutes as a grid area, and send the entry time and the starting and ending points to the traffic manager; the traffic big data center continuously predicts future traffic information based on the information collected by urban traffic. Plan the path within the grid;
城市路网典型的划分为100乘100的网格区域并将车辆的行驶方向划分成东南西北四个方向区域，然后以10分钟为一个时间区间将车辆的GPS数据根据经纬度和行驶方向分配到所属的网格内，得到城市路网中车辆往各个方向行驶的密度图：A typical urban road network is divided into grid areas of 100 by 100, and the driving direction of vehicles is divided into four areas in the southeast, northwest, and then the GPS data of the vehicle is assigned to the corresponding grid according to the longitude, latitude and driving direction in a time interval of 10 minutes, and the density map of vehicles traveling in various directions in the urban road network is obtained:
采用卷积神经网络和残差神经结构来进行交通预测，将实时收集的车流密度图(前30±15分钟的路网密度数据)作为神经网络的输入就用于预测下一个时间区间即10分钟后的车流密度图，并将密度图与路网中的每一条道路进行匹配，最后根据道路密度与道路平均速度的关系v＝vmax(1-λ/λmax)，计算每条道路的平均行驶速度，得到各条道路的速度信息(某时刻t从i到j的道路的平均速度)：Convolutional neural network and residual neural structure are used for traffic prediction, and the real-time collected traffic flow density map (the road network density data of the first 30±15 minutes) is used as the input of the neural network to predict the traffic flow density map in the next time interval, that is, 10 minutes later, and the density map is matched with each road in the road network. Finally, according to the relationship between road density and road average speed v=vmax (1-λ/λmax ), the average driving speed of each road is calculated, and the speed information of each road (at a certain time t from i the average speed of the road to j):
其中vmax是道路的限速，λ和λmax分别是道路的车流密度与车流密度最大值。Where vmax is the speed limit of the road, and λ and λmax are the traffic density and the maximum value of the traffic density of the road respectively.
在路径规划部分，栅格化精细路径规划分为栅格内最优路径计算以及从起点经各节点到终点的估算距离计算两个方面，当某车辆t时刻从起点O到终点D，根据两点的距离建立椭圆的路径搜索空间即从起点O到终点D可能经过的节点(节点代表不同道路的连接点，在一个栅格区域内存在多个节点)，对于搜索空间内的节点计算其到终点的估算距离；路径规划的栅格与交通预测的栅格不太一样，路径规划的栅格是根据车辆位置划分的10分钟路程内即10min*vmax内的小栅格区域综合，然后节点是道路的连接点，即道路是从A点到B点的这种节点，然后边缘点就是节点位于大栅格的最外围并且可以从这个节点离开当前栅格区域的节点。In the path planning part, the grid-based fine path planning is divided into two aspects: the calculation of the optimal path in the grid and the calculation of the estimated distance from the start point to the end point through each node. When a vehicle travels from the start point O to the end point D at time t, an elliptical path search space is established according to the distance between the two points, that is, the nodes that may pass through from the start point O to the end point D (nodes represent the connection points of different roads, and there are multiple nodes in a grid area), and the estimated distance to the end point is calculated for the nodes in the search space. The small grid area within 10 minutes divided by the vehicle position is integrated, that is, the small grid area within 10min*vmax, and then the node is the connection point of the road, that is, the road is the node from point A to point B, and the edge point is the node that is located at the outermost edge of the large grid and can leave the current grid area from this node.
每次车辆划分当前位置为原点，半径为10分钟路程内的所有路网预测小栅格区域作为车辆路径规划的大栅格区域，根据预测信息计算当前位置到栅格内边缘点(边缘点即为点位于栅格的边缘位置，并存在从该点离开栅格的道路)的最短时间及路径，再综合边缘节点到终点的估算距离，判断从哪个边缘节点离开栅格区域，当车辆即将离开栅格区域时重新划分栅格，进行路径计算过程，直到终点在栅格内时终止；Each time the vehicle divides the current position as the origin and all road network prediction small grid areas within a radius of 10 minutes are used as the large grid area for vehicle path planning, the shortest time and path from the current position to the edge point in the grid (the edge point is the point at the edge of the grid, and there is a road leaving the grid from this point) is calculated according to the prediction information, and then the estimated distance from the edge node to the end point is integrated to determine which edge node leaves the grid area. end within time;
首先定义栅格区域的节点集为A＝{a1,a2,a3…}以及搜索空间B＝{b1,b2,b3…}First define the node set of the grid area as A={a1 ,a2 ,a3 ...} and the search space B={b1 ,b2 ,b3 ...}
综上所述，本实施例提供的每次车辆进行栅格化的路径规划问题可以写成：To sum up, the path planning problem for each vehicle gridded in this embodiment can be written as:
其中ai,分别是栅格区域内的普通节点和边缘节点,/>为当前位置到栅格区域边缘节点ai的最短时间，/>为边缘节点到终点的估算距离，α为0，1之间的权值，该权值的意思是栅格内根据预测信息进行的路径规划占较大的权重，这是因为栅格内的路径规划是根据预测信息计算的精确的时间，而估算距离是根据静态交通信息计算的值，比根据预测信息计算的准确度低。where ai , They are ordinary nodes and edge nodes in the grid area, /> is the shortest time from the current position to the edge node ai of the grid area, /> is the estimated distance from the edge node to the end point, and α is a weight between 0 and 1. The weight means that the path planning based on the prediction information in the grid occupies a relatively large weight. This is because the path planning in the grid is calculated based on the accurate time of the prediction information, and the estimated distance is calculated based on the static traffic information, which is less accurate than the calculation based on the prediction information.
每次车辆划分栅格都进行一次上述问题的求解，当ai中包括终点时终止。Each time the vehicle is divided into grids, the above problem is solved once, and it is terminated when ai includes the end point.
本发明是一种基于路网栅格化道路车流预测的精细化路径规划方法。采用城市交通密度方向划分，利用神经网络预测交通密度以及车辆导航的栅格化路径规划算法。由于道路的平均速度与道路的车流量密度紧密相关，车辆能够根据预测的城市道路车流量密度阶段性地选择不同道路，从而寻找每个阶段中时间代价最小的行驶路径。相比于传统的交通导航方法，本发明基于人工智能技术来预测城市道路车流密度，通过栅格化的路径规划算法避免车辆重复拥堵情况，能够找出通往目的地所需时间更短的道路，提高行驶过程的平均速度，进一步达到交通诱导的作用。The invention is a refined path planning method based on road network gridded road traffic flow prediction. Using urban traffic density direction division, using neural network to predict traffic density and grid path planning algorithm for vehicle navigation. Since the average speed of the road is closely related to the traffic flow density of the road, vehicles can choose different roads in stages according to the predicted traffic flow density of urban roads, so as to find the driving route with the least time cost in each stage. Compared with the traditional traffic navigation method, the present invention predicts the traffic density of urban roads based on artificial intelligence technology, avoids repeated congestion of vehicles through a gridded path planning algorithm, can find a road with a shorter time to the destination, improves the average speed of the driving process, and further achieves the effect of traffic guidance.
有益效果：与现有技术相比，本发明的优点如下，首先，该城市路网栅格化道路车辆密度预测方法及栅格化的精细路径规划方法能够有效地避免车辆往拥堵区域和路段方向行驶；其次，城市级的路网栅格化道路车流密度预测方法是本专利的突出贡献之一；再次，车辆进行路径规划时进行栅格化的阶段性路径规划，降低了计算的复杂度，提高了车辆选择道路的灵活性，极大程度上提高了整个交通系统的鲁棒性。Beneficial effects: Compared with the prior art, the present invention has the following advantages. First, the urban road network gridded road vehicle density prediction method and the gridded fine route planning method can effectively prevent vehicles from driving in the direction of congested areas and road sections; secondly, the city-level road network gridded road traffic density prediction method is one of the outstanding contributions of this patent; thirdly, the gridded staged path planning for vehicles reduces the complexity of calculation, improves the flexibility of vehicles choosing roads, and greatly improves the robustness of the entire traffic system.
附图说明Description of drawings
图1是本发明实施例所采用的栅格化精细路径规划场景图。FIG. 1 is a scene diagram of a gridded fine path planning used in an embodiment of the present invention.
图2是本发明实施例城市路网栅格化道路车流密度预测框图。Fig. 2 is a block diagram of gridded road traffic density prediction in an urban road network according to an embodiment of the present invention.
图3是本发明实施例栅格区域内最优路径算法实现框图。Fig. 3 is a block diagram of implementing an optimal path algorithm in a grid area according to an embodiment of the present invention.
图4是本发明实施例各节点到终点估算距离算法实现框图。FIG. 4 is a block diagram of an algorithm for estimating distances from each node to a destination in an embodiment of the present invention.
图5基于预测信息的栅格化精细路径规划方法和传统路径规划方法的平均出行时间比较示意图；Figure 5. Schematic diagram of the comparison of the average travel time between the rasterized fine path planning method based on forecast information and the traditional path planning method;
图6基于预测信息的栅格化精细路径规划方法和传统路径规划方法的平均出行速度比较示意图；Figure 6. Schematic diagram of the comparison of the average travel speed between the rasterized fine path planning method based on forecast information and the traditional path planning method;
图7基于预测信息的栅格化精细路径规划方法和传统路径规划方法的平均计算时间比较示意图。Fig. 7 is a schematic diagram comparing the average calculation time between the rasterized fine path planning method based on prediction information and the traditional path planning method.
具体实施方式Detailed ways
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图对本发明的实施例作详细说明：本实施例在以本发明技术方案为前提下进行实施，给出了详细的实施方式和具体的操作过程。应当理解，此处所描述的具体实例仅仅用以解释本发明，但本发明的保护范围不限于下述的实施例。In order to make the purpose, technical solutions and advantages of the present invention clearer, the embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings: this embodiment is implemented on the premise of the technical solutions of the present invention, and detailed implementation methods and specific operating procedures are provided. It should be understood that the specific examples described here are only used to explain the present invention, but the protection scope of the present invention is not limited to the following examples.
实施例Example
本实施例采用了图1的城市交通路网场景，提出了一种城市路网栅格化道路车辆密度预测方法及栅格化的精细路径规划方法。首先在该交通场景下，需要在城市设立交通大数据中心。交通大数据中心负责整个城市交通数据的收集，将收集的信息进行整理分析，并利用神经网络进行城市路网栅格化的道路车辆密度预测，然后根据预测的交通信息为需要进行路径规划的车辆提供服务。This embodiment adopts the urban traffic road network scene in FIG. 1 , and proposes a gridded urban road vehicle density prediction method and a gridded fine path planning method. First of all, in this traffic scenario, it is necessary to set up a traffic big data center in the city. The traffic big data center is responsible for the collection of traffic data in the whole city, organizes and analyzes the collected information, and uses the neural network to predict the road vehicle density of the urban road network grid, and then provides services for vehicles that need to be routed according to the predicted traffic information.
本实施例的基本目标是通过利用基于人工智能的交通预测技术得到路网交通预测信息，并用于车辆的路径选择。为了避免长时间预测导致的准确度下降和减少路径规划的计算量，交通大数据中心每次只需要对以车辆为中心半径为10分钟路程的栅格区域内的路径进行规划。以交通大数据中心作为收集交通信息，进行交通预测并为车辆提供路径规划的中心控制管理机制实现步骤如下：当一辆车辆需要从起点到终点进行路径规划时，划分以起点为原点，半径为10分钟路程的扇形区域内的道路作为栅格区域，向交通管理器发送进入时刻以及起点终点；交通大数据中心不断根据城市交通收集的信息预测未来的交通信息，在收到车辆的路径需求信息后根据栅格内的预测信息以及栅格边缘点到终点的估算距离为车辆进行栅格内路径的规划。上述车辆路径规划机制的特点在于，车辆需要进行路径规划时交通大数据中心只需要进行栅格内的路径计算，计算复杂度低，此外交通大数据中心每次只为车辆规划10分钟的路径，可以使得车辆的路径具备可调整性并充分利用预测的交通信息。The basic goal of this embodiment is to obtain road network traffic prediction information by using artificial intelligence-based traffic prediction technology, and use it for vehicle route selection. In order to avoid the decrease in accuracy caused by long-term forecasting and reduce the calculation amount of path planning, the traffic big data center only needs to plan the path in the grid area with a radius of 10 minutes from the center of the vehicle each time. The traffic big data center is used as the central control management mechanism to collect traffic information, make traffic forecasts and provide vehicles with path planning. The implementation steps are as follows: When a vehicle needs to plan a path from the starting point to the destination, divide the road in the fan-shaped area with the starting point as the origin and a radius of 10 minutes as a grid area, and send the entry time and the starting and ending points to the traffic manager; the traffic big data center continuously predicts future traffic information based on the information collected by urban traffic. Plan the path in the grid. The characteristic of the above-mentioned vehicle route planning mechanism is that when a vehicle needs to perform route planning, the traffic big data center only needs to calculate the route within the grid, and the calculation complexity is low. In addition, the traffic big data center only plans a 10-minute route for the vehicle at a time, which can make the vehicle's route adjustable and make full use of the predicted traffic information.
在本实施例中可将城市路网栅格化道路车辆密度预测方法及栅格化的精细路径规划方法分为交通预测，路径规划两个部分。In this embodiment, the urban road network gridded road vehicle density prediction method and the gridded fine path planning method can be divided into two parts: traffic prediction and path planning.
在交通预测部分，我们将某城市路网典型的划分为100乘100的网格区域并将车辆的行驶方向划分成东南西北四个方向区域，然后以10分钟为一个时间区间将车辆的GPS数据根据经纬度和行驶方向分配到所属的网格内，得到城市路网中车辆往各个方向行驶的密度图：In the traffic prediction part, we typically divide a city’s road network into 100 by 100 grid areas and divide the driving direction of vehicles into four areas in the southeast, northwest, and then use 10 minutes as a time interval to assign the vehicle’s GPS data to the corresponding grid according to the latitude and longitude and driving direction, and obtain the density map of vehicles traveling in various directions in the urban road network:
其中(m,n)为第m行第n列的网格区域，是t时刻车辆k的信息，包括位置信息以及速度信息/>意味着车辆t时刻在网格区域(m,n)内且行驶方向在方向区间si内。因此位于行驶方向si的网格区域(m,n)内的车辆密度可用xt(si,m,n)来表示。最后我们可以得到每个时间间隔的城市路网车流密度分布情况：Where (m,n) is the grid area of row m and column n, is the information of vehicle k at time t, including position information and speed information /> It means that the vehicle is in the grid area (m,n) at time t and the driving direction is in the direction intervalsi . Therefore, the vehicle density in the grid area (m,n) in the driving direction si can be represented by xt (si ,m,n). Finally, we can get the traffic density distribution of the urban road network at each time interval:
{Xt|t＝t0,t0-1…}{Xt |t＝t0 ,t0 -1...}
其次我们建立一个神经网络来进行交通的预测。在本发明中我们采用了卷积神经网络和残差神经结构来进行交通预测，卷积神经网络是一种功能强大的神经网络，通过共享参数等可以大量减少神经网络的训练复杂度，而残差神经结构在训练深层次网络中表现出很好的效果。一层卷积神经网络可以用下式表示：Secondly, we build a neural network to predict traffic. In the present invention, we use a convolutional neural network and a residual neural structure for traffic prediction. The convolutional neural network is a powerful neural network, which can greatly reduce the training complexity of the neural network by sharing parameters, and the residual neural structure shows good results in training deep-level networks. A layer of convolutional neural network can be represented by the following formula:
Xl+1＝fl(Wl*Xl+bl)Xl+1 ＝fl (Wl *Xl +bl )
其中*代表卷积运算，Xl是神经网络的输入也是上一层神经网络的输出，Xl+1则是该层神经网络的输出，Wl和bl是该层神经网络的线性运算参数，而fl是该层神经网络的非线性运算函数，在本研究中我们采用的是ReLu函数。Among them, * represents the convolution operation, Xl is the input of the neural network and the output of the neural network of the previous layer, Xl+1 is the output of the neural network of this layer, Wl and bl are the linear operation parameters of the neural network of this layer, and fl is the nonlinear operation function of the neural network of this layer, and we use the ReLu function in this study.
而残差神经结构由三层卷积神经网络组成，并且它的输入与输出是相连的，因此它的公式如下：The residual neural structure consists of a three-layer convolutional neural network, and its input and output are connected, so its formula is as follows:
和/>分别是残差神经结构的输入和输出，/>指的是残差神经结构的具体构造，其包含了3个卷积神经网络和ReLu函数。 and /> are the input and output of the residual neural structure, respectively, /> Refers to the specific construction of the residual neural structure, which includes 3 convolutional neural networks and ReLu functions.
然后根据历史交通信息以及误差函数对神经网络进行训练，训练目标是使得误差函数的值最小，误差函数如下：Then the neural network is trained according to the historical traffic information and the error function. The training goal is to minimize the value of the error function. The error function is as follows:
其中xt(si,m,n)和分别是交通信息的真实值以及预测值，S是方向区间的个数，比如在东南西北四个方向区间的情况下S＝4。where xt (si ,m,n) and are the actual value and predicted value of the traffic information, respectively, and S is the number of direction intervals, for example, S=4 in the case of four direction intervals in southeast, north, and south.
最后将实时收集的车流密度图作为神经网络的输入就可以预测下一个时间区间的车流密度图，并将密度图与路网中的每一条道路进行匹配，最后根据道路密度与道路平均速度的关系v＝vmax(1-λ/λmax)，计算每条道路的平均行驶速度，得到各条道路的速度信息(某时刻t从i到j的道路的平均速度)：Finally, the real-time collected traffic flow density map is used as the input of the neural network to predict the traffic flow density map of the next time interval, and the density map is matched with each road in the road network. Finally, according to the relationship between road density and road average speed v=vmax (1-λ/λmax ), the average driving speed of each road is calculated to obtain the speed information of each road (the average speed of the road from i to j at a certain time t):
V(i,j,t)V(i,j,t)
城市路网栅格化的交通车流密度预测效果如图2所示。Figure 2 shows the traffic flow density prediction effect of gridded urban road network.
在路径规划部分，栅格化精细路径规划可分为栅格内最优路径计算以及从起点经各节点到终点的估算距离计算两个方面，当某车辆t时刻从起点O到终点D，根据两点的距离建立椭圆的路径搜索空间即从起点O到终点D可能经过的节点(节点代表不同道路的连接点，在一个栅格区域内存在多个节点)，对于搜索空间内的节点计算其到终点的估算距离。每次车辆划分当前位置为原点，半径为10分钟路程内的所有路网预测小栅格区域作为车辆路径规划的大栅格区域，根据预测信息计算当前位置到栅格内边缘点(边缘点即为点位于栅格的边缘位置，并存在从该点离开栅格的道路)的最短时间及路径，再综合边缘节点到终点的估算距离，判断从哪个边缘节点离开栅格区域，当车辆即将离开栅格区域时重新划分栅格，进行路径计算过程，直到终点在栅格内时终止。在这我们首先定义栅格区域的节点集为A＝{a1,a2,a3…}以及搜索空间B＝{b1,b2,b3…}In the path planning part, the grid-based fine path planning can be divided into two aspects: the calculation of the optimal path in the grid and the calculation of the estimated distance from the starting point to the end point through each node. When a vehicle travels from the starting point O to the end point D at time t, an elliptical path search space is established according to the distance between the two points, that is, the nodes that may pass through from the starting point O to the end point D (nodes represent the connection points of different roads, and there are multiple nodes in a grid area), and the estimated distance to the end point is calculated for the nodes in the search space. Each time the vehicle divides the current position as the origin and all road network prediction small grid areas within a radius of 10 minutes are used as the large grid area for vehicle path planning, the shortest time and path from the current position to the edge point in the grid (the edge point is the point at the edge of the grid, and there is a road leaving the grid from this point) is calculated according to the prediction information, and then the estimated distance from the edge node to the end point is integrated to determine which edge node leaves the grid area. Terminate within time. Here we first define the node set of the grid area as A={a1 ,a2 ,a3 ...} and the search space B={b1 ,b2 ,b3 ...}
综上所述，本实施例提供的每次车辆进行栅格化的路径规划问题可以写成：To sum up, the path planning problem for each vehicle gridded in this embodiment can be written as:
其中ai,分别是栅格区域内的普通节点和边缘节点,/>为当前位置到栅格区域边缘节点ai的最短时间，/>为边缘节点到终点的估算距离，α为0，1之间的权值，该权值的意思是栅格内根据预测信息进行的路径规划占较大的权重，这是因为栅格内的路径规划是根据预测信息计算的精确的时间，而估算距离是根据静态交通信息计算的值，比根据预测信息计算的准确度低。where ai , They are ordinary nodes and edge nodes in the grid area, /> is the shortest time from the current position to the edge node ai of the grid area, /> is the estimated distance from the edge node to the end point, and α is a weight between 0 and 1. The weight means that the path planning based on the prediction information in the grid occupies a relatively large weight. This is because the path planning in the grid is calculated based on the accurate time of the prediction information, and the estimated distance is calculated based on the static traffic information, which is less accurate than the calculation based on the prediction information.
路径规划的栅格与交通预测的栅格不太一样，路径规划的栅格是根据车辆位置划分的10分钟路程内即10min*vmax内的小栅格区域综合，然后节点是道路的连接点，即道路是从A点到B点的这种节点，然后边缘点就是节点位于大栅格的最外围并且可以从这个节点离开当前栅格区域的节点。The grid of route planning is different from the grid of traffic prediction. The grid of route planning is based on the synthesis of small grid areas within 10 minutes of distance divided by the vehicle position, that is, within 10min*vmax. Then the nodes are the connection points of the road, that is, the road is the node from point A to point B, and the edge point is the node that is located at the outermost periphery of the large grid and can leave the current grid area from this node.
每次车辆划分栅格都进行一次上述问题的求解，当ai中包括终点时终止。Each time the vehicle is divided into grids, the above problem is solved once, and it is terminated when ai includes the end point.
求解上述问题包括两方面计算，各节点到终点的估算距离计算以及栅格内最优路径计算，其求解过程分别如图3和图4所示。Solving the above problems includes two aspects of calculation, the calculation of the estimated distance from each node to the end point and the calculation of the optimal path in the grid. The solving process is shown in Figure 3 and Figure 4 respectively.
对于栅格内的最优路径计算，参考dijistra算法，计算从起点到栅格内边缘点的最短时间，每次从栅格内未选择过的节点中选择离起点所需时间最短的节点，更新起点到其邻居节点的最短时间：For the calculation of the optimal path in the grid, refer to the dijistra algorithm to calculate the shortest time from the starting point to the edge point in the grid, each time select the node with the shortest time from the starting point from the unselected nodes in the grid, and update the shortest time from the starting point to its neighbor nodes:
其中ai为所选的节点，为该节点的邻居节点，Li,j为两点间的距离，V(i,j,T(ai))为车辆在T(ai)时刻根据交通预测所得到的两节点间道路的平均行驶速度。where ai is the selected node, is the neighbor node of this node, Li,j is the distance between two points, V(i,j,T(ai )) is the average driving speed of the road between two nodes obtained by the vehicle at T(ai ) according to the traffic forecast.
然后将计算完毕的节点ai列为已选择节点，再重复进行下一个离起点所需时间最短的节点的选择，直到所有节点均被迭代完毕或者离起点所需时间最短的节点的值为无穷大，即得到起点到各边缘点的最短时间T。Then list the calculated nodes ai as the selected nodes, and then repeat the selection of the next node with the shortest time from the starting point until all nodes are iterated or the value of the node with the shortest time from the starting point is infinite, that is, the shortest time T from the starting point to each edge point is obtained.
对于各节点到终点的估算距离的计算，对于搜索空间内的节点，我们首先采用分层的方法将整个节点集根据节点到终点之间的节点数进行分层，然后计算每一层的节点到终点的估算距离。具体来说，从终点D开始，将节点集B中的节点进行分层，若节点距离终点D之间至少还存在n个节点，那该节点即属于第n层节点，然后根据静态的道路信息，从第0层节点开始，每次计算第j层中的节点距离终点的估算距离，如下式：For the calculation of the estimated distance from each node to the destination, for the nodes in the search space, we first use a hierarchical method to stratify the entire node set according to the number of nodes between the node and the destination, and then calculate the estimated distance from each layer of nodes to the destination. Specifically, starting from the end point D, the nodes in the node set B are layered. If there are at least n nodes between the node and the end point D, the node belongs to the nth layer node. Then, according to the static road information, starting from the 0th layer node, the estimated distance between the node in the jth layer and the end point is calculated each time, as follows:
其中为第j层中的节点，/>为/>在第j-1层中的邻居节点，根据上述公式最终可以得到各点到终点D的估算距离Q。in is the node in the jth layer, /> for /> For the neighbor nodes in the j-1th layer, the estimated distance Q from each point to the destination D can be finally obtained according to the above formula.
最后根据基于交通预测信息的从起点到边缘节点的最短时间T以及基于静态交通信息的边缘节点到终点的估算距离D可以求解并求的车辆在栅格区域内的具体行驶路径。Finally, according to the shortest time T from the starting point to the edge node based on the traffic forecast information and the estimated distance D from the edge node to the end point based on the static traffic information, the specific driving path of the vehicle in the grid area can be solved and obtained.
为了使本实施例更具直观性并且比较基于交通预测信息的栅格化精细路径规划算法与传统路径规划算法的性能优劣，图5，6，7展示了基于交通预测的栅格化精细路径规划算法和传统最短距离路径规划算法以及依靠实时信息的最短路径算法的平均出行时间，速度，计算时间的比较示意图。传统最短路径规划算法只考虑所选道路的物理距离，选择总距离最短的道路，依靠实时信息的最短路径算法是根据当前的交通状况，计算出按照当前交通状况所需的总体出行时间最短的路径。可以很明显发现基于交通预测的栅格化精细路径规划算法的出行时间明显小于传统最短距离路径规划算法，与依靠实时信息的最短路径算法的出行时间接近甚至略小于依靠实时信息的最短路径算法。基于交通预测的栅格化精细路径规划算法的出行平均速度大于传统路径规划算法和依靠实时信息的最短路径算法，而出行的距离比传统最短距离路径规划要略大。这是因为传统的路径规划是选择距离最短的路径，而不考虑实时交通信息，而依靠实时信息的最短路径算法考虑了当前的交通信息，但不能完全适应瞬息万变的交通状况，而基于交通预测的栅格化精细路径规划算法可以有效预测道路的堵塞状况，避免选择拥堵或即将拥堵的道路，可以寻找到可能行驶距离稍长，但拥堵程度低的道路，达到减少出行时间和提高出行平均速度的效果。另外可以发现基于交通预测的栅格化精细路径规划算法在每天早上5-8点附近时优势特别明显，这是由于该时间段属于由出行低谷到出行高峰的过渡区间，在这个时间段里道路的车流密度变化迅速，一般的路径算法无法适应如此快速变化的城市交通，而基于交通预测的栅格化精细路径规划算法通过预测交通信息的方法可以很好地在快速变化的城市交通路网中找到所需时间更短的道路。In order to make this embodiment more intuitive and compare the performance of the gridded fine path planning algorithm based on traffic forecast information and the traditional path planning algorithm, Figures 5, 6, and 7 show the comparison diagrams of the average travel time, speed, and calculation time of the gridded fine path planning algorithm based on traffic prediction and the traditional shortest distance path planning algorithm, as well as the shortest path algorithm relying on real-time information. The traditional shortest path planning algorithm only considers the physical distance of the selected road and selects the road with the shortest total distance. The shortest path algorithm relying on real-time information calculates the path with the shortest overall travel time according to the current traffic conditions based on the current traffic conditions. It can be clearly found that the travel time of the grid-based fine path planning algorithm based on traffic forecasting is significantly shorter than that of the traditional shortest distance path planning algorithm, and is close to or even slightly shorter than that of the shortest path algorithm relying on real-time information. The average travel speed of the rasterized fine route planning algorithm based on traffic prediction is greater than that of the traditional route planning algorithm and the shortest route algorithm relying on real-time information, while the travel distance is slightly larger than that of the traditional shortest distance route planning algorithm. This is because traditional route planning selects the shortest route without considering real-time traffic information. The shortest route algorithm relying on real-time information takes into account current traffic information, but cannot fully adapt to changing traffic conditions. The grid-based fine route planning algorithm based on traffic forecasting can effectively predict road congestion, avoid choosing congested or imminent congested roads, and find roads that may travel slightly longer but are less congested, thereby reducing travel time and increasing the average travel speed. In addition, it can be found that the grid-based fine route planning algorithm based on traffic prediction has a particularly obvious advantage around 5-8 o'clock every morning. This is because this time period belongs to the transition period from the travel trough to the travel peak. During this time period, the traffic flow density of the road changes rapidly, and the general route algorithm cannot adapt to such a rapidly changing urban traffic. However, the grid-based fine route planning algorithm based on traffic prediction can find a road that takes less time in the rapidly changing urban traffic network by predicting traffic information.
以上所述仅为本发明的较佳实施例而已，并不用以限制本发明，凡在本发明的精神和原则之内所作的任何修改、等同替换和改进等，均应包含在本发明的保护范围之内。The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011252260.XACN112629533B (en) | 2020-11-11 | 2020-11-11 | A Refined Path Planning Method Based on Gridded Road Traffic Flow Prediction |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011252260.XACN112629533B (en) | 2020-11-11 | 2020-11-11 | A Refined Path Planning Method Based on Gridded Road Traffic Flow Prediction |
| Publication Number | Publication Date |
|---|---|
| CN112629533A CN112629533A (en) | 2021-04-09 |
| CN112629533Btrue CN112629533B (en) | 2023-07-25 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011252260.XAActiveCN112629533B (en) | 2020-11-11 | 2020-11-11 | A Refined Path Planning Method Based on Gridded Road Traffic Flow Prediction |
| Country | Link |
|---|---|
| CN (1) | CN112629533B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113449402B (en)* | 2021-06-22 | 2022-08-05 | 武汉大学 | A method for predicting the efficiency gain of the road network after the broken road is opened up |
| CN113651061B (en)* | 2021-08-18 | 2023-04-14 | 上海德启信息科技有限公司 | A monitoring method and monitoring system for logistics flow status |
| CN113848897B (en)* | 2021-09-17 | 2023-12-12 | 宁波大学 | Method, system and related products for path planning for unmanned surface vessels |
| CN114463977B (en)* | 2022-02-10 | 2023-06-23 | 北京工业大学 | A route planning method based on vehicle-road coordination multi-source data fusion traffic flow prediction |
| CN114877901B (en)* | 2022-03-31 | 2024-05-28 | 北京工业大学 | Urban emergency path planning method based on map rasterization fusion and A-star search |
| CN115662114B (en)* | 2022-10-08 | 2024-07-05 | 北京中软政通信息技术有限公司 | Intelligent traffic system for relieving congestion based on big data and operation method thereof |
| CN115630988B (en)* | 2022-12-22 | 2023-10-27 | 北京大学深圳研究生院 | Comprehensive land transportation accessibility measurement method and device |
| CN116132998B (en)* | 2023-03-30 | 2023-07-25 | 江西师范大学 | A City Edge Server Deployment Method Based on Intersection Centrality |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104567906A (en)* | 2015-01-14 | 2015-04-29 | 合肥革绿信息科技有限公司 | Beidou-based urban road network vehicle path planning method and device |
| CN105258704A (en)* | 2014-06-16 | 2016-01-20 | 中国科学院沈阳自动化研究所 | Multi-scale space-time hot point path detection method based on rapid road network modeling |
| CN107084732A (en)* | 2011-02-03 | 2017-08-22 | 通腾发展德国公司 | Method to generate expected average travel speed |
| CN108106623A (en)* | 2017-09-08 | 2018-06-01 | 同济大学 | A kind of unmanned vehicle paths planning method based on flow field |
| CN108877268A (en)* | 2018-08-07 | 2018-11-23 | 南京大学 | One kind is towards unpiloted no traffic lights crossroad intelligent dispatching method |
| WO2019048070A1 (en)* | 2017-09-06 | 2019-03-14 | Swiss Reinsurance Company Ltd. | Electronic system for dynamic, quasi-realtime measuring and identifying driver maneuvers solely based on mobile phone telemetry, and a corresponding method thereof |
| CN109959388A (en)* | 2019-04-09 | 2019-07-02 | 南京大学 | A refined route planning method for intelligent traffic based on grid extension model |
| CN110009906A (en)* | 2019-03-25 | 2019-07-12 | 上海交通大学 | Dynamic route planning method based on traffic prediction |
| CN110046742A (en)* | 2019-02-28 | 2019-07-23 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Perceive planing method of the fluid on feasible path |
| CN110111575A (en)* | 2019-05-16 | 2019-08-09 | 北京航空航天大学 | A kind of Forecast of Urban Traffic Flow network analysis method based on Complex Networks Theory |
| CN110299005A (en)* | 2019-06-10 | 2019-10-01 | 浙江大学 | A kind of city large-scale road network traffic speed prediction technique based on Deep integrating study |
| CN110691406A (en)* | 2019-10-10 | 2020-01-14 | 南京大学 | A D2D Spectrum Efficient Sharing Method for Network-connected Unmanned Vehicle Safe Communication |
| CN111197991A (en)* | 2020-01-15 | 2020-05-26 | 西安电子科技大学 | A method for predicting the best driving path of vehicles based on deep neural network |
| CN111223301A (en)* | 2020-03-11 | 2020-06-02 | 北京理工大学 | A traffic flow prediction method based on graph attention convolutional network |
| CN111337044A (en)* | 2020-03-24 | 2020-06-26 | 北京交通发展研究院 | Urban road path planning method based on traffic weight |
| CN111811525A (en)* | 2020-06-09 | 2020-10-23 | 广东国地规划科技股份有限公司 | Road network generation method and system based on remote sensing image and floating car track |
| JP2020180870A (en)* | 2019-04-25 | 2020-11-05 | 山内 和博 | Automobiles and car driving support systems |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10612930B2 (en)* | 2016-07-26 | 2020-04-07 | Apple Inc. | Normalization of device probe data for path identification |
| US20190004524A1 (en)* | 2016-08-31 | 2019-01-03 | Faraday&Future Inc. | System and method for planning a vehicle path |
| US11755018B2 (en)* | 2018-11-16 | 2023-09-12 | Uatc, Llc | End-to-end interpretable motion planner for autonomous vehicles |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107084732A (en)* | 2011-02-03 | 2017-08-22 | 通腾发展德国公司 | Method to generate expected average travel speed |
| CN105258704A (en)* | 2014-06-16 | 2016-01-20 | 中国科学院沈阳自动化研究所 | Multi-scale space-time hot point path detection method based on rapid road network modeling |
| CN104567906A (en)* | 2015-01-14 | 2015-04-29 | 合肥革绿信息科技有限公司 | Beidou-based urban road network vehicle path planning method and device |
| WO2019048070A1 (en)* | 2017-09-06 | 2019-03-14 | Swiss Reinsurance Company Ltd. | Electronic system for dynamic, quasi-realtime measuring and identifying driver maneuvers solely based on mobile phone telemetry, and a corresponding method thereof |
| CN108106623A (en)* | 2017-09-08 | 2018-06-01 | 同济大学 | A kind of unmanned vehicle paths planning method based on flow field |
| CN108877268A (en)* | 2018-08-07 | 2018-11-23 | 南京大学 | One kind is towards unpiloted no traffic lights crossroad intelligent dispatching method |
| CN110046742A (en)* | 2019-02-28 | 2019-07-23 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Perceive planing method of the fluid on feasible path |
| CN110009906A (en)* | 2019-03-25 | 2019-07-12 | 上海交通大学 | Dynamic route planning method based on traffic prediction |
| CN109959388A (en)* | 2019-04-09 | 2019-07-02 | 南京大学 | A refined route planning method for intelligent traffic based on grid extension model |
| JP2020180870A (en)* | 2019-04-25 | 2020-11-05 | 山内 和博 | Automobiles and car driving support systems |
| CN110111575A (en)* | 2019-05-16 | 2019-08-09 | 北京航空航天大学 | A kind of Forecast of Urban Traffic Flow network analysis method based on Complex Networks Theory |
| CN110299005A (en)* | 2019-06-10 | 2019-10-01 | 浙江大学 | A kind of city large-scale road network traffic speed prediction technique based on Deep integrating study |
| CN110691406A (en)* | 2019-10-10 | 2020-01-14 | 南京大学 | A D2D Spectrum Efficient Sharing Method for Network-connected Unmanned Vehicle Safe Communication |
| CN111197991A (en)* | 2020-01-15 | 2020-05-26 | 西安电子科技大学 | A method for predicting the best driving path of vehicles based on deep neural network |
| CN111223301A (en)* | 2020-03-11 | 2020-06-02 | 北京理工大学 | A traffic flow prediction method based on graph attention convolutional network |
| CN111337044A (en)* | 2020-03-24 | 2020-06-26 | 北京交通发展研究院 | Urban road path planning method based on traffic weight |
| CN111811525A (en)* | 2020-06-09 | 2020-10-23 | 广东国地规划科技股份有限公司 | Road network generation method and system based on remote sensing image and floating car track |
| Title |
|---|
| Big data driven vehicular networks;Cheng, N., Lyu, F., Chen, J., Xu, W., Zhou, H., Zhang, S., & Shen, X;《IEEE Network》;第32卷(第6期);160-167* |
| Deep learning sensor fusion for autonomous vehicle perception and localization: A review;Fayyad, J., Jaradat, M. A., Gruyer, D., & Najjaran, H;《Sensors》;第20卷(第15期);1-4* |
| Toward collision-free and efficient coordination for automated vehicles at unsignalized intersection;Qian B, Zhou H, Lyu F, et al;《IEEE Internet of Things Journal》;第6卷(第6期);10408-10420* |
| 城市道路未来车速预测模型研究;杨马英;楼挺;李一飞;;《小型微型计算机系统》(第12期);190-194* |
| 基于典型经验路径库的路径规划算法;谢海莹;;《交通运输研究》;第2卷(第01期);17-19* |
| 空天地一体化网络技术：探索与展望;焦椤方,许云霆,张天祺等;《物联网学报》;第4卷(第3期);3-6* |
| 车辆导航系统中基于街区分块的分层路网路径规划;张照生;杨殿阁;张德鑫;连小珉;;《中国机械工程》(第23期);3255-3257* |
| Publication number | Publication date |
|---|---|
| CN112629533A (en) | 2021-04-09 |
| Publication | Publication Date | Title |
|---|---|---|
| CN112629533B (en) | A Refined Path Planning Method Based on Gridded Road Traffic Flow Prediction | |
| CN109959388B (en) | Intelligent traffic refined path planning method based on grid expansion model | |
| CN113065074B (en) | Track destination prediction method based on knowledge graph and self-attention mechanism | |
| CN111197991B (en) | A method for predicting the best driving path of vehicles based on deep neural network | |
| CN100442018C (en) | Quasi-dynamic route optimization method for vehicle navigation system with delay risk avoidance | |
| CN107944605A (en) | A kind of dynamic traffic paths planning method based on data prediction | |
| CN114117700B (en) | Research method of urban public transportation network optimization based on complex network theory | |
| CN115409256B (en) | Route recommendation method for avoiding congestion area based on travel time prediction | |
| CN106207290A (en) | A kind of charging electric vehicle aid decision optimization method based on multi-source data | |
| CN108681796B (en) | Urban outside road passenger transport hub site selection method based on POI data and Dijkstra algorithm | |
| CN105489000A (en) | Night-shift bus stop and path selection method | |
| CN112418514B (en) | Method for optimizing campus bus route planning by using ant colony system | |
| CN111879329B (en) | Customized public transport passable shortest path calculation method based on A-x algorithm | |
| CN112991801B (en) | An optimal safe path acquisition method based on time-varying road conditions | |
| CN112330013A (en) | A charging guidance and pricing method for electric vehicles based on dynamic circuit-electric coupling network | |
| CN109579861A (en) | A kind of method for path navigation and system based on intensified learning | |
| Liu et al. | A distributed Markovian parking assist system | |
| CN116194935B (en) | Method and apparatus for determining a navigation profile of a vehicle in a geographic area | |
| CN109087509A (en) | A kind of road grid traffic operating status prediction technique | |
| CN110674990B (en) | Method and system for instant delivery route selection with sliding window update mechanism | |
| CN114842641B (en) | Multi-mode chain traffic distribution method for province domain | |
| Ji-hua et al. | A hierarchical path planning method using the experience of taxi drivers | |
| CN108256662A (en) | The Forecasting Methodology and device of arrival time | |
| CN118262513A (en) | Travel recommendation method based on knowledge graph and traffic speed prediction | |
| CN112183871A (en) | Urban Traffic Guidance System Based on Air Index |
| 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 |