Traffic flow allocation method and system suitable for vehicle-road collaboration in intelligent transportation operation
By introducing weight formulas and real-time monitoring technology into the navigation system, the problem that the existing navigation system ignores the needs and traffic capacity of multiple drivers is solved, and more efficient traffic flow distribution and personalized navigation services are achieved, reducing traffic congestion and Improved driving experience.
Patent Information
- Application Number
- CN202310915156.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-07-25
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2043-07-25
AI Technical Summary
The existing navigation system ignores the destination needs of multiple drivers in a high-density urban road network, resulting in overcrowding of roads, mainly focusing on the shortest or fastest paths and neglecting traffic capacity, which cannot effectively improve road usage efficiency and driving experience. .
A traffic flow distribution method suitable for intelligent traffic operation vehicle-road coordination is adopted, by calculating the weight formula of the road W(ri) = α/(T(ri)+1) + β×C(ri) + γ×D( ri) + σ/(E(ri)+1), where T(ri) is the proportion of road vehicles, C(ri) is the capacity, D(ri) is the driver's demand, and E(ri) is the total demand, Use vehicle Internet of Things and image recognition technology to monitor road conditions in real time to generate the optimal path.
It improves road usage efficiency, optimizes driver's driving path, reduces traffic congestion, provides personalized navigation services, and adaptively adjusts parameters to cope with different road conditions and needs.
Smart Images

Figure CN117173911B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent transportation, and more specifically, to a traffic flow allocation method and system suitable for vehicle-road collaboration in intelligent transportation operation. Background Art
[0002] Navigation services play a vital role in modern transportation systems. By providing optimal driving routes, navigation services can help drivers plan their trips efficiently, avoid traffic jams, and save time and fuel. However, existing navigation services usually have some problems, especially in high-density urban road networks.
[0003] First, existing navigation systems mainly focus on the driving needs of a single driver and often ignore the needs of other drivers. Although navigation systems take into account the real-time traffic conditions of the road, such as traffic volume, traffic speed, etc., they usually do not fully consider the destination needs of all drivers. In other words, when multiple drivers' destinations are in the same area, existing navigation systems may provide them with the same or similar driving routes, which may cause some roads to be too crowded and cause traffic congestion.
[0004] Secondly, existing navigation systems mainly consider the shortest driving distance or the fastest driving time when providing driving routes for drivers. However, this is not always the best choice. For example, if a road is already very congested, it may take a long time for drivers to pass even if it is the shortest route. Similarly, if a road has poor traffic capacity, such as the road has accidents, is narrow or has potholes, drivers may feel uncomfortable even if it is the fastest route.
[0005] Therefore, existing navigation services often seem to be unable to cope with large-scale and complex urban traffic problems. In order to improve the efficiency of road use, reduce traffic congestion, and enhance the driving experience of drivers, a new navigation technology solution is needed. This solution can comprehensively consider the needs of all drivers, as well as the real-time conditions and traffic capacity of the road to provide a better driving route. Summary of the invention
[0006] The technical problem to be solved by the present invention is to provide a traffic flow allocation method and system suitable for vehicle-road collaboration in intelligent traffic operation, so as to solve the problems mentioned in the background technology.
[0007] In order to achieve the above object, the present invention adopts the following technical solutions:
[0008] A traffic flow allocation method suitable for vehicle-road collaboration in intelligent transportation operation includes the following steps:
[0009] S1: Calculate the weight of road ri for any driver A:
[0010] W(ri)=α / (T(ri)+1)+β×C(ri)+γ×D(ri)+σ / (E(ri)+1)
[0011] in:
[0012] ri represents the i-th road;
[0013] T(ri) is the ratio of the actual number of vehicles on road ri to the number of vehicles that can be placed on road ri;
[0014] C(ri) represents the traffic capacity of road ri itself. The better the traffic capacity, the larger C(ri) is.
[0015] D(ri) represents the degree of demand for road ri for driver A. The higher the demand, the greater D(ri);
[0016] E(ri) is the superposition of the demand degree D(ri) of road ri for all drivers;
[0017] α, β, γ, and σ are constants greater than 0;
[0018] The calculation process of D(ri) corresponding to any driver is as follows:
[0019] S1.1: Collect the driver’s current location and destination information through the vehicle IoT system;
[0020] S1.2: Generate multiple possible navigation routes from the current location to the destination through a navigation algorithm;
[0021] S1.3: For each generated navigation route, determine whether it passes through the road ri and calculate D(ri) using the following formula:
[0022]
[0023] Among them, m is the number of navigation routes generated, Match j (ri) represents the indicator function. If the jth navigation route passes through road ri, take Match j (ri) = 1, if the jth navigation route does not pass through road ri, take Match j (ri) = 0; s j represents the length of the j-th navigation route;
[0024] Among them, α, β, γ, σ are specially calculated for road ri and driver A, specifically: collect historical data H = {(T1, C1, D1, E1), (T2, C2, D2, E2), ... (Tk, Ck, Dk, Ek) ..., (Tn, Cn, Dn, En)}, where (Tk, Ck, Dk, Ek) represents the kth group of data in multiple groups (T(ri), C(ri), D(ri), E(ri)) data corresponding to road ri and driver A, calculate ((Tk+1), 1 / Ck, 1 / Dk, (Ek+1)) for each group of data and set it equal to (αk, βk, γk, σk), calculate the average value of αk from 1 to n and use it as α, calculate the average value of βk from 1 to n and use it as β, calculate the average value of γk from 1 to n and use it as γ, calculate the average value of σk from 1 to n and use it as σ;
[0025] S2: Consider each road ri as an edge and each intersection as a node, and construct a navigation route map. The weight of edge ri is W(ri) corresponding to driver A calculated in step S1; then, the following algorithm is used to find the optimal path from the starting point to the end point of the driver and send it to driver A:
[0026] S2.1: Mark a shortest distance value for each node, set the shortest distance value of the starting point to 0, and set the shortest distance values of all other nodes to infinity;
[0027] S2.2: Mark all nodes and all edges as unprocessed;
[0028] S2.3: Select a node N from the unprocessed nodes according to the shortest distance value;
[0029] S2.4: For each unprocessed adjacent edge H of node N, calculate W(H)+R(N). If W(H)+R(N)<R(M), then update the shortest distance of M to W(H)+R(N), and mark edge H and node N as processed; where W(H) is the weight of H, R(N) is the shortest distance of N, and M is the other end point of H relative to N; if there is no unprocessed adjacent edge for node N, directly mark N as processed;
[0030] S2.5: Repeat steps S2.3 and S2.4 until the endpoint is marked as processed;
[0031] S2.6: Starting from the starting point, find the points marked as processed in chronological order and connect them as the shortest path from the starting point to the end point as the optimal path.
[0032] Preferably, the calculation process of T(ri) includes the following steps:
[0033] a. Set up multiple camera devices along the road ri to monitor the road ri in real time;
[0034] b. Use image recognition technology to identify the number of vehicles in the monitoring image of each camera device;
[0035] c. Add the number of vehicles monitored by each camera device to obtain the actual number of vehicles on road ri;
[0036] d. Calculate the number of vehicles that can be placed on the road ri, including measuring the width and length of the road ri, and then estimating the maximum number of vehicles that can be placed based on the standard vehicle size;
[0037] e. Divide the actual number of vehicles obtained in step c by the number of vehicles that can be placed in step d to obtain T(ri).
[0038] Preferably, C(ri) is obtained by the following steps:
[0039] Multiple camera devices are set up along the road ri to monitor the road ri in real time;
[0040] Using image recognition technology, identify road obstacles in the monitoring image of each camera device, including at least road obstacles caused by construction, accidents, and landslides;
[0041] Calculate the lateral length of the lane occupied by each obstacle. If the lateral lengths of multiple obstacles are connected, treat these multiple obstacles as one obstacle and calculate their total lateral lengths. Finally, the lateral lengths of N obstacles L1 to LN are obtained.
[0042] Measure the total lateral length LT of the road ri;
[0043] C(ri) is calculated according to the following formula:
[0044]
[0045] Where N is the number of obstacles and Lj is the lateral length of j obstacles.
[0046] The present invention also discloses a traffic flow distribution system suitable for vehicle-road collaboration in intelligent traffic operation, comprising:
[0047] (1) A vehicle number detection module is used to detect and calculate the actual number of vehicles on each road ri through sensor devices installed on the road, and calculate the ratio T(ri) of the actual number of vehicles on ri to the number of vehicles that can be placed on the road;
[0048] (2) Road status monitoring module, which monitors road ri in real time through multiple camera devices arranged along road ri, and uses image recognition technology to identify road obstacles in the monitoring image of each camera device, including construction, accidents, and landslides, and calculates C(ri) based on the recognition results using the above-mentioned traffic flow allocation method applicable to intelligent transportation operation vehicle-road collaboration;
[0049] (3) a demand calculation module, which is used to collect the driver's current location and destination information through the vehicle Internet of Things system, generate multiple possible navigation routes from the current location to the destination through a navigation algorithm, and determine whether each route passes through the road ri, so as to calculate D(ri) and E(ri) according to the above-mentioned traffic flow allocation method applicable to vehicle-road collaboration in intelligent transportation operation;
[0050] (4) a weight calculation module, used to calculate the weight of each road ri according to the calculated T(ri), C(ri) and D(ri), according to the formula W(ri) = α / (T(ri)+1)+β×C(ri)+γ×D(ri)+σ / (E(ri)+1), where α, β, γ, and σ are constants greater than 0 calculated according to the above-mentioned traffic flow allocation method applicable to intelligent traffic operation vehicle-road collaboration;
[0051] (5) a path planning module, which is used to regard each road ri as an edge and each intersection as a node, and construct a navigation route map. The weight of the edge ri is W(ri) calculated by the weight calculation module, and then find the optimal path from the starting point to the end point of the driver according to the method described in S2 above, and send it to the driver;
[0052] The vehicle quantity detection module, road status monitoring module, demand calculation module, weight calculation module, and path planning module are interconnected and communicated with each other.
[0053] Preferably, the vehicle quantity detection module includes a camera device and an image processing algorithm center, the camera device is used to capture real-time images of the road, and the image processing algorithm center is used to identify and calculate the number of vehicles.
[0054] Preferably, the road status monitoring module includes a camera device and a deep learning processing unit, the camera device is used to capture real-time images of the road, and the deep learning processing unit is used to implement image recognition functions, identify road obstacles and calculate the road's traffic capacity.
[0055] Preferably, the demand calculation module, the weight calculation module, and the path planning module are all arranged on a high-performance server and a cloud computing platform to process real-time data and quickly complete complex computing tasks.
[0056] The advantages of the present invention over the prior art are:
[0057] Improve road use efficiency:
[0058] Through the weight formula W(ri)=α / (T(ri)+1)+β×C(ri)+γ×D(ri)+σ / (E(ri)+1), the present invention takes into account both the specific needs of each driver and the overall road conditions and the needs of the driver. The formula includes the actual number of vehicles on the road T(ri), the road's own traffic capacity C(ri), the driver's demand level D(ri), and the total demand of all drivers for the road E(ri). In this way, road network resources can be used more effectively and the efficiency of road use can be improved.
[0059] Optimizing the driver's driving path: When solving the driver's optimal path, the present invention adopts a weight-based shortest path algorithm. This algorithm calculates weights according to each driver's needs and road conditions, and then determines the optimal path based on these weights. In this way, each driver can be provided with the most suitable driving path for them, improving the driver's driving experience.
[0060] Reduce traffic congestion: The present invention introduces a decay factor 1 / (E(ri)+1) into the weight formula, which means that if the total demand E(ri) of a road is high, the weight of this road will be reduced. This can guide drivers to avoid roads with high demand, thereby reducing the congestion of these roads and reducing traffic congestion.
[0061] Adaptive adjustment parameters: The weight formula of the present invention includes four parameters α, β, γ, σ, which are automatically adjusted according to historical data. This allows the solution of the present invention to adaptively respond to different road conditions and driver needs, further improving the efficiency of road network use.
[0062] Providing personalized navigation services: Through the refined calculation of each driver's demand D(ri) and the demand-based shortest path algorithm, the present invention can provide each driver with personalized navigation services. This can not only improve the driver's driving experience, but also improve the user satisfaction of the navigation system. BRIEF DESCRIPTION OF THE DRAWINGS
[0063] Figure 1 It is a schematic diagram of the overall method flow of the present invention;
[0064] Figure 2 It is a schematic diagram of calculating road weights of the present invention;
[0065] Figure 3 It is a schematic diagram of the optimal path algorithm of the present invention. DETAILED DESCRIPTION
[0066] The specific implementation of the present invention will be described below in conjunction with the accompanying drawings.
[0067] Embodiment 1 of the present invention, the specific design of the system of the present invention is as follows:
[0068] (1) Vehicle quantity detection module:
[0069] The present invention first selects appropriate sensor equipment, such as cameras, magnetic induction or infrared vehicle detection sensors, which will be installed on the curbs on both sides of the road or the isolation belt in the middle of the road. For places that are difficult to install directly, such as bridges or elevated roads, they can be installed on fixed traffic signs or street light poles. The installation location needs to ensure that the sensor can cover all lanes so as to accurately calculate the number of vehicles at each moment. These sensors send data to the central processing unit wirelessly or wiredly, and the central processing unit will calculate the number of vehicles on the lane in real time and use the data to calculate T(ri).
[0070] (2) Road status monitoring module:
[0071] Multiple high-resolution camera devices can be arranged along the road ri. These cameras should be installed on street lamp poles or special monitoring poles to cover the entire road width as much as possible to ensure that there are no blind spots. The camera transmits image data to the deep learning processing unit in real time through a high-speed optical fiber network. The deep learning processing unit will use deep learning technology to analyze the image, identify various road obstacles including construction, accidents, and landslides, and calculate C(ri) based on the recognition results.
[0072] (3) Demand calculation module:
[0073] The present invention collects the driver's current location and destination information through the vehicle Internet of Things system, which can be achieved through the vehicle GPS system. After obtaining this information, the module will use the existing navigation algorithm (such as the existing Dijkstra algorithm) to generate multiple possible navigation routes, and then determine whether each route passes through the road ri, and calculate D(ri) and E(ri).
[0074] (4) Weight calculation module:
[0075] This module is responsible for calculating the weight of each road ri. It calculates W(ri) based on T(ri), C(ri), and D(ri), where α, β, γ, and σ are constants greater than 0 calculated from historical data using the coefficient calculation algorithm described later. This module can be implemented on a high-performance computing server.
[0076] (5) Path planning module:
[0077] This module performs path planning based on the results of the weight calculation module. It first considers each road ri as an edge and each intersection as a node to construct a navigation route map, and then uses the subsequent optimal path algorithm of the present invention for the above weights to find the optimal path from the starting point to the end point, and sends it to the driver. This step can be completed through the vehicle information system or mobile phone APP.
[0078] In order to achieve communication between the vehicle and the module, the driver's device (such as the vehicle information system or mobile phone APP) needs to install the corresponding software to receive the results of the path planning module and give the driver clear navigation instructions.
[0079] All these modules are interconnected and communicated through high-speed networks to transmit and process data in real time. Considering the real-time requirements of the system, the network should adopt high-speed, low-latency communication technologies such as optical fiber or 5G. Data centers need to have sufficient storage and computing power to process large amounts of image data and execute complex algorithms.
[0080] In actual operation, the system also needs to consider fault recovery and redundancy design. For example, backup sensors and camera devices can be deployed in key locations to prevent data loss caused by equipment failure. At the same time, data centers also need to have backup and failover capabilities to ensure service continuity and data integrity.
[0081] In addition, the system also needs a management and maintenance module to monitor the system's operating status, including device status, network status, data processing efficiency, etc., as well as to regularly maintain and update the system. This module can be completed by both manual and automated tools.
[0082] So far, the present invention has completed the design of an intelligent traffic management system including a vehicle quantity detection module, a road state monitoring module, a demand calculation module, a weight calculation module, and a path planning module. Through this system, the method of the present invention under the following embodiment 2 can be used to achieve intelligent management of urban road traffic, improve road utilization, reduce traffic congestion, and improve the travel efficiency of drivers.
[0083] In Example 2, the method of the present invention is described. Figures 1 to 3 Shown are schematic diagrams of the method of the present invention, wherein Figure 1 It is a schematic diagram of the overall method flow of the present invention; Figure 2 It is a schematic diagram of calculating road weights of the present invention; Figure 3 It is a schematic diagram of the optimal path algorithm of the present invention.
[0084] The method of the present invention specifically comprises the following steps:
[0085] S1: Calculate the weight of road ri for any driver A:
[0086] W(ri)=α / (T(ri)+1)+β×C(ri)+γ×D(ri)+σ / (E(ri)+1)
[0087] in:
[0088] ri represents the i-th road;
[0089] T(ri) is the ratio of the actual number of vehicles on road ri to the number of vehicles that can be placed on road ri;
[0090] C(ri) represents the traffic capacity of road ri itself. The better the traffic capacity, the larger C(ri) is.
[0091] D(ri) represents the degree of demand for road ri for driver A. The higher the demand, the greater D(ri);
[0092] E(ri) is the superposition of the demand degree D(ri) of road ri for all drivers;
[0093] α, β, γ, and σ are constants greater than 0;
[0094] The calculation process of D(ri) corresponding to any driver is as follows:
[0095] S1.1: Collect the driver’s current location and destination information through the vehicle IoT system;
[0096] S1.2: Generate multiple possible navigation routes from the current location to the destination through a navigation algorithm;
[0097] S1.3: For each generated navigation route, determine whether it passes through the road ri and calculate D(ri) using the following formula:
[0098]
[0099] Among them, m is the number of navigation routes generated, Match j (ri) represents the indicator function. If the jth navigation route passes through road ri, take Match j (ri) = 1, if the jth navigation route does not pass through road ri, take Match j (ri) = 0; s j represents the length of the j-th navigation route;
[0100] The calculation process of T(ri) includes the following steps:
[0101] a. Set up multiple camera devices along the road ri to monitor the road ri in real time;
[0102] b. Use image recognition technology to identify the number of vehicles in the monitoring image of each camera device;
[0103] c. Add the number of vehicles monitored by each camera device to obtain the actual number of vehicles on road ri;
[0104] d. Calculate the number of vehicles that can be placed on the road ri, including measuring the width and length of the road ri, and then estimating the maximum number of vehicles that can be placed based on the standard vehicle size;
[0105] e. Divide the actual number of vehicles obtained in step c by the number of vehicles that can be placed in step d to obtain T(ri).
[0106] Among them, C(ri) is obtained by the following steps:
[0107] Multiple camera devices are set up along the road ri to monitor the road ri in real time;
[0108] Using image recognition technology, identify road obstacles in the monitoring image of each camera device, including at least road obstacles caused by construction, accidents, and landslides;
[0109] Calculate the lateral length of the lane occupied by each obstacle (perpendicular to the direction of vehicle travel). If the lateral lengths of multiple obstacles are connected (for example, if two cars are connected horizontally and occupy the same lane in an accident, the two cars are regarded as one obstacle, and the lateral length is the length of the lane), then these multiple obstacles are regarded as one obstacle and their total lateral length is calculated; finally, the lateral lengths of N obstacles L1~LN are obtained;
[0110] Measure the total lateral length LT of the road ri;
[0111] C(ri) is calculated according to the following formula:
[0112]
[0113] Where N is the number of obstacles and Lj is the lateral length of j obstacles.
[0114] In addition, α, β, γ, and σ are calculated specifically for road ri and driver A. Specifically, historical data H = {(T1, C1, D1, E1), (T2, C2, D2, E2), ... (Tk, Ck, Dk, Ek) ..., (Tn, Cn, Dn, En)} are collected, where (Tk, Ck, Dk, Ek) represents the kth group of data in the multiple groups (T(ri), C(ri), D(ri), E(ri)) corresponding to road ri and driver A, and each group of data is calculated separately. ((Tk+1), 1 / Ck, 1 / Dk, (Ek+1)) and let it be equal to (αk, βk, γk, σk), calculate the mean of αk from 1 to n and use it as α, calculate the mean of βk from 1 to n and use it as β, calculate the mean of γk from 1 to n and use it as γ, calculate the mean of σk from 1 to n and use it as σ; the choice of this coefficient can make each item in the weight tend to the same value after multiplying the coefficient, thereby avoiding too large a proportion of a certain item, and thus balancing the contribution of each item to the weight.
[0115] S2: Consider each road ri as an edge and each intersection as a node, and construct a navigation route map. The weight of edge ri is W(ri) corresponding to driver A calculated in step S1; then, the following algorithm is used to find the optimal path from the starting point to the end point of the driver and send it to driver A:
[0116] S2.1: Mark a shortest distance value for each node, set the shortest distance value of the starting point to 0, and set the shortest distance values of all other nodes to infinity;
[0117] S2.2: Mark all nodes and all edges as unprocessed;
[0118] S2.3: Select a node N from the unprocessed nodes according to the shortest distance value;
[0119] S2.4: For each unprocessed adjacent edge H of node N, calculate W(H)+R(N). If W(H)+R(N)<R(M), then update the shortest distance of M to W(H)+R(N), and mark edge H and node N as processed; where W(H) is the weight of H, R(N) is the shortest distance of N, and M is the other end point of H relative to N; if there is no unprocessed adjacent edge for node N, directly mark N as processed;
[0120] S2.5: Repeat steps S2.3 and S2.4 until the endpoint is marked as processed;
[0121] S2.6: Starting from the starting point, find the points marked as processed in chronological order and connect them as the shortest path from the starting point to the end point as the optimal path.
[0122] In the following embodiment 3, an example is used to illustrate the optimal path search after the weight is calculated in the present invention.
[0123] Assume that the present invention has 5 nodes, labeled A, B, C, D and E, with the starting point being A and the end point being E. The following are the connections between these nodes and the corresponding weights (the distance of the edge mentioned later refers to the weight):
[0124] The weight from A to B is 4;
[0125] The weights of A to C are 2;
[0126] The weight from B to C is 1;
[0127] The weight of B to D is 5;
[0128] The weight of C to D is 8;
[0129] The weight of C to E is 10;
[0130] The weight of D to E is 2;
[0131] The present invention seeks the shortest path from A to E. Initially, the present invention sets the shortest distance of A to 0, and the shortest distances of all other nodes to infinity. Therefore, the present invention has:
[0132] A's shortest distance: 0;
[0133] The shortest distance to B: infinity;
[0134] The shortest distance of C: infinity;
[0135] The shortest distance of D: infinity;
[0136] The shortest distance of E: infinity;
[0137] Then the present invention starts the algorithm:
[0138] The present invention selects the unprocessed node with the smallest shortest distance, that is, A. For each adjacent edge of A (AB and AC), the total distance from A to the other end node is calculated (the shortest distance of A plus the distance from A to the other end point). For AB, the total distance is 0+4; for AC, the total distance is 0+2. Because both total distances are less than the current shortest distance between B and C (infinity), the present invention updates the shortest distance between B and C. Then the present invention marks A as processed.
[0139] After update:
[0140] B's shortest distance: 4;
[0141] The shortest distance of C: 2;
[0142] Next, the present invention selects the node with the shortest distance among the unprocessed nodes, which is C. Looking at the adjacent edges of C, there are only CD and CE, because AC has been processed before. For CD, the total distance from A through C to D is 2 (the distance from A to C) + 8 (the weight from C to D) = 10, which is greater than the current shortest distance of D (infinity), so the present invention updates the shortest distance of D. For CE, the total distance from A through C to E is 2 (the distance from A to C) + 10 (the weight from C to E) = 12, which is also less than the current shortest distance of E (infinity), so the present invention updates the shortest distance of E. Then the present invention marks C as processed.
[0143] After update:
[0144] Minimum distance of D: 10;
[0145] Minimum distance of E: 12;
[0146] Next, the present invention selects B because B has the smallest shortest distance. Then, the present invention checks B's adjacent edges BC and BD because AB has been processed before. For BC, the total distance from A through B to C is 4 (distance from A to B) + 1 (weight from B to C) = 5, but this is greater than the current shortest distance of C, 2, so the present invention does not update the shortest distance of C. For BD, the total distance from A through B to D is 4 (distance from A to B) + 5 (weight from B to D) = 9, which is less than the current shortest distance of D, 10, so the present invention updates the shortest distance of D. The present invention then marks B as processed.
[0147] After update:
[0148] Shortest distance of D: 9
[0149] Now, select the node with the shortest distance among the unprocessed nodes again, which is D. D has only one unprocessed adjacent edge, which is DE. The total distance from A to E via D is 9 (the distance from A to D) + 2 (the weight from D to E) = 11, which is less than the current shortest distance of E, 12, so the present invention updates the shortest distance of E. Then the present invention marks D as processed.
[0150] After update:
[0151] Minimum distance of E: 11;
[0152] Finally, the present invention has only one unprocessed node left, which is E. Since E has no unprocessed adjacent edges, the present invention directly marks E as processed.
[0153] Now all nodes have been processed, and A, B, D, and E are marked as processed in chronological order, so the present invention finds the shortest path from A to E: A->B->D->E.
[0154] This algorithm can be used to find the shortest path from the starting point to all other nodes, and its time complexity is relatively low (O(V^2), V is the number of nodes, if a priority queue or heap is used, the time complexity can be further reduced to O(P+V log V), P is the number of edges), which makes it very useful in practical applications, especially in transportation networks, telecommunication networks, etc.
[0155] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.
Claims
1. A traffic flow allocation method suitable for vehicle-road collaboration in intelligent transportation operation, characterized in that: The steps include: S1: Calculate the weight of road ri for any driver A: W(ri)=α / (T(ri)+1)+β×C(ri)+γ×D(ri)+σ / (E(ri)+1) in: ri represents the i-th road; T(ri) is the ratio of the actual number of vehicles on road ri to the number of vehicles that can be placed on road ri; C(ri) represents the traffic capacity of road ri itself. The better the traffic capacity, the larger C(ri) is. D(ri) represents the degree of demand for road ri for driver A. The higher the demand, the greater D(ri); E(ri) is the superposition of the demand degree D(ri) of road ri for all drivers; α, β, γ, and σ are constants greater than 0; The calculation process of D(ri) corresponding to any driver is as follows: S1.1: Collect the driver’s current location and destination information through the vehicle IoT system; S1.2: Generate multiple possible navigation routes from the current location to the destination through a navigation algorithm; S1.3: For each generated navigation route, determine whether it passes through road ri and calculate D(ri) using the following formula: Among them, m is the number of navigation routes generated, Match j (ri) represents the indicator function. If the jth navigation route passes through road ri, take Match j (ri) = 1, if the jth navigation route does not pass through road ri, take Match j (ri) = 0; s j represents the length of the j-th navigation route; Among them, α, β, γ, σ are specially calculated for road ri and driver A, specifically: collect historical data H = {(T1, C1, D1, E1), (T2, C2, D2, E2), ... (Tk, Ck, Dk, Ek) ..., (Tn, Cn, Dn, En)}, where (Tk, Ck, Dk, Ek) represents the kth group of data in multiple groups (T(ri), C(ri), D(ri), E(ri)) data corresponding to road ri and driver A, calculate ((Tk+1), 1 / Ck, 1 / Dk, (Ek+1)) for each group of data and set it equal to (αk, βk, γk, σk), calculate the average value of αk from 1 to n and use it as α, calculate the average value of βk from 1 to n and use it as β, calculate the average value of γk from 1 to n and use it as γ, calculate the average value of σk from 1 to n and use it as σ; S2: Consider each road ri as an edge and each intersection as a node, and construct a navigation route map. The weight of edge ri is W(ri) corresponding to driver A calculated in step S1; then, the following algorithm is used to find the optimal path from the starting point to the end point of the driver and send it to driver A: S2.1: Mark a shortest distance value for each node, set the shortest distance value of the starting point to 0, and set the shortest distance values of all other nodes to infinity; S2.2: Mark all nodes and all edges as unprocessed; S2.3: Select a node N from the unprocessed nodes according to the shortest distance value; S2.4: For each unprocessed adjacent edge H of node N, calculate W(H)+R(N). If W(H)+R(N)<R(M), then update the shortest distance of M to W(H)+R(N), and mark edge H and node N as processed; where W(H) is the weight of H, R(N) is the shortest distance of N, and M is the other end point of H relative to N; if there is no unprocessed adjacent edge for node N, directly mark N as processed; S2.5: Repeat steps S2.3 and S2.4 until the endpoint is marked as processed; S2.6: Starting from the starting point, find the points marked as processed in chronological order and connect them as the shortest path from the starting point to the end point as the optimal path.
2. The traffic flow allocation method for intelligent traffic operation vehicle-road collaboration according to claim 1 is characterized in that: The calculation process of T(ri) includes the following steps: a. Set up multiple camera devices along the road ri to monitor the road ri in real time; b. Use image recognition technology to identify the number of vehicles in the monitoring image of each camera device; c. Add the number of vehicles monitored by each camera device to obtain the actual number of vehicles on road ri; d. Calculate the number of vehicles that can be placed on the road ri, including measuring the width and length of the road ri, and then estimating the maximum number of vehicles that can be placed based on the standard vehicle size; e. Divide the actual number of vehicles obtained in step c by the number of vehicles that can be placed in step d to obtain T(ri).
3. The traffic flow allocation method for vehicle-road collaboration in intelligent transportation operation according to claim 1 is characterized in that: C(ri) is obtained by the following steps: Multiple camera devices are set up along the road ri to monitor the road ri in real time; Using image recognition technology, identify road obstacles in the monitoring image of each camera device, including at least road obstacles caused by construction, accidents, and landslides; Calculate the lateral length of the lane occupied by each obstacle. If the lateral lengths of multiple obstacles are connected, treat these multiple obstacles as one obstacle and calculate their total lateral lengths. Finally, the lateral lengths of N obstacles L1 to LN are obtained. Measure the total lateral length LT of the road ri; C(ri) is calculated according to the following formula: Where N is the number of obstacles and Lj is the lateral length of j obstacles.
4. A traffic flow distribution system suitable for intelligent traffic operation vehicle-road collaboration, characterized in that: include: (1) A vehicle number detection module is used to detect and calculate the actual number of vehicles on each road ri through sensor devices installed on the road, and calculate the ratio T(ri) of the actual number of vehicles on ri to the number of vehicles that can be placed on the road; (2) a road status monitoring module, which monitors the road ri in real time through multiple camera devices arranged along the road ri, identifies road obstacles in the monitoring image of each camera device through image recognition technology, including construction, accidents, and landslides, and calculates C(ri) according to claim 1 based on the recognition results; (3) a demand calculation module, used to collect the driver's current location and destination information through the vehicle Internet of Things system, generate multiple possible navigation routes from the current location to the destination through a navigation algorithm, and determine whether each route passes through the road ri, so as to calculate D(ri) and E(ri) according to the method of claim 1; (4) a weight calculation module, used to calculate the weight of each road ri according to the calculated T(ri), C(ri) and D(ri) according to the formula W(ri)=α / (T(ri)+1)+β×C(ri)+γ×D(ri)+σ / (E(ri)+1), where α, β, γ, and σ are constants greater than 0 calculated from historical data according to the method of claim 1; (5) a path planning module, which is used to regard each road ri as an edge and each intersection as a node, to construct a navigation route map, the weight of the edge ri is W(ri) calculated by the weight calculation module, and then find the optimal path from the starting point to the end point of the driver according to the method described in S2 of claim 1, and send it to the driver; The vehicle quantity detection module, road status monitoring module, demand calculation module, weight calculation module, and path planning module are interconnected and communicated with each other.
5. The system according to claim 4, characterized in that: The vehicle quantity detection module includes a camera device and an image processing algorithm center, wherein the camera device is used to capture real-time images of the road, and the image processing algorithm center is used to identify and calculate the number of vehicles.
6. The system according to claim 4, characterized in that: The road status monitoring module includes a camera device and a deep learning processing unit. The camera device is used to capture real-time images of the road, and the deep learning processing unit is used to implement image recognition functions, identify road obstacles and calculate the road's traffic capacity.
7. The traffic flow distribution system of intelligent traffic operation vehicle-road cooperation according to claim 4 is characterized in that: The demand calculation module, the weight calculation module, and the path planning module are all arranged on a high-performance server and a cloud computing platform, and are used to process real-time data and quickly complete complex computing tasks.
