Road traffic flow monitoring method, device and system using AI platform

Abstract

The invention relates to the technical field of traffic, in particular to a road traffic flow monitoring method, device and system using an AI platform, and the method comprises the steps: combining the traffic flow of each subordinate road connected with each road at each sampling moment according to the traffic flow change difference of different roads at continuous sampling moments, and calculating the traffic flow of each subordinate road at each sampling moment; determining the contribution weight of each subordinate road connected with each road at each sampling moment, then determining the traffic flow change degree of each road at each sampling moment, and determining the traffic flow prediction weight of each road at each sampling moment by combining the traffic flow difference at different monitoring points of each road at each sampling moment; and thus, the predicted traffic flow at a plurality of future moments is obtained. According to the invention, traffic flow data is accurately predicted, traffic routes are planned and adjusted in advance, vehicles are guided to avoid congested road sections in peak periods, and balanced distribution of traffic flows is realized.

Description

Technical Field

[0001] The present invention relates to the field of traffic technology, and in particular to a road traffic flow monitoring method, device and system using an AI platform. Background Art

[0002] With the acceleration of urbanization and the continued growth of motor vehicle ownership, road traffic congestion, frequent accidents, and inefficient management are becoming increasingly prominent. Traditional traffic flow monitoring, which primarily relies on fixed cameras, geomagnetic coils, or manual inspections, struggles to continuously monitor traffic density and driving trajectories across the entire road network. Furthermore, response times are delayed, and manually scheduled traffic light control strategies cannot dynamically adapt to sudden changes in traffic flow during peak hours. Intelligently monitoring traffic conditions through an AI platform integrates multi-source sensing devices, optimizes an edge-cloud collaborative computing architecture, and introduces adaptive learning mechanisms to achieve comprehensive, all-time intelligent monitoring of road traffic flow.

[0003] Existing problem: When monitoring road traffic flow, because traffic flow changes in real time, traffic flow changes in different areas and main roads are different, and traffic conditions are affected by local factors (such as traffic accidents, traffic control, etc.), it is inaccurate to judge the changes in vehicle traffic flow based solely on traffic monitoring at fixed points. Summary of the Invention

[0004] The present invention provides a road traffic flow monitoring method, device and system using an AI platform to solve existing problems.

[0005] The present invention provides a method, device, and system for monitoring road traffic flow using an AI platform, employing the following technical solutions: One embodiment of the present invention provides a method for monitoring road traffic flow using an AI platform, the method comprising the following steps: At each sampling moment, the traffic volume of each road, the traffic volume at each monitoring point on each road, and the traffic volume entering from one road to another connected road are obtained; each road connects several upper-level roads and lower-level roads; Based on the difference in traffic volume changes of different roads at consecutive sampling moments, combined with the traffic volume of each subordinate road connected to each road at each sampling moment, the contribution weight of each subordinate road connected to each road at each sampling moment is determined; Determine the traffic flow variation of each road at each sampling moment based on the contribution weight, the traffic flow from each lower-level road connected to each road at each sampling moment, and the number of all upper-level roads connected to each road; Based on the traffic flow variability and the difference in traffic flow at different monitoring points on each road at each sampling moment, the traffic flow prediction weight of each road at each sampling moment is determined; the traffic flow prediction weight and traffic flow of each road at all sampling moments are used as the input of the prediction model, and the predicted traffic flow at several future moments is output.

[0006] Furthermore, the determination of the contribution weight of each subordinate road connected to each road at each sampling moment includes the following specific steps: Preset first constant , get from From the sampling moment to the All sampling moments within the period of the sampling moment The traffic flow sequence composed of the traffic flow of the roads; In the traffic flow sequence, calculate the Traffic volume minus the The difference between the traffic flows of each vehicle is used to obtain the normalized value of the sum of the differences of all adjacent traffic flows, which is recorded as At the sampling moment The traffic volume variation coefficient of the road; According to The difference in traffic flow variation coefficients of different roads at the sampling time is used to determine the Traffic flow influencing factors of any two roads at a sampling time; Get from From the sampling moment to the All sampling moments within the period of the sampling moment The road connecting The sum of the traffic volumes of the subordinate roads is recorded as the first sum; The first At the sampling moment The road connecting The lower level road and the The normalized value of the product of the traffic flow impact factor of the road and the first sum value is recorded as At the sampling moment The road connecting The contribution weight of the subordinate roads.

[0007] Furthermore, the determination The traffic flow influencing factors of any two roads at a sampling time include the following specific steps: Get the The average of the traffic flow variation coefficients of all roads at the sampling time is recorded as the overall traffic flow variation coefficient; In from From the sampling moment to the During the period of each sampling moment, the inversely proportional normalized value of the absolute value of the difference between the traffic flow variation coefficients of any two roads at each sampling moment is obtained, which is recorded as the similarity value at each sampling moment. The sum of the similarity values ​​at all sampling moments is recorded as the comprehensive similarity value. The normalized value of the product of the comprehensive similarity value and the overall traffic flow variation coefficient is recorded as the first The traffic flow influencing factor of any two roads at a sampling time.

[0008] Furthermore, the determination of the traffic flow variation degree of each road at each sampling moment includes the following specific steps: Get the The normalized value of the number of all superior roads connected by a road is recorded as the target number; Get the At the sampling moment The road connecting The contribution weight of the lower-level road is the same as that of the At the sampling moment, The road connecting The lower level road enters the connected The product of the traffic volume of the roads is recorded as At the sampling moment The road connecting The first product of the lower-level roads, the At the sampling moment The sum of the first products of all subordinate roads connected by the road is recorded as the second sum; According to the second sum and the target number, determine the At the sampling moment The traffic flow variability of a road.

[0009] Further, the second sum and the target number are determined. At the sampling moment The traffic flow change degree of a road includes the following specific steps: The normalized value of the product of the second sum and the target number is recorded as At the sampling moment The traffic flow variability of a road.

[0010] Furthermore, the specific steps of determining the traffic flow prediction weight of each road at each sampling moment include the following: According to At the sampling moment On the road The difference between the traffic flow at the monitoring point and the traffic flow at all other monitoring points, combined with the At the sampling moment The traffic flow change degree of the road is determined At the sampling moment The first road The second product of monitoring points; Preset second constant , in the From the sampling moment to the During the period of sampling time, obtain the The first road The normalized value of the sum of the second products of monitoring points is recorded as At the sampling moment The first road Traffic flow anomaly degree at each monitoring point; According to At the sampling moment The traffic flow abnormality of all monitoring points on the road, and At the sampling moment The traffic volume of the road is determined At the sampling moment Traffic flow prediction weight for each road.

[0011] Furthermore, the determination At the sampling moment The first road The second product of the monitoring points includes the following specific steps: In the At the sampling moment On the road, get The mean of the difference between the traffic flow at the first monitoring point and the traffic flow at all other monitoring points is recorded as the first mean. The inverse proportional normalized value of the first mean is compared with the first At the sampling moment The product of the traffic flow change degree of the road is recorded as At the sampling moment The first road The second product of the monitoring points.

[0012] Furthermore, according to At the sampling moment The traffic flow abnormality of all monitoring points on the road, and At the sampling moment The traffic volume of the road is determined At the sampling moment The traffic flow prediction weight of each road includes the following specific steps: Get the At the sampling moment The mean of the traffic flow anomaly of all monitoring points on the road is recorded as the third mean. At the sampling moment The normalized value of the product of the traffic volume of the road is recorded as At the sampling moment Traffic flow prediction right for a road.

[0013] The present invention also proposes a road traffic flow monitoring system using an AI platform, which adopts any one of the above-mentioned methods for monitoring road traffic flow using an AI platform. The system includes the following modules: The traffic flow data acquisition module is used to obtain the traffic flow of each road, the traffic flow at each monitoring point on each road, and the traffic flow from one road to another connected road at each sampling moment; each road connects a number of upper-level roads and lower-level roads; The contribution weight analysis module is used to determine the contribution weight of each subordinate road connected to each road at each sampling moment based on the difference in traffic volume changes of different roads at consecutive sampling moments and the traffic volume of each subordinate road connected to each road at each sampling moment; a traffic flow change analysis module for determining the traffic flow change degree of each road at each sampling moment based on the contribution weight, the traffic flow from each subordinate road connected to each road to each connected road at each sampling moment, and the number of all superior roads connected to each road; The traffic flow prediction module is used to determine the traffic flow prediction weight of each road at each sampling moment based on the traffic flow variability and the difference in traffic flow at different monitoring points on each road at each sampling moment; the traffic flow prediction weight and traffic flow of each road at all sampling moments are used as the input of the prediction model, and the predicted traffic flow at several future moments is output.

[0014] The present invention also proposes a road traffic flow monitoring device using an AI platform, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the aforementioned method for monitoring road traffic flow using an AI platform.

[0015] The beneficial effects of the technical solution of the present invention are: In an embodiment of the present invention, the contribution weight of each subordinate road connected to each road at each sampling moment is determined based on the difference in traffic flow changes between different roads at consecutive sampling moments, combined with the traffic flow of each subordinate road connected to each road at each sampling moment. The traffic flow variability of each road at each sampling moment is determined based on the traffic flow entering each connected road from each subordinate road at each sampling moment, as well as the number of all superior roads connected to each road. Based on the temporal changes in traffic flow on the road and the traffic flow inflow and outflow from other roads connected to the road, the traffic flow variability of the road is determined, which is used to obtain a traffic flow prediction weight, ensure the accuracy of traffic flow prediction, and thus ensure the timeliness of road traffic flow monitoring. Furthermore, the traffic flow prediction weight of each road at each sampling moment is determined based on the traffic flow differences at different monitoring points on each road at each sampling moment. The traffic flow prediction weight and traffic flow of each road at all sampling moments are used as inputs to a prediction model, which outputs predicted traffic flows at several future moments. Thus, the present invention plans and adjusts traffic routes in advance by accurately predicting traffic flow data, guiding vehicles to avoid congested sections during peak hours, and achieving a balanced distribution of traffic flow. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 This is a flowchart of the steps of a road traffic flow monitoring method using an AI platform according to the present invention; Figure 2 This is a structural block diagram of a road traffic flow monitoring system using an AI platform according to the present invention. DETAILED DESCRIPTION

[0018] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation methods, structures, features and effects of a road traffic flow monitoring method, device and system using an AI platform proposed by the present invention. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics of one or more embodiments may be combined in any suitable form.

[0019] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0020] The following describes in detail a method, device, and system for monitoring road traffic flow using an AI platform provided by the present invention with reference to the accompanying drawings.

[0021] See also Figure 1 , which shows a flowchart of a method for monitoring road traffic flow using an AI platform provided by one embodiment of the present invention, the method comprising the following steps: Step S001: At each sampling moment, obtain the traffic flow of each road, the traffic flow at each monitoring point of each road, and the traffic flow from one road to another connected road; each road connects several upper-level roads and lower-level roads.

[0022] It should be noted that when monitoring road traffic flow, it is necessary to collect multi-dimensional data to fully reflect the traffic operation status. In this embodiment, the main data categories and collection methods are as follows: (1) Fixed collection equipment (video detection system): using smart cameras with more than 2 million pixels, installed on pillars, with a height of 6 to 8 meters and a spacing of 300 to 500 meters, that is, each pillar is a monitoring point, and then multi-target tracking is achieved based on algorithms such as YOLOv5 / 8. This is a well-known technology and the specific method will not be introduced here. (2) Mobile collection means (floating vehicle data): car GPS trajectory (sampling frequency is once every 10 to 30 seconds) and mobile phone signaling data (needs to be desensitized).

[0023] Therefore, within any urban road monitoring area, the traffic flow of each road at each sampling moment, the traffic flow at each monitoring point of each road at each sampling moment, and the traffic flow from one road to another connected road at each sampling moment can be obtained; each road connects several upper-level roads and lower-level roads.

[0024] The sampling frequency is once per minute, i.e., the number of vehicles passing each road per minute, the number of vehicles passing each road per minute at each monitoring point, and the number of vehicles entering from one road to another connected road per minute at intersections are collected. In this embodiment, urban roads are monitored. Urban roads are generally divided into expressways, main roads, secondary roads, and branch roads based on their status and function in the urban transportation system. The upper-level roads of a branch road are expressways, main roads, and secondary roads, and the lower-level roads of a branch road are other branches connected to it. The upper-level roads of a secondary road are expressways and main roads, and the lower-level roads of a secondary road are branches. The upper-level road of a main road is an expressway, and the lower-level roads of a main road are secondary roads and branches. The upper-level road of an expressway is other expressways connected to it, and the lower-level roads of an expressway are main roads, secondary roads, and branches.

[0025] Step S002: Determine the contribution weight of each subordinate road connected to each road at each sampling moment based on the difference in traffic volume changes of different roads at consecutive sampling moments and the traffic volume of each subordinate road connected to each road at each sampling moment.

[0026] It should be noted that monitoring road traffic flow is a core requirement for modern traffic management and smart city development. Real-time traffic flow monitoring can accurately identify road congestion points, abnormal events (such as accidents or illegal parking), and traffic efficiency bottlenecks. It provides data support for dynamic signal timing, variable lane control, and emergency dispatch, thereby improving the overall traffic efficiency of the road network. Furthermore, by analyzing traffic flow, vehicle speed, and vehicle type distribution, dangerous driving behaviors such as speeding and driving against traffic can be promptly detected, preventing traffic accidents. Furthermore, during large-scale events or emergencies, real-time traffic flow data can assist in planning evacuation routes and ensure public safety. Long-term traffic flow data can reveal road usage patterns, providing a scientific basis for road network expansion, bus route optimization, and parking facility layout, thereby avoiding resource waste. Therefore, this embodiment achieves road traffic flow monitoring by sensing information about different areas and road conditions, analyzing the mutual impact of traffic flow between different areas, and predicting changes in traffic flow on roads at different times.

[0027] It is further important to note that urban roads are divided into different levels based on their traffic functions, service levels, and design speeds, including expressways, trunk roads, secondary roads, and branch roads. These roads have varying traffic volumes and service attributes. For example, trunk roads are designed to carry the majority of a city's traffic, connecting various functional zones (such as commercial and residential areas), serving short- and medium-distance travel, and feature numerous lanes and high traffic volumes. Trunk roads, secondary roads, and branch roads interact with each other at different times of day. For example, during rush hour, traffic on branch roads tends to converge on trunk roads, with a greater concentration in central business districts and office areas. In residential areas, traffic disperses from trunk roads to secondary roads and branch roads, resulting in varying traffic volumes in different areas.

[0028] Preferably, in one embodiment of the present invention, the method for obtaining the contribution weight of each subordinate road connected to each road at each sampling moment includes: Preset first constant The value is 60, and this is used as an example for description.

[0029] Get from From the sampling moment to the All sampling moments within the period of the sampling moment The traffic flow sequence composed of the traffic flow of the roads.

[0030] In the traffic flow sequence, calculate the Traffic volume minus the The difference between the traffic flows of the adjacent vehicles and the sum of the differences between the traffic flows of the adjacent vehicles The normalized value of At the sampling moment The traffic volume variation coefficient of the road.

[0031] It should be noted that: according to the above method, the traffic flow change coefficient of each road under each sampling can be obtained. Since road traffic flow monitoring is a long-term monitoring, there will be several days or even months of historical monitoring data when the current analysis is performed, which ensures that the historical data can be supported when the data analysis is performed. In this embodiment, in order to describe the functional changes of different roads, by obtaining the changes in traffic flow on different roads, if the traffic flow change coefficient is a positive number, it means that in the first The traffic volume on this road section increased within one hour before the sampling time. On the contrary, if the traffic volume change coefficient is negative, it means that the traffic volume on the road section increased within one hour before the sampling time. The traffic volume on this road section decreased within an hour before the sampling time. As The normalized value of It is a linear normalization function used to normalize data values ​​to between 0 and 1.

[0032] Get the The average of the traffic flow variation coefficients of all roads at a sampling time is recorded as the overall traffic flow variation coefficient.

[0033] In from From the sampling moment to the During the sampling time period, obtain the absolute value of the difference between the traffic flow change coefficients of any two roads at each sampling time. The inverse proportional normalized value is recorded as the similarity value at each sampling moment, and the sum of the similarity values ​​at all sampling moments is recorded as the comprehensive similarity value. The product of the comprehensive similarity value and the overall traffic flow change coefficient is The normalized value of The traffic flow influencing factor of any two roads at a sampling time.

[0034] It should be noted that: in this embodiment, As The inverse normalized value of As Normalized value. Because the position of the road is fixed, when the adjacent roads have an impact, the traffic flow will change at the same time. When the overall traffic flow change coefficient is larger, it means that the traffic flow in the urban road monitoring area increases and the traffic flow image between roads is larger. Within one hour before the sampling time, the smaller the difference in the traffic flow variation coefficients of any two roads is, the more it means that the traffic flows of the two roads affect each other.

[0035] It should be further explained that roads are distributed hierarchically, different road nodes are distributed in a grid pattern, the traffic volume on roads at different levels is different, and for a main road, different branches have different degrees of influence on it. Therefore, it is necessary to obtain the contribution weights of different roads based on the traffic volume impact factor of the road.

[0036] Get from From the sampling moment to the All sampling moments within the period of the sampling moment The road connecting The sum of the traffic volumes of the subordinate roads is recorded as the first sum.

[0037] The first At the sampling moment The road connecting The lower level road and the The product of the traffic flow impact factor of the road and the first sum value The normalized value of At the sampling moment The road connecting The contribution weight of the subordinate roads.

[0038] What needs to be explained is: As Normalized value of . During peak hours, the traffic volume of branch roads is large, which may also cause traffic congestion on the main road. Therefore, when predicting traffic volume, the greater the contribution weight is, the greater the first sum value is. The first sampling time within an hour before the The road connecting The greater the traffic volume of the lower-level road, and at this time The road connecting The lower level road and the The greater the traffic flow impact factor of a road, the greater its contribution weight.

[0039] Step S003: Determine the traffic flow variation of each road at each sampling moment based on the contribution weight, the traffic flow from each lower-level road connected to each road at each sampling moment, and the number of all upper-level roads connected to each road.

[0040] It should be noted that because urban roads are complex networks with numerous intersecting trunk and branch roads, and different urban areas serve different functions, road traffic volume can vary. For example, if a certain area is crowded with businesses, traffic volume can be high during weekdays, with particularly high volumes during morning and evening rush hours. Large shopping malls and tourist attractions, on the other hand, attract large crowds of visitors on weekends and holidays, resulting in high traffic volumes on surrounding roads. Furthermore, during rush hour, traffic flows from other areas into areas with a high concentration of businesses, resulting in short-term high traffic volumes. Since most vehicles entering these areas travel via the city's main roads and expressways, it is necessary to determine the traffic flow variability of the current road based on changes in traffic flow over different time periods.

[0041] Preferably, in one embodiment of the present invention, the method for obtaining the traffic flow variation degree of each road at each sampling moment includes: Get the The number of all superior roads connected by the road The normalized value of is recorded as the target number.

[0042] Get the At the sampling moment The road connecting The contribution weight of the lower-level road is the same as that of the At the sampling moment, The road connecting The lower level road enters the connected The product of the traffic volume of the roads is recorded as At the sampling moment The road connecting The first product of the lower-level roads, the At the sampling moment The sum of the first products of all the subordinate roads connected by the road is recorded as the second sum, and the product of the second sum and the target number is The normalized value of At the sampling moment The traffic flow variability of a road.

[0043] What needs to be explained is: As The normalized value of As The more the number of upper-level roads is, the higher the The greater the change in vehicle outflow on the road, the greater the change in vehicle outflow on the road. At the sampling moment, The contribution weight of the lower-level road connected by the first road is greater, and the contribution weight of the lower-level road connected by the first road is greater. The greater the traffic volume on a road, the The greater the change in the number of vehicles merging on the road, the product of the second sum and the target number is used as the first At the sampling moment The traffic flow variability of a road.

[0044] According to the above method, the traffic flow variation degree of each road at each sampling moment can be obtained.

[0045] Step S004: Determine the traffic flow prediction weight of each road at each sampling moment based on the traffic flow variability and the traffic flow difference at different monitoring points on each road at each sampling moment; use the traffic flow prediction weight and traffic flow of each road at all sampling moments as the input of the prediction model, and output the predicted traffic flow at several future moments.

[0046] It should be noted that changes in road traffic volume are also affected by various factors, such as road construction. When a main road is under construction, it affects traffic efficiency, and vehicles may choose to detour to reduce congestion, which will affect the traffic volume on roads around the road. Road construction generally lasts for a long time, which has a greater impact on traffic flow. Therefore, we collect traffic flow data over a period of time, analyze the changes in current road traffic flow, and then predict road traffic flow.

[0047] Preferably, in one embodiment of the present invention, the method for obtaining the predicted traffic flow at a future moment includes: In the At the sampling moment On the road, get The mean of the difference between the traffic flow at the monitoring point and the traffic flow at all other monitoring points is recorded as the first mean. The inverse normalized value of At the sampling moment The product of the traffic flow change degree of the road is recorded as At the sampling moment The first road The second product of the monitoring points.

[0048] Among them, As The inversely proportional normalized value of .

[0049] Preset second constant The value is 1440, and this is used as an example for description.

[0050] In from From the sampling moment to the During the period of sampling time, obtain the The first road The sum of the second products of the monitoring points The normalized value of At the sampling moment The first road The traffic flow abnormality of each monitoring point.

[0051] What needs to be explained is: As The normalized value of . If the traffic volume at a certain monitoring point on the same road is smaller than that at other monitoring points (the first mean is smaller) within the day before a sampling moment, it means that traffic at that monitoring point is affected by construction or other factors. The greater the variability of the traffic flow at this time, that is, due to construction detours, vehicles are constantly entering and exiting on both sides of the construction road section, and the more abnormal the traffic flow on the road is.

[0052] Get the At the sampling moment The mean of the traffic flow anomaly of all monitoring points on the road is recorded as the third mean. At the sampling moment The product of the traffic volume of the roads The normalized value of At the sampling moment Traffic flow prediction weight for each road.

[0053] It should be noted that the greater the traffic abnormality and the larger the traffic volume, the greater the impact on road traffic, and the more attention needs to be paid during prediction to ensure the accuracy of the prediction.

[0054] For any road, the ARIMA prediction model is used to obtain the predicted traffic flow at several future moments based on the traffic flow prediction weights and traffic flow at all sampling moments.

[0055] It should be noted that the ARIMA forecasting model, which stands for Autoregressive Integrated Moving Average Model (ARIMA), is a well-known technique, and its specific methods are not described here. The model's input data consists of the predicted traffic weights and volume at all sampling times within the previous day, and its output is the predicted traffic volume for the next 30 minutes. This forecasted volume allows for pre-planning and adjustment of traffic routes, guiding vehicles to avoid congested areas during peak hours and achieving a balanced distribution of traffic flow.

[0056] Second, see Figure 2 , which shows a road traffic flow monitoring system using an AI platform provided by one embodiment of the present invention, the system includes the following modules: The traffic flow data acquisition module is used to obtain the traffic flow of each road, the traffic flow at each monitoring point on each road, and the traffic flow from one road to another connected road at each sampling moment; each road connects a number of upper-level roads and lower-level roads; The contribution weight analysis module is used to determine the contribution weight of each subordinate road connected to each road at each sampling moment based on the difference in traffic volume changes of different roads at consecutive sampling moments and the traffic volume of each subordinate road connected to each road at each sampling moment; a traffic flow change analysis module for determining the traffic flow change degree of each road at each sampling moment based on the contribution weight, the traffic flow from each subordinate road connected to each road to each connected road at each sampling moment, and the number of all superior roads connected to each road; The traffic flow prediction module is used to determine the traffic flow prediction weight of each road at each sampling moment based on the traffic flow variability and the difference in traffic flow at different monitoring points on each road at each sampling moment; the traffic flow prediction weight and traffic flow of each road at all sampling moments are used as the input of the prediction model, and the predicted traffic flow at several future moments is output.

[0057] In a third aspect, the present invention also provides a road traffic flow monitoring device using an AI platform, comprising a memory, a processor, and a computer program stored in the memory and runnable on the processor, wherein the processor executes the computer program stored in the memory to implement the steps of the aforementioned method for road traffic flow monitoring using an AI platform.

[0058] So far, the present invention is completed.

[0059] In summary, in an embodiment of the present invention, based on the difference in traffic volume changes between different roads at consecutive sampling moments, combined with the traffic volume of each subordinate road connected to each road at each sampling moment, the contribution weight of each subordinate road connected to each road at each sampling moment is determined. Combined with the traffic volume entering each road connected from each subordinate road at each sampling moment, and the number of all superior roads connected to each road, the traffic volume variation of each road at each sampling moment is determined. Combined with the traffic volume difference at different monitoring points on each road at each sampling moment, the traffic volume prediction weight of each road at each sampling moment is determined. The traffic volume prediction weight and traffic volume of each road at all sampling moments are used as inputs to the prediction model, and the predicted traffic volume at several future moments is output. By accurately predicting traffic volume data, the present invention can plan and adjust traffic routes in advance, guide vehicles to avoid congested sections during peak hours, and achieve a balanced distribution of traffic flow.

[0060] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A road traffic flow monitoring method using an AI platform, characterized in that: The method comprises the following steps: At each sampling moment, the traffic volume of each road, the traffic volume at each monitoring point on each road, and the traffic volume entering from one road to another connected road are obtained; each road connects several upper-level roads and lower-level roads; Based on the difference in traffic volume changes of different roads at consecutive sampling moments, combined with the traffic volume of each subordinate road connected to each road at each sampling moment, the contribution weight of each subordinate road connected to each road at each sampling moment is determined; Determine the traffic flow variation of each road at each sampling moment based on the contribution weight, the traffic flow from each lower-level road connected to each road at each sampling moment, and the number of all upper-level roads connected to each road; Based on the traffic flow variability and the difference in traffic flow at different monitoring points on each road at each sampling moment, the traffic flow prediction weight of each road at each sampling moment is determined; the traffic flow prediction weight and traffic flow of each road at all sampling moments are used as the input of the prediction model, and the predicted traffic flow at several future moments is output.

2. The method for monitoring road traffic flow using an AI platform according to claim 1, characterized in that: The specific steps of determining the contribution weight of each subordinate road connected to each road at each sampling moment are as follows: Preset first constant , get from From the sampling moment to the All sampling moments within the period of the sampling moment The traffic flow sequence composed of the traffic flow of the roads; In the traffic flow sequence, calculate the Traffic volume minus the The difference between the traffic flows of each vehicle is used to obtain the normalized value of the sum of the differences of all adjacent traffic flows, which is recorded as At the sampling moment The traffic volume variation coefficient of the road; According to The difference in traffic flow variation coefficients of different roads at the sampling time is used to determine the Traffic flow influencing factors of any two roads at a sampling time; Get from From the sampling moment to the All sampling moments within the period of the sampling moment The road connecting The sum of the traffic volumes of the subordinate roads is recorded as the first sum; The first At the sampling moment The road connecting The lower level road and the The normalized value of the product of the traffic flow impact factor of the road and the first sum value is recorded as At the sampling moment The road connecting The contribution weight of the subordinate roads.

3. The method for monitoring road traffic flow using an AI platform according to claim 2, characterized in that: The determination of The traffic flow influencing factors of any two roads at a sampling time include the following specific steps: Get the The average of the traffic flow variation coefficients of all roads at the sampling time is recorded as the overall traffic flow variation coefficient; In the From the sampling moment to the During the period of each sampling moment, the inversely proportional normalized value of the absolute value of the difference between the traffic flow variation coefficients of any two roads at each sampling moment is obtained, which is recorded as the similarity value at each sampling moment. The sum of the similarity values ​​at all sampling moments is recorded as the comprehensive similarity value. The normalized value of the product of the comprehensive similarity value and the overall traffic flow variation coefficient is recorded as the first The traffic flow influencing factor of any two roads at a sampling time.

4. The method for monitoring road traffic flow using an AI platform according to claim 1, characterized in that: The specific steps of determining the traffic flow variation degree of each road at each sampling moment are as follows: Get the The normalized value of the number of all superior roads connected by a road is recorded as the target number; Get the At the sampling moment The road connecting The contribution weight of the subordinate roads is the same as that of the At the sampling moment, The road connecting The lower level road enters the connected The product of the traffic volume of the roads is recorded as At the sampling moment The road connecting The first product of the lower-level roads, the At the sampling moment The sum of the first products of all subordinate roads connected by the road is recorded as the second sum; According to the second sum and the target number, determine the At the sampling moment The traffic flow variability of a road.

5. The method for monitoring road traffic flow using an AI platform according to claim 4, characterized in that: The second sum and the target number are determined. At the sampling moment The traffic flow change degree of a road includes the following specific steps: The normalized value of the product of the second sum and the target number is recorded as At the sampling moment The traffic flow variability of a road.

6. The method for monitoring road traffic flow using an AI platform according to claim 1, characterized in that: The specific steps of determining the traffic flow prediction weight of each road at each sampling moment are as follows: According to At the sampling moment On the road The difference between the traffic flow at the monitoring point and the traffic flow at all other monitoring points, combined with the At the sampling moment The traffic flow change degree of the road is determined At the sampling moment The first road The second product of monitoring points; Preset second constant , in the From the sampling moment to the During the period of sampling time, obtain the The first road The normalized value of the sum of the second products of monitoring points is recorded as At the sampling moment The first road Traffic flow anomaly degree at each monitoring point; According to At the sampling moment The traffic flow abnormality of all monitoring points on the road, and At the sampling moment The traffic volume of the road is determined At the sampling moment Traffic flow prediction weight for each road.

7. The method for monitoring road traffic flow using an AI platform according to claim 6, characterized in that: The determination of At the sampling moment The first road The second product of the monitoring points includes the following specific steps: In the At the sampling moment On the road, get The mean of the difference between the traffic flow at the first monitoring point and the traffic flow at all other monitoring points is recorded as the first mean. The inverse proportional normalized value of the first mean is compared with the first At the sampling moment The product of the traffic flow change degree of the road is recorded as At the sampling moment The first road The second product of the monitoring points.

8. The method for monitoring road traffic flow using an AI platform according to claim 6, characterized in that: According to the At the sampling moment The traffic flow abnormality of all monitoring points on the road, and At the sampling moment The traffic volume of the road is determined At the sampling moment The traffic flow prediction weight of each road includes the following specific steps: Get the At the sampling moment The mean of the traffic flow anomaly of all monitoring points on the road is recorded as the third mean. At the sampling moment The normalized value of the product of the traffic volume of the road is recorded as At the sampling moment Traffic flow prediction right for a road.

9. A road traffic flow monitoring system using an AI platform, using a road traffic flow monitoring method using an AI platform as claimed in any one of claims 1 to 8, characterized in that: The system includes the following modules: The traffic flow data acquisition module is used to obtain the traffic flow of each road, the traffic flow at each monitoring point on each road, and the traffic flow from one road to another connected road at each sampling moment; each road connects a number of upper-level roads and lower-level roads; The contribution weight analysis module is used to determine the contribution weight of each subordinate road connected to each road at each sampling moment based on the difference in traffic volume changes of different roads at consecutive sampling moments and the traffic volume of each subordinate road connected to each road at each sampling moment; a traffic flow change analysis module for determining the traffic flow change degree of each road at each sampling moment based on the contribution weight, the traffic flow from each subordinate road connected to each road to each connected road at each sampling moment, and the number of all superior roads connected to each road; The traffic flow prediction module is used to determine the traffic flow prediction weight of each road at each sampling moment based on the traffic flow variability and the difference in traffic flow at different monitoring points on each road at each sampling moment; the traffic flow prediction weight and traffic flow of each road at all sampling moments are used as the input of the prediction model, and the predicted traffic flow at several future moments is output.

10. A road traffic flow monitoring device using an AI platform, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is executed by a processor, the steps of a road traffic flow monitoring method using an AI platform as described in any one of claims 1 to 8 are implemented.

Images