Monitoring and adjustment method of intelligent transportation system based on Internet of Things

By adaptively adjusting the ARIMA model parameters, combined with the changing trends and periodic characteristics of vehicle flow data, the problem of low accuracy of vehicle flow data processing in the existing technology is solved, more accurate traffic light adjustment is achieved, and road traffic is improved efficiency.

CN118762524BActive Publication Date: 2024-11-19洛阳墨潇网络科技有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411238520.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-05
Publication Date
2024-11-19
Estimated Expiration
2044-09-05

AI Technical Summary

Technical Problem

The ARIMA model used in existing smart transportation systems lacks noise resistance when processing traffic data, making it difficult to capture the complex relationship between the data, resulting in low accuracy of monitoring and regulation.

Method used

By obtaining traffic data for multiple days in real time, analyzing the changing trends and periodic characteristics, adaptively adjusting the ARIMA model parameters, combining the degree of abnormality and the correlation between adjacent intersections, a more robust traffic prediction model is built for adjustment Traffic light duration.

Benefits of technology

It improves the accuracy of traffic flow prediction and the noise resistance of the model, enhances the accuracy of traffic light adjustment, and improves road traffic efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118762524B_ABST
    Figure CN118762524B_ABST
Patent Text Reader

Abstract

The present application relates to the field of traffic light control, and specifically to a monitoring and adjustment method for a smart transportation system based on the Internet of Things, including obtaining a first traffic flow data sequence in different directions at each intersection in real time, analyzing the changing trend and periodic characteristics of the traffic flow, and determining the abnormal degree of each traffic flow; according to the abnormal degree of each traffic flow and the correlation between the traffic flow at adjacent intersections, each parameter in the adaptive ARIMA model is used to obtain a traffic flow prediction model for each intersection in each direction; the second traffic flow data sequence is input into the corresponding traffic flow prediction model to obtain the prediction results for each intersection in different directions, and then the duration of the traffic lights at each intersection is adjusted. The present application obtains more accurate traffic flow prediction results through the various parameters in the adaptive ARIMA model, and improves the accuracy of the traffic light duration adjustment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of traffic signal light control, and specifically to a monitoring and control method for an intelligent transportation system based on the Internet of Things. Background Art

[0002] The Internet of Things is a technology that connects any object to the network through information sensing devices. Objects exchange and communicate information through information dissemination media to achieve intelligent identification, positioning, tracking, supervision and other functions. With the rapid increase in the number of cars, more and more road traffic problems frequently occur. The Internet of Things technology has also been applied to traffic planning and management. For example, the duration of traffic lights is controlled according to real-time traffic information such as traffic flow to improve the efficiency of intersections and avoid wasting road traffic capacity. Road traffic conditions are constantly changing, and various special situations may occur at any time. Therefore, adaptive control of traffic lights under complex traffic backgrounds is an urgent problem to be solved.

[0003] In the existing technology, the smart transportation system collects road traffic data in real time through IoT devices, uses the ARIMA (Autoregressive Integrated Moving Average Model) model to analyze the traffic data patterns, and automatically adjusts the duration of traffic lights based on the analysis results to alleviate traffic congestion. However, the parameters in the ARIMA model in the existing methods are usually selected within a range of values, lacking noise resistance, and the pre-set parameters make it difficult to extract effective information from various dynamically changing data. Especially for traffic data, traffic data often presents complex nonlinear relationships and strong periodicity, and is easily affected by emergencies. The ARIMA model cannot capture the relationship between data when processing data, resulting in low accuracy in system monitoring and adjustment. Summary of the invention

[0004] In order to solve the technical problem that the preset parameter values ​​in the ARIMA model lead to low accuracy of system monitoring and adjustment, the purpose of this application is to provide a monitoring and adjustment method for a smart transportation system based on the Internet of Things. The technical solutions adopted are as follows:

[0005] An embodiment of the present application provides a method for monitoring and adjusting a smart transportation system based on the Internet of Things, the method comprising the following steps:

[0006] Acquire in real time a first traffic flow data sequence in different directions at all intersections in the monitored urban area for several days, wherein the first traffic flow data sequence includes the traffic flow in each preset time period;

[0007] Analyze the change trend and periodicity characteristics of the traffic flow according to each traffic flow in the first traffic flow data sequence to which each traffic flow belongs, and determine the abnormality degree of each traffic flow;

[0008] According to the abnormal degree of each traffic flow and the correlation between the traffic flows of adjacent intersections, each parameter in the adaptive ARIMA model is used to obtain a traffic flow prediction model for each intersection in each direction.

[0009] Obtain a second traffic flow data sequence corresponding to each intersection in different directions with a currently set duration, input the second traffic flow data sequence into a corresponding traffic flow prediction model, and obtain prediction results for each intersection in different directions; and adjust the duration of the traffic lights at each intersection according to the prediction results.

[0010] Furthermore, the step of analyzing the change trend and periodicity characteristics of the traffic flow according to each traffic flow in the first traffic flow data sequence to which each traffic flow belongs, and determining the abnormality degree of each traffic flow includes:

[0011] Taking any vehicle flow rate as a target vehicle flow rate, and determining an abnormal factor of a change trend of the target vehicle flow rate according to the target vehicle flow rate and a first preset number of vehicle flows closest to the target vehicle flow rate;

[0012] Divide the first vehicle flow data sequence into vehicle flow data subsequences corresponding to each day, determine the vehicle flow in the vehicle flow data subsequence that is at the same position as the target vehicle flow, record it as the periodic vehicle flow, and then determine the abnormal factor of the change trend of the periodic vehicle flow;

[0013] Determine the periodic abnormal factor of the target traffic flow according to the target traffic flow, the traffic flow of each period and the abnormal factor of the change trend of the traffic flow of each period;

[0014] The abnormal degree of the target vehicle flow is determined by combining the abnormal trend factor and the period abnormal factor of the target vehicle flow, and both the abnormal trend factor and the period abnormal factor are positively correlated with the abnormal degree.

[0015] Furthermore, the determining of the abnormal factor of the change trend of the target vehicle flow according to the target vehicle flow and the first preset number of vehicle flows closest to the target vehicle flow includes:

[0016] Taking a first preset number of vehicle flows closest to the target vehicle flow as adjacent vehicle flows, and obtaining each adjacent vehicle flow of the target vehicle flow;

[0017] The calculation formula of the abnormal factor of the change trend of target traffic flow is:

[0018] ; In the formula, is the abnormal factor of the change trend of the target traffic flow, () is a linear normalization function, v is the vth neighboring vehicle flow of the target vehicle flow, V is the number of neighboring vehicle flows, is the target traffic volume, is the vth neighboring traffic flow of the target traffic flow, To find the absolute value function, is the ceiling function.

[0019] Furthermore, the determining of the periodic abnormality factor of the target vehicle flow according to the target vehicle flow, each periodic vehicle flow and the abnormality factor of the change trend of each periodic vehicle flow includes:

[0020] Perform negative correlation processing on the abnormal factors of the change trend of the traffic flow in each period, obtain the preliminary weight of the traffic flow in each period, and take the ratio of the preliminary weight to the cumulative sum of all preliminary weights as the final weight;

[0021] The product of each periodic traffic flow and its corresponding final weight is accumulated over all periodic traffic flows as the weighted value of all periodic traffic flows, and the difference between the target traffic flow and the weighted value of all periodic traffic flows is calculated as the initial periodic abnormality factor;

[0022] The initial cycle anomaly factor is normalized and used as the cycle anomaly factor of the target vehicle flow.

[0023] Furthermore, the parameters in the adaptive ARIMA model according to the abnormality of each traffic flow and the correlation between the traffic flows at adjacent intersections include:

[0024] Take any intersection as the target intersection, take any direction of the target intersection as the target direction, select the previous intersection and the next intersection of the target intersection in the target direction, call the direction of vehicles moving toward the target direction the convergence direction, and call the direction of vehicles moving away from the target direction the divergence direction;

[0025] According to each vehicle flow rate and its abnormality degree in each converging direction of the previous intersection, each vehicle flow rate and its abnormality degree in the target direction and diverging direction of the target intersection, and each vehicle flow rate and its abnormality degree in the converging direction and diverging direction of the next intersection, the parameters of each AR part and each MA part of the vehicle flow prediction model in the target direction of the target intersection are determined;

[0026] Selecting a second preset number of intersections that are located before the target intersection and closest to the target intersection in the target direction as related intersections of the target intersection, and obtaining actual distances between each related intersection of the target intersection and the target intersection;

[0027] According to each traffic flow and its abnormality at each related intersection of the target intersection in each converging direction, and the actual distance between each related intersection of the target intersection and the target intersection, the constant term of the traffic flow prediction model of the target intersection in the target direction is determined.

[0028] Furthermore, the parameters of each AR part and each MA part of the vehicle flow prediction model at the target intersection in the target direction are determined, including:

[0029] The collection time of any vehicle flow in the target direction at the target intersection is recorded as t; according to the vehicle flow in the target direction and the diverging direction of the target intersection at the collection time t, the vehicle flow in each converging direction of the previous intersection at the collection time tb, and the vehicle flow in the converging direction and the diverging direction of the next intersection at the collection time t+c, the first noise factor of the vehicle flow in the target direction of the target intersection at the collection time t is determined; wherein b is the average driving time from the previous intersection to the target intersection, and c is the average driving time from the target intersection to the next intersection;

[0030] Determine the second noise factor of the traffic flow in the target direction at the target intersection at the collection time t according to the abnormal degree of the traffic flow in the target direction at the collection time t, the abnormal degree of the traffic flow in each converging direction at the previous intersection at the collection time tb, and the abnormal degree of the traffic flow in the converging direction and the diverging direction at the next intersection at the collection time t+c;

[0031] Calculate the product of the abnormal degree of the vehicle flow in the target direction at the target intersection at the collection time t and the first noise factor and the second noise factor, normalize the product, and use the normalized product as the noise degree, thereby obtaining the noise degree of each vehicle flow in the target direction at the target intersection;

[0032] The noise levels of each vehicle flow at the target intersection in the target direction are negatively correlated to obtain initial parameters; the ratio of the initial parameters to the cumulative sum of all initial parameters is used as the parameters of the AR part and the MA part of the corresponding vehicle flow, and then the parameters of each AR part and the parameters of each MA part of the vehicle flow prediction model at the target intersection in the target direction are obtained; wherein, the parameters of the AR part and the parameters of the MA part of the same vehicle flow are equal.

[0033] Furthermore, the first noise factor of the vehicle flow at the target intersection in the target direction at the acquisition time t is determined, including:

[0034] ; In the formula, is the first noise factor of the traffic flow in the target direction at the target intersection at the acquisition time t, is the traffic flow in the target direction at the target intersection at the collection time t, , and are the target direction, the first diversion direction, and the second diversion direction of the target intersection respectively. is the cumulative sum of the traffic flow in the converging direction and diverging direction at the next intersection after the collection time t+c, C is the next intersection after the target intersection, c is the average driving time from the target intersection to the next intersection, is the cumulative sum of the traffic volume in each converging direction of the previous intersection at the collection time tb, b is the average driving time from the previous intersection to the target intersection, B is the previous intersection of the target intersection, is the traffic flow in the first diverging direction of the target intersection at the collection time t, is the traffic flow in the second diverging direction of the target intersection at the collection time t, To find the absolute value function, is a hyperparameter.

[0035] Furthermore, the second noise factor of the vehicle flow at the target intersection in the target direction at the acquisition time t is determined, including:

[0036] ; In the formula, is the second noise factor of the traffic flow in the target direction at the target intersection at the acquisition time t, is the abnormal degree of traffic flow in the target direction at the target intersection at the acquisition time t, is the average value of the abnormal degree of traffic flow in the converging and diverging directions at the next intersection after the collection time t+c, is the average value of the abnormal degree of traffic flow in each converging direction at the previous intersection at the collection time tb, To find the absolute value function.

[0037] Further, the constant term of the vehicle flow prediction model of the target intersection in the target direction is determined according to each vehicle flow and its abnormality at each related intersection of the target intersection in each converging direction and the actual distance between each related intersection of the target intersection and the target intersection, including:

[0038] ; In the formula, is the constant term of the traffic flow prediction model of the target intersection in the target direction, N is the number of related intersections corresponding to the target intersection, X is the sequence number of the related intersections corresponding to the target intersection, max[] is the maximum value function, T is the next moment after the current moment, is the accumulated value of the abnormal degree of traffic flow in each converging direction at the Xth related intersection at the collection time Tx, is the accumulated value of the noise level of the traffic flow in each converging direction at the Xth related intersection at the collection time Tx, x is the average driving time from the Xth target intersection to the target intersection, is the accumulated value of the traffic flow in each converging direction at the Xth related intersection at the collection time Tx, is the average value of the accumulated traffic flow in each converging direction at the Xth related intersection at all collection times, is the actual distance between the Xth related intersection and the target intersection, and exp() is an exponential function with a natural constant as the base.

[0039] Furthermore, the step of acquiring each vehicle flow in the first vehicle flow data sequence includes:

[0040] For any direction of any intersection, first determine the ratio of the green light duration at the intersection to the preset time period, then determine half of the original traffic flow data of the intersection in that direction, and take the ratio of half of the original traffic flow data to the ratio as the traffic flow of the intersection in that direction.

[0041] This application has the following beneficial effects:

[0042] The monitoring and adjustment method of the intelligent transportation system based on the Internet of Things provided in this application involves the field of traffic light control, and is mainly used for the duration adjustment of traffic lights. First, the abnormal degree of each traffic flow is determined by using the changing trend and periodic characteristics of the traffic flow data, which is helpful to analyze the abnormal traffic conditions that may occur based on the data characteristics of the traffic flow. The numerical accuracy and reliability of the abnormal degree determined based on the data characteristics of the two angles are higher; in addition, the abnormal degree provides data support for the subsequent adaptive prediction model parameters, while helping to enhance the ability to analyze traffic flow data in different situations; according to the abnormal degree of traffic flow and the correlation between the traffic flow at adjacent intersections, the parameters in the adaptive ARIMA model can better capture the relationship between the traffic flow data at each intersection, enhance the robustness of the model, and at the same time improve the anti-noise ability of the model, which is conducive to obtaining more accurate prediction results; based on the prediction results of more accurate traffic flow, the duration of the traffic light is adjusted, which helps to enhance the accuracy of system monitoring and adjustment, and further improve the traffic efficiency of the road. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 This is a flow chart of a method for monitoring and adjusting a smart transportation system based on the Internet of Things according to an embodiment of the present application;

[0045] Figure 2 A flow chart of determining the abnormality degree of target vehicle flow in an embodiment of the present application;

[0046] Figure 3 Schematic diagram of the distribution of target intersections in the embodiment of the present application. DETAILED DESCRIPTION

[0047] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the technical solutions proposed by the present application are described in detail below in conjunction with the accompanying drawings and preferred embodiments. 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 in one or more embodiments may be combined in any suitable form.

[0048] 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 application belongs.

[0049] The application scenarios targeted by the embodiments of this application are:

[0050] When monitoring the traffic flow on roads and adjusting the duration of traffic lights, the existing prediction model is used to analyze the data patterns and automatically adjust the duration of traffic lights according to the results to alleviate traffic congestion. However, the existing prediction model setting parameters leads to poor generalization ability for scenes, lack of noise resistance, and difficulty in extracting effective information from various dynamically changing data. In other words, the existing prediction model has low robustness, resulting in low accuracy in monitoring and adjusting traffic.

[0051] In order to improve the accuracy of monitoring and regulation, a monitoring and regulation method of a smart transportation system based on the Internet of Things is provided, such as Figure 1 As shown, the following steps are included:

[0052] S1, real-time acquisition of the first traffic flow data sequence in different directions at all intersections in the monitored urban area for several days.

[0053] In this embodiment, the road conditions are monitored in real time by cameras, sensors, radar equipment or Internet of Things devices such as GPS navigation arranged at all intersections in the monitored urban area. After obtaining the road conditions, the pending traffic flow is determined by combining the traffic flow data of all intersections in different directions; then, the traffic flow data is transmitted to the data processing system to facilitate the subsequent adjustment of the duration of traffic lights based on the data processing results.

[0054] When collecting pending traffic flow, the pending traffic flow in different directions at different intersections can be collected at the same time. The intersection can be a crossroads, and the direction can refer to the direction of vehicle travel; one pending traffic flow corresponds to one second, and the cumulative value of the pending traffic flow per second within the preset time period is used as the initial traffic flow; one preset time period corresponds to one initial traffic flow, and the preset time period can be set to 3 minutes, and the initial traffic flow for 7 days is collected. Among them, the preset time period needs to be greater than the cycle length of all traffic lights. The preset time period and the number of days for data collection can be set by the implementer according to the specific actual situation, and no specific restrictions are made here.

[0055] It should be noted that the ratio of traffic light duration in each preset time period is different. In order to avoid the loss of authenticity of traffic flow data due to uneven traffic light duration, which in turn affects the results of the prediction model, the initial traffic flow needs to be uniformly quantified according to the ratio of traffic light duration.

[0056] Ideally, the ratio of traffic light duration to green light duration can be 1:1. Therefore, the initial traffic flow obtained when the green light duration accounts for half of the preset period is ideal data. The specific implementation steps for obtaining the traffic flow at the intersection in different directions may include:

[0057] For any direction of any intersection, first determine the ratio of the green light duration at the intersection to the preset time period, then determine half of the original traffic flow data of the intersection in that direction, and take the ratio of half of the original traffic flow data to the ratio as the traffic flow of the intersection in that direction.

[0058] As an example, the calculation formula for obtaining the traffic flow at the intersection in this direction can be:

[0059] ; In the formula, F is the traffic flow of the intersection in this direction, f is the original traffic flow data of the intersection in this direction, is half of the original traffic flow data of the intersection in that direction, and a is the ratio of the green light duration of the intersection to the duration of the preset period.

[0060] After unifying all the original traffic flow data, all the traffic flows are obtained, and the traffic flows in the same direction at the same intersection within 7 days are sequenced to obtain the first traffic flow data sequence. Each direction at each intersection has its corresponding first traffic flow data sequence, and the first traffic flow data sequence includes the traffic flow in each preset time period.

[0061] So far, this embodiment has obtained real-time traffic information at different intersections in different directions.

[0062] S2, analyzing the change trend and periodicity characteristics of the traffic flow according to each traffic flow in the first traffic flow data sequence to which each traffic flow belongs, and determining the abnormality degree of each traffic flow.

[0063] In a real traffic environment, the traffic conditions on the road are highly random, and affected by factors such as insufficient accuracy of data acquisition equipment, monitoring environment, and unexpected traffic accidents, there are inevitably abnormal data with low authenticity in the first traffic flow data sequence. When constructing a traffic flow prediction model based on abnormal data, it is easy to reduce the prediction accuracy of the model, which is not conducive to the adjustment and control of the traffic light duration. In order to avoid the low accuracy of the prediction results to a certain extent, it is necessary to quantify the abnormal conditions of each traffic flow in order to provide data support for the subsequent determination of model parameters.

[0064] In this embodiment, the calculation process of the abnormality degree of all vehicle flows is consistent. In order to reduce unnecessary description, any vehicle flow is taken as an example to determine the abnormality degree, and any vehicle flow is taken as the target vehicle flow to determine the abnormality degree of the target vehicle flow, such as Figure 2 As shown, the specific implementation steps may include:

[0065] The first step is to determine the abnormal factor of the change trend of the target vehicle flow according to the target vehicle flow and a first preset number of vehicle flows closest to the target vehicle flow.

[0066] First of all, it should be noted that traffic flow is a type of data with strong correlation in time series. The change of traffic flow often follows certain change rules. If the traffic flow suddenly shows a large fluctuation in a short period of time, it means that the traffic flow is more likely to be abnormal. Therefore, the abnormal trend factor can characterize the abnormal situation of traffic flow. The larger the abnormal trend factor, the greater the abnormal degree of traffic flow. Therefore, the abnormal trend factor of traffic flow is one of the key indicators for determining the degree of abnormality in the subsequent determination.

[0067] In this embodiment, the first preset number of vehicle flows closest to the target vehicle flow is first used as the adjacent vehicle flows, and each adjacent vehicle flow of the target vehicle flow can be obtained. Then, according to the difference between the target vehicle flow and each adjacent vehicle flow, the abnormal factor of the change trend of the target vehicle flow is determined. Among them, the first preset number can be set to 10, and the implementer can set the size of the first preset number according to the specific actual situation.

[0068] As an example, the calculation formula of the abnormal factor of the change trend of the target vehicle flow can be:

[0069] ; In the formula, is the abnormal factor of the change trend of the target traffic flow, norm() is the linear normalization function, v is the vth neighboring traffic flow of the target traffic flow, V is the number of neighboring traffic flows, is the target traffic volume, is the vth neighboring traffic flow of the target traffic flow, To find the absolute value function, is the ceiling function.

[0070] In the calculation formula of the abnormal factor of the change trend, It indicates the difference between the target traffic flow and its 10 neighboring traffic flows. The larger the difference, the greater the abnormal factor of the change trend, and vice versa. The distance weight can be represented. The closer the adjacent traffic flow is to the target traffic flow, the greater the weight is. The sequence number v of the adjacent traffic flow in the first traffic flow data sequence is divided by 2 and then rounded up to ensure that the distance weights of the two adjacent traffic flows located on the left and right of the target traffic flow and at the same distance are consistent.

[0071] In the second step, the first traffic flow data sequence is divided into traffic flow data subsequences corresponding to each day, and the traffic flow in the traffic flow data subsequence with the same position as the target traffic flow is determined, recorded as the periodic traffic flow, and then the abnormal factor of the change trend of the periodic traffic flow is determined.

[0072] Affected by people's daily living habits, daily traffic flow shows a strong periodicity, that is, the changes in traffic flow at the same time every day are basically similar. If a certain traffic flow performs poorly in periodicity, the possibility of abnormality in this traffic flow is greater.

[0073] In this embodiment, the first traffic flow data sequence includes traffic flow data for 7 days. The traffic flow of the same day is divided together to obtain the traffic flow data subsequence corresponding to each day. The traffic flow at the same position as the target traffic flow is determined in the traffic flow data subsequence. For example, if the target traffic flow is collected at 12 o'clock, the traffic flow at 12 o'clock every day is obtained and recorded as the periodic traffic flow. In order to facilitate the subsequent determination of the periodic abnormal factors, the abnormal factors of the change trend of each periodic traffic flow are determined with reference to the calculation process of the abnormal factors of the change trend of the target traffic flow.

[0074] It should be noted that the collection time of each vehicle flow in the vehicle flow data subsequence of each day is the same, and the number of vehicle flow data in different vehicle flow data subsequences is the same.

[0075] The third step is to determine the periodic abnormality factor of the target traffic flow according to the target traffic flow, the traffic flow of each period, and the abnormality factor of the change trend of the traffic flow of each period. The specific implementation steps may include:

[0076] In the first sub-step, negative correlation processing is performed on the abnormal factors of the change trend of the traffic flow in each period to obtain the preliminary weight of the traffic flow in each period, and the ratio of the preliminary weight to the cumulative sum of all preliminary weights is used as the final weight.

[0077] In this embodiment, the exponential function exp(-) can be used for negative correlation processing. Negative correlation means that the dependent variable will decrease as the independent variable increases, and the dependent variable will increase as the independent variable decreases. It can be a subtraction relationship, a division relationship, etc., which is determined by actual applications.

[0078] In the second sub-step, the cumulative sum of the products of each periodic traffic flow and its corresponding final weight is used as the weighted value of all periodic traffic flows, and the difference between the target traffic flow and the weighted values ​​of all periodic traffic flows is calculated as the initial periodic abnormality factor.

[0079] In the third sub-step, the initial period anomaly factor is normalized, and the normalized initial period anomaly factor is used as the period anomaly factor of the target vehicle flow.

[0080] In this embodiment, the normalization process can be specifically a linear normalization function, and the normalization in subsequent steps can all use a linear normalization function. In other embodiments of the present application, other normalization methods can be selected according to the specific range of numerical values, which will not be described in detail.

[0081] As an example, the calculation formula of the periodic abnormality factor of the target vehicle flow can be:

[0082] ; In the formula, is the periodic anomaly factor of the target traffic flow, norm[] is the linear normalization function, is the target traffic flow, i is the sequence number of the periodic traffic flow, I is the number of periodic traffic flows, is the i-th cycle traffic flow of the target traffic flow, is the abnormal factor of the change trend of the target traffic flow in the i-th period, is the initial weight of the target traffic flow in the ith period, is the final weight of the target traffic flow in the i-th period, is the weighted value of all periodic traffic flows of the target traffic flow, To find the absolute value function.

[0083] In the calculation formula of the cycle anomaly factor, the final weight can represent the abnormal change trend of the target traffic flow in the i-th cycle. The larger the change trend anomaly factor, the greater the abnormal degree of data change, indicating that the i-th cycle traffic flow accounts for a smaller proportion in the weighted value calculation process; It can represent the difference between the target traffic flow and the traffic flow at the same time every day. The larger the difference, the greater the degree of periodic anomaly of the target traffic flow, and vice versa. The periodic anomaly factor can characterize the degree of abnormality of the periodic characteristics of the target traffic flow.

[0084] The fourth step is to determine the abnormal degree of the target traffic flow by combining the abnormal factor of the change trend of the target traffic flow and the abnormal factor of the period.

[0085] In this embodiment, the degree of abnormality can characterize the abnormal situation of traffic flow when measuring data and the abnormal situation when it is affected by unexpected traffic accidents. The lower the degree of abnormality, the better the quality of traffic flow when predicted; the change trend abnormality factor and the periodic abnormality factor are both positively correlated with the degree of abnormality. The positive correlation indicates that the dependent variable will increase as the independent variable increases, and the dependent variable will decrease as the independent variable decreases. The specific relationship can be a multiplication relationship, an addition relationship, and the power of an exponential function, which is determined by the actual application.

[0086] As an example, the calculation formula for the abnormality degree of the target vehicle flow may be:

[0087] ; In the formula, is the abnormal degree of target traffic flow, is the abnormal factor of the change trend of the target traffic flow, is the periodic anomaly factor of the target traffic flow.

[0088] So far, this embodiment has obtained the abnormality degree of each vehicle flow in each first vehicle flow data sequence.

[0089] S3, according to the abnormal degree of each traffic flow and the correlation between the traffic flows of adjacent intersections, each parameter in the adaptive ARIMA model is used to obtain a traffic flow prediction model for each intersection in each direction.

[0090] It should be noted that the known ARIMA model can be regarded as a direct combination of the AR model (Autoregressive Models) and the MA model (Moving Average Model), and there are three parameters that need to be considered in the ARIMA model, namely: the parameters of the AR part, which are used to represent the weight of a certain original time series data during prediction; the parameters of the MA part, which are used to represent the weight of the residual or error of a certain original time series data during prediction; the constant term, which represents the basic value of the original time series data. Among them, the parameters of the AR part and the MA part can be analyzed by the abnormal degree of traffic flow, and the constant term needs to be obtained in combination with the correlation between the traffic flows at adjacent intersections.

[0091] First, the parameters of each AR part and each MA part of the traffic flow prediction model in each direction at each intersection are determined. The specific implementation steps may include:

[0092] In the first step, any intersection is taken as the target intersection, and any direction of the target intersection is taken as the target direction. The previous intersection and the next intersection of the target intersection are selected in the target direction. The direction in which vehicles travel toward the target direction is called the convergence direction, and the direction in which vehicles travel away from the target direction is called the divergence direction.

[0093] In this embodiment, it is assumed that all roads in the monitored urban area are interconnected. Each first traffic flow data sequence has its corresponding traffic flow prediction model. For the convenience of description, a traffic flow prediction model is analyzed, so any intersection of all intersections is selected as the target intersection, and any direction of the target intersection is selected as the target direction to construct a traffic flow prediction model in the target direction of the target intersection.

[0094] In the subsequent steps, when analyzing the parameters of each AR part and the parameters of each MA part, for ease of understanding, the previous intersection is defined as the previous intersection of the target intersection in the target direction, the next intersection is defined as the next intersection of the target intersection in the target direction, the convergence direction is defined as the direction in which vehicles are traveling toward the target direction, and the divergence direction is defined as the direction in which vehicles are traveling away from the target direction.

[0095] in, Figure 3 is a distribution diagram of the target intersection. Figure 3 middle, represents the target direction of the target intersection, Indicates the first divergence direction of the target intersection. Indicates the second diversion direction of the target intersection. , and Indicates the three convergence directions of the previous intersection of the target intersection, Indicates the convergence direction of the next intersection after the target intersection. and Indicates the two diverging directions of the next intersection after the target intersection.

[0096] In the second step, the parameters of each AR part and each MA part of the traffic flow prediction model for the target intersection in the target direction are determined based on the traffic flows and their abnormality levels in each converging direction of the previous intersection, the traffic flows and their abnormality levels in the target direction and diverging directions of the target intersection, and the traffic flows and their abnormality levels in the converging direction and diverging directions of the next intersection.

[0097] In the AR part, each item is the original time series data, and the parameters of the time series data can represent the relationship between the time series data and the predicted value; in the MA part, each item is the residual size of the original time series data, and the residual parameters describe the relationship between the residual and the predicted value. When the authenticity of the time series data is strong, the reference value in the prediction is higher, and the corresponding parameters of the two AR parts and the MA part are also larger. Therefore, the size of the corresponding parameters can be calculated by the abnormality of the traffic flow. The greater the abnormality of the traffic flow, the lower the credibility of the prediction, and the smaller the parameters of the AR part and the MA part. The parameters of the AR part and the MA part of the same traffic flow are equal.

[0098] However, the degree of abnormality of traffic flow cannot directly represent the authenticity of the data. When special circumstances occur on the road, such as traffic accidents, events, etc., the traffic flow may suddenly change, resulting in a sudden increase in the degree of abnormality of traffic flow. The corresponding traffic flow data when special circumstances occur are also real data, which has a relatively important impact on the prediction of traffic flow in the future. The weight of traffic flow in special circumstances should be appropriately increased in the prediction process. Therefore, it is necessary to quantify the noise degree of each traffic flow. The greater the noise degree, the less likely it is that the traffic flow is a real abnormality, that is, the less likely the abnormality of the traffic flow is caused by special circumstances.

[0099] Depend on Figure 3 It can be seen that the traffic flow in the target direction at the target intersection at the collection time t is strongly correlated with the traffic flows in the three converging directions of the previous intersection of the target intersection, and one converging direction and two diverging directions of the next intersection of the target intersection. When the traffic flow in the target direction at the collection time t shows a high degree of abnormality, if the traffic flows at the previous intersection and the next intersection also produce similar abnormalities, it means that the traffic flow in the target direction at the collection time t is The change is a real exception, that is, an exception caused by a special situation. Otherwise, it indicates It is very likely that the noise data is generated during the measurement. Therefore, the first noise factor and the second noise factor can be quantified by analyzing the similar situations of the changes in the traffic volume and the abnormal degree of the adjacent intersections. The specific implementation steps may include:

[0100] In the first sub-step, the collection time of any vehicle flow in the target direction at the target intersection is recorded as t; based on the vehicle flow in the target direction and the diverging direction of the target intersection at the collection time t, the vehicle flow in each converging direction of the previous intersection at the collection time tb, and the vehicle flow in the converging direction and the diverging direction of the next intersection at the collection time t+c, the first noise factor of the vehicle flow in the target direction at the collection time t is determined.

[0101] As an example, the calculation formula for determining the first noise factor of the vehicle flow in the target direction at the target intersection at the collection time t may be:

[0102] ; In the formula, is the first noise factor of the traffic flow in the target direction at the target intersection at the acquisition time t, is the traffic flow in the target direction at the target intersection at the collection time t, , and are the target direction, the first diversion direction, and the second diversion direction of the target intersection respectively. is the cumulative sum of the traffic flow in the converging direction and diverging direction at the next intersection after the collection time t+c, C is the next intersection after the target intersection, c is the average driving time from the target intersection to the next intersection, is the cumulative sum of the traffic volume in each converging direction of the previous intersection at the collection time tb, b is the average driving time from the previous intersection to the target intersection, B is the previous intersection of the target intersection, is the traffic flow in the first diverging direction of the target intersection at the collection time t, is the traffic flow in the second diverging direction of the target intersection at the collection time t, To find the absolute value function, is a hyperparameter.

[0103] In the calculation formula of the first noise factor, the hyperparameter To prevent the special case when the denominator of the fraction is zero, the hyperparameter can be set to 1; and It can represent the consistency between the traffic flow in the target direction and the traffic flow on the way. The closer the ratio is to 1, the stronger the consistency is. and The smaller it is, the more consistent the traffic flow in the target direction is with the oncoming traffic flow, and the smaller the first noise factor is, that is, the possibility that the traffic flow in the target direction of the target intersection at the collection time t is noise data is smaller.

[0104] In the second sub-step, the second noise factor of the traffic flow in the target direction at the target intersection at the collection time t is determined based on the abnormal degree of traffic flow in the target direction at the collection time t, the abnormal degree of traffic flow in each converging direction at the previous intersection at the collection time tb, and the abnormal degree of traffic flow in the converging direction and the diverging direction at the next intersection at the collection time t+c.

[0105] As an example, the calculation formula for determining the second noise factor of the vehicle flow in the target direction at the target intersection at the collection time t may be:

[0106] ; In the formula, is the second noise factor of the traffic flow in the target direction at the target intersection at the acquisition time t, is the abnormal degree of traffic flow in the target direction at the target intersection at the acquisition time t, is the average value of the abnormal degree of traffic flow in the converging and diverging directions at the next intersection after the collection time t+c, is the average value of the abnormal degree of traffic flow in each converging direction at the previous intersection at the collection time tb, To find the absolute value function.

[0107] In the calculation formula of the second noise factor, and Respectively The degree of abnormality and , The greater the difference, the more similar the abnormal state of the traffic flow in the target direction is to that in the traveling direction, and the smaller the second noise factor is, that is, the smaller the possibility that the traffic flow in the target direction of the target intersection at the collection time t is noise data.

[0108] The third sub-step is to calculate the product of the abnormal degree of the vehicle flow in the target direction at the target intersection at the collection time t and the first noise factor and the second noise factor, normalize the product, and use the normalized product as the noise degree, thereby obtaining the noise degree of each vehicle flow in the target direction at the target intersection.

[0109] As an example, the calculation formula for the noise level of the traffic flow in the target direction at the target intersection at the acquisition time t can be:

[0110] ; In the formula, is the noise level of the traffic flow in the target direction at the target intersection at the acquisition time t, norm[] is the linear normalization function, is the abnormal degree of traffic flow in the target direction at the target intersection at the acquisition time t, is the first noise factor of the traffic flow in the target direction at the target intersection at the acquisition time t, It is the second noise factor of the traffic flow in the target direction at the target intersection at the collection time t.

[0111] In the calculation formula of the noise level, the smaller the abnormality of the traffic flow in the target direction at the target intersection at the collection time t, the more consistent the traffic flow in the target direction is with the traffic flow in the traveling direction, and the more similar the states of the traffic flow in the target direction are with the traffic flow in the traveling direction, the smaller the noise level of the traffic flow in the target direction at the collection time t is.

[0112] After obtaining the noise level of each vehicle flow at the target intersection in the target direction, the parameters of the AR part and the MA part of the corresponding vehicle flow in the vehicle flow prediction model are determined based on the noise level. The specific implementation steps may include:

[0113] Since the greater the noise level of the traffic flow, the lower the reference value in the prediction model, the noise level of each traffic flow at the target intersection in the target direction is negatively correlated to obtain the initial parameters; the ratio of the initial parameters to the cumulative sum of all initial parameters is used as the parameters of the AR part and the MA part of the corresponding traffic flow, and then the parameters of each AR part and each MA part of the traffic flow prediction model at the target intersection in the target direction are obtained. Among them, each traffic flow has its corresponding AR part parameters and MA part parameters.

[0114] As an example, the calculation formula for the parameters of the AR part and the MA part of the traffic flow prediction model of the target intersection in the target direction at the acquisition time t can be:

[0115] ; In the formula, is the parameter of the AR part or the parameter of the MA part in the traffic flow prediction model of the target intersection in the target direction at the collection time t, Q is the total number of collection times, exp() is an exponential function with a natural constant as the base, is the noise level of the traffic flow in the target direction at the target intersection at the acquisition time t, is the initial parameters of the AR part or the initial parameters of the MA part in the traffic flow prediction model of the target intersection in the target direction at the acquisition time t, is the cumulative sum of all initial parameters.

[0116] So far, this embodiment has obtained the parameters of the AR part and the MA part of the traffic flow prediction model for each intersection in each direction.

[0117] Secondly, the constant term of the traffic flow prediction model for each intersection in each direction is determined.

[0118] It should be noted that the constant term in the ARIMA model usually represents the overall level of the time series, which can be understood as an offset to compensate for possible systematic deviations in the prediction model. In a real road traffic environment, the randomness of traffic flow changes is very strong. When predicting, we cannot only consider the existing data, but also take into account possible deviations. Therefore, it is necessary to dynamically adjust the constant term according to the changes in traffic flow data to improve the flexibility and prediction accuracy of the traffic flow prediction model. The basis for adjusting the constant term is the traffic flow deviation in all relevant directions before the current prediction direction. The more correlated the traffic flow, the greater the impact.

[0119] In this embodiment, taking the prediction of the traffic flow in the target direction of the target intersection at the next moment when the collection moment is the current moment as an example, the parameter items of the traffic flow prediction model of the target intersection in the target direction are determined, and the specific implementation steps may include:

[0120] In the first step, a second preset number of intersections located before and closest to the target intersection are selected in the target direction as related intersections of the target intersection, and the actual distance between each related intersection of the target intersection and the target intersection is obtained.

[0121] In this embodiment, the second preset number can be set to 5. The implementer can set the size of the second preset number according to the actual situation, and no specific limitation is made. The closer the actual distance between the relevant intersection and the target intersection, the greater the impact of the traffic flow of the relevant intersection on the traffic flow of the target intersection. Therefore, it is necessary to obtain the actual distance between each relevant intersection and the target intersection to provide data support for the subsequent calculation of the constant term. The unit of the actual distance can be kilometers.

[0122] The second step is to determine the constant term of the traffic flow prediction model of the target intersection in the target direction according to the traffic flow and its abnormality at each related intersection of the target intersection in each convergence direction and the actual distance between each related intersection of the target intersection and the target intersection.

[0123] As an example, the calculation formula for determining the constant term of the vehicle flow prediction model at the target intersection in the target direction may be:

[0124] ; In the formula, is the constant term of the traffic flow prediction model of the target intersection in the target direction, N is the number of related intersections corresponding to the target intersection, X is the sequence number of the related intersections corresponding to the target intersection, max[] is the maximum value function, T is the next moment after the current moment, is the accumulated value of the abnormal degree of traffic flow in each converging direction at the Xth related intersection at the collection time Tx, is the accumulated value of the noise level of the traffic flow in each converging direction at the Xth related intersection at the collection time Tx, x is the average driving time from the Xth target intersection to the target intersection, is the accumulated value of the traffic flow in each converging direction at the Xth related intersection at the collection time Tx, is the average value of the accumulated traffic flow in each converging direction at the Xth related intersection at all collection times, is the actual distance between the Xth related intersection and the target intersection, and exp() is an exponential function with a natural constant as the base.

[0125] In the calculation formula of the constant term, means The difference between the abnormality degree and the noise degree can be represented as the real abnormality degree when a special situation occurs at the relevant intersection. The larger the real abnormality degree, the more likely the traffic flow in each convergence direction of the relevant intersection is to deviate, and the larger the constant term should be. It can represent the deviation of the accumulated value of the traffic flow in each converging direction at the Xth related intersection at the collection time Tx. The larger it is, the larger the constant term will be; It can represent the correlation between the Xth related intersection and the target intersection. The smaller the actual distance is, the stronger the correlation is. The higher the credibility of the deviation analyzed based on the Xth related intersection is, the larger the constant term is.

[0126] At this point, all parameter sizes of the traffic flow prediction model at the target intersection in the target direction have been obtained. Subsequently, the traffic flow prediction model can be adaptively constructed according to the parameters to obtain the traffic flow prediction model for all intersections in each direction to predict the traffic flow data at the next moment from the current moment.

[0127] S4, obtaining the second traffic flow data sequence corresponding to each intersection in different directions with the current set duration, inputting the second traffic flow data sequence into the corresponding traffic flow prediction model, and obtaining the prediction results of each intersection in different directions; according to the prediction results, adjusting the duration of the traffic lights at each intersection.

[0128] In this embodiment, the traffic flow prediction model is an ARIMA model with adaptively determined parameters, and each direction of each intersection has its corresponding traffic flow prediction model. The current setting duration can be set to 20 minutes, and the sequence consisting of the traffic flow in the same direction of the same intersection within the current setting duration at the current moment is used as the second traffic flow data sequence, and the second traffic flow data sequence is input into the corresponding traffic flow prediction model, and the prediction results of each intersection in different directions can be obtained, that is, the traffic flow of each intersection in each direction at the next moment of the current moment, that is, the predicted traffic flow. Among them, the implementation process of the ARIMA model is a prior art, which is not within the scope of protection of this application, and will not be elaborated here; the current setting duration can be set by the implementer according to the specific actual situation, and no specific limitation is made here. It should be noted that the form of the input data of the traffic flow prediction model is consistent with the data form of the output result, and the input data can be the traffic flow corresponding to 3 minutes in step S1, or the pending traffic flow corresponding to each second. The form of the input data of the traffic flow prediction model can be set by the implementer according to the specific actual situation, and no specific limitation is made here.

[0129] After obtaining the prediction results, the predicted traffic flow data is used to adjust the duration of the traffic light. The specific adjustment basis can be the traffic flow ratio in different directions of the intersection. Assuming that the initial ratio of the green light duration in the east-west direction to the north-south direction of the target intersection is d, the calculation formula for the traffic flow adjustment coefficient when adjusting the traffic light of the target intersection can be:

[0130] ; In the formula, is the traffic flow adjustment coefficient of the target intersection, A is the target intersection, is the prediction result of the east-west direction of the target intersection at the next moment from the current moment. is the prediction result of the target intersection in the north-south direction at the next moment from the current moment, d is the initial ratio of the green light duration in the east-west direction to the north-south direction of the target intersection, To find the absolute value function.

[0131] After obtaining the traffic flow adjustment coefficient of the target intersection, the adjustment threshold is set as ,when Greater than or equal to When , it means that the deviation between the traffic flow ratio and the initial ratio is too large, and the traffic light duration of the target intersection at the next moment of the current moment needs to be adjusted to ensure that the ratio of the green light duration in the east-west and north-south directions of the target intersection is equal to the traffic flow ratio in the east-west and north-south directions of the target intersection; when Less than When the traffic light duration at the target intersection is not adjusted.

[0132] Of course, there are many other ways to adjust the duration of traffic lights based on the prediction results, which will not be described in detail in this embodiment. The implementer can select a suitable adjustment method according to the specific actual situation.

[0133] Thus, the present embodiment realizes accurate adjustment of the duration of traffic lights, thereby improving the traffic efficiency of the road.

[0134] The embodiment of the present application provides a monitoring and adjustment method for an intelligent transportation system based on the Internet of Things. The ARIMA algorithm is used to combine the time series characteristics of vehicle flow data and the correlation between different vehicle flows to adaptively predict the model parameters, and a vehicle flow prediction model is constructed to obtain a relatively accurate vehicle flow prediction result. This provides decision support for the strategy of traffic light control, and at the same time helps to achieve real-time regulation of the duration of traffic lights at each intersection, thereby improving the accuracy of system monitoring and regulation, and further improving the traffic efficiency of the road.

[0135] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A monitoring and adjustment method for a smart transportation system based on the Internet of Things, characterized in that: The following steps are involved: Acquire in real time a first traffic flow data sequence in different directions at all intersections in the monitored urban area for several days, wherein the first traffic flow data sequence includes the traffic flow in each preset time period; Analyze the change trend and periodicity characteristics of the traffic flow according to each traffic flow in the first traffic flow data sequence to which each traffic flow belongs, and determine the abnormality degree of each traffic flow; According to the abnormal degree of each traffic flow and the traffic flow of adjacent intersections, each parameter in the adaptive ARIMA model is used to obtain a traffic flow prediction model for each intersection in each direction; Obtaining a second traffic flow data sequence corresponding to each intersection in different directions with a currently set duration, inputting the second traffic flow data sequence into a corresponding traffic flow prediction model, and obtaining prediction results for each intersection in different directions; adjusting the duration of the traffic light at each intersection according to the prediction results; The step of analyzing the change trend and periodicity characteristics of the traffic flow according to each traffic flow in the first traffic flow data sequence to which each traffic flow belongs, and determining the abnormality degree of each traffic flow includes: Taking any vehicle flow rate as a target vehicle flow rate, and determining an abnormal factor of a change trend of the target vehicle flow rate according to the target vehicle flow rate and a first preset number of vehicle flows closest to the target vehicle flow rate; Divide the first vehicle flow data sequence into vehicle flow data subsequences corresponding to each day, determine the vehicle flow in the vehicle flow data subsequence that is at the same position as the target vehicle flow, record it as the periodic vehicle flow, and then determine the abnormal factor of the change trend of the periodic vehicle flow; Determine the periodic abnormal factor of the target traffic flow according to the target traffic flow, the traffic flow of each period and the abnormal factor of the change trend of the traffic flow of each period; The abnormal degree of the target vehicle flow is determined by combining the abnormal trend factor and the period abnormal factor of the target vehicle flow, and both the abnormal trend factor and the period abnormal factor are positively correlated with the abnormal degree.

2. The method for monitoring and regulating an intelligent transportation system based on the Internet of Things according to claim 1 is characterized in that: The step of determining the abnormal factor of the change trend of the target vehicle flow rate according to the target vehicle flow rate and the first preset number of vehicle flows closest to the target vehicle flow rate includes: Taking a first preset number of vehicle flows closest to the target vehicle flow as adjacent vehicle flows, and obtaining each adjacent vehicle flow of the target vehicle flow; The calculation formula of the abnormal factor of the change trend of target traffic flow is: ; In the formula, is the abnormal factor of the change trend of the target traffic flow, () is a linear normalization function, v is the vth neighboring vehicle flow of the target vehicle flow, V is the number of neighboring vehicle flows, is the target traffic volume, is the vth neighboring traffic flow of the target traffic flow, To find the absolute value function, is the ceiling function.

3. The method for monitoring and regulating an intelligent transportation system based on the Internet of Things according to claim 1 is characterized in that: Determining the periodic abnormality factor of the target vehicle flow according to the target vehicle flow, each periodic vehicle flow and the abnormality factor of the change trend of each periodic vehicle flow includes: Perform negative correlation processing on the abnormal factors of the change trend of the traffic flow in each period, obtain the preliminary weight of the traffic flow in each period, and take the ratio of the preliminary weight to the cumulative sum of all preliminary weights as the final weight; The product of each periodic traffic flow and its corresponding final weight is accumulated over all periodic traffic flows as the weighted value of all periodic traffic flows, and the difference between the target traffic flow and the weighted value of all periodic traffic flows is calculated as the initial periodic abnormality factor; The initial cycle anomaly factor is normalized and used as the cycle anomaly factor of the target vehicle flow.

4. The method for monitoring and regulating an intelligent transportation system based on the Internet of Things according to claim 1 is characterized in that: The parameters in the adaptive ARIMA model according to the abnormal degree of each traffic flow and the traffic flow of adjacent intersections include: Take any intersection as the target intersection, take any direction of the target intersection as the target direction, select the previous intersection and the next intersection of the target intersection in the target direction, call the direction of vehicles moving toward the target direction the convergence direction, and call the direction of vehicles moving away from the target direction the divergence direction; According to each vehicle flow rate and its abnormality degree in each converging direction of the previous intersection, each vehicle flow rate and its abnormality degree in the target direction and diverging direction of the target intersection, and each vehicle flow rate and its abnormality degree in the converging direction and diverging direction of the next intersection, the parameters of each AR part and each MA part of the vehicle flow prediction model in the target direction of the target intersection are determined; Selecting a second preset number of intersections that are located before the target intersection and closest to the target intersection in the target direction as related intersections of the target intersection, and obtaining actual distances between each related intersection of the target intersection and the target intersection; According to each traffic flow and its abnormality at each related intersection of the target intersection in each converging direction, and the actual distance between each related intersection of the target intersection and the target intersection, the constant term of the traffic flow prediction model of the target intersection in the target direction is determined.

5. The method for monitoring and regulating an intelligent transportation system based on the Internet of Things according to claim 4 is characterized in that: The parameters of each AR part and each MA part of the vehicle flow prediction model for determining the target intersection in the target direction include: The collection time of any vehicle flow in the target direction at the target intersection is recorded as t; according to the vehicle flow in the target direction and the diverging direction of the target intersection at the collection time t, the vehicle flow in each converging direction of the previous intersection at the collection time tb, and the vehicle flow in the converging direction and the diverging direction of the next intersection at the collection time t+c, the first noise factor of the vehicle flow in the target direction of the target intersection at the collection time t is determined; wherein b is the average driving time from the previous intersection to the target intersection, and c is the average driving time from the target intersection to the next intersection; Determine the second noise factor of the traffic flow in the target direction at the target intersection at the collection time t according to the abnormal degree of the traffic flow in the target direction at the collection time t, the abnormal degree of the traffic flow in each converging direction at the previous intersection at the collection time tb, and the abnormal degree of the traffic flow in the converging direction and the diverging direction at the next intersection at the collection time t+c; Calculate the product of the abnormal degree of the vehicle flow in the target direction at the target intersection at the collection time t and the first noise factor and the second noise factor, normalize the product, and use the normalized product as the noise degree, thereby obtaining the noise degree of each vehicle flow in the target direction at the target intersection; The noise levels of each vehicle flow at the target intersection in the target direction are negatively correlated to obtain initial parameters; the ratio of the initial parameters to the cumulative sum of all initial parameters is used as the parameters of the AR part and the MA part of the corresponding vehicle flow, and then the parameters of each AR part and the parameters of each MA part of the vehicle flow prediction model at the target intersection in the target direction are obtained; wherein, the parameters of the AR part and the parameters of the MA part of the same vehicle flow are equal.

6. The method for monitoring and regulating an intelligent transportation system based on the Internet of Things according to claim 5 is characterized in that: The first noise factor of the vehicle flow in the target direction at the target intersection at the acquisition time t is determined, including: ; In the formula, is the first noise factor of the traffic flow in the target direction at the target intersection at the acquisition time t, is the traffic flow in the target direction at the target intersection at the collection time t, , and are the target direction, the first diversion direction, and the second diversion direction of the target intersection respectively. is the cumulative sum of the traffic flow in the converging direction and diverging direction at the next intersection after the collection time t+c, C is the next intersection after the target intersection, c is the average driving time from the target intersection to the next intersection, is the cumulative sum of the traffic volume in each converging direction of the previous intersection at the collection time tb, b is the average driving time from the previous intersection to the target intersection, B is the previous intersection of the target intersection, is the traffic flow in the first diverging direction of the target intersection at the collection time t, is the traffic flow in the second diverging direction of the target intersection at the collection time t, To find the absolute value function, is a hyperparameter.

7. The method for monitoring and regulating an intelligent transportation system based on the Internet of Things according to claim 6 is characterized in that: The second noise factor of the vehicle flow in the target direction at the target intersection at the acquisition time t is determined, including: ; In the formula, is the second noise factor of the traffic flow in the target direction at the target intersection at the acquisition time t, is the abnormal degree of traffic flow in the target direction at the target intersection at the acquisition time t, is the average value of the abnormal degree of traffic flow in the converging and diverging directions at the next intersection after the collection time t+c, is the average value of the abnormal degree of traffic flow in each converging direction at the previous intersection at the collection time tb, To find the absolute value function.

8. The method for monitoring and regulating an intelligent transportation system based on the Internet of Things according to claim 5 is characterized in that: The method of determining the constant term of the vehicle flow prediction model of the target intersection in the target direction according to each vehicle flow and its abnormality at each related intersection of the target intersection in each converging direction and the actual distance between each related intersection of the target intersection and the target intersection comprises: ; In the formula, is the constant term of the traffic flow prediction model of the target intersection in the target direction, N is the number of related intersections corresponding to the target intersection, X is the sequence number of the related intersections corresponding to the target intersection, max[] is the maximum value function, T is the next moment after the current moment, is the accumulated value of the abnormal degree of traffic flow in each converging direction at the Xth related intersection at the collection time Tx, is the accumulated value of the noise level of the traffic flow in each converging direction at the Xth related intersection at the collection time Tx, x is the average driving time from the Xth target intersection to the target intersection, is the accumulated value of the traffic flow in each converging direction at the Xth related intersection at the collection time Tx, is the average value of the accumulated traffic flow in each converging direction at the Xth related intersection at all collection times, is the actual distance between the Xth related intersection and the target intersection, and exp() is an exponential function with a natural constant as the base.

9. The method for monitoring and regulating an intelligent transportation system based on the Internet of Things according to claim 1, characterized in that: The step of acquiring each vehicle flow in the first vehicle flow data sequence comprises: For any direction of any intersection, first determine the ratio of the green light duration at the intersection to the preset time period, then determine half of the original traffic flow data of the intersection in that direction, and take the ratio of half of the original traffic flow data to the ratio as the traffic flow of the intersection in that direction.