A traffic situation assessment method based on traffic flow parameter prediction

Abstract

This invention discloses a traffic situation assessment method based on traffic flow parameter prediction. The method collects raw traffic flow data of a target road or road segment using traffic detection equipment. This raw traffic flow data includes at least traffic volume, average speed, and road occupancy. The collected data is preprocessed to obtain a continuous and stable traffic flow time series. Based on the traffic flow time series, a traffic flow parameter feature set is constructed, and a traffic flow parameter prediction model is established to predict traffic flow parameters within a preset future time window. Furthermore, based on the predicted traffic flow parameters, a traffic situation assessment model is constructed to comprehensively evaluate the traffic operation status of the target road or road segment within the time window, obtaining a traffic situation assessment result. The corresponding traffic situation level is then determined according to preset rules. The traffic situation assessment result can be used for traffic guidance dissemination, signal control optimization, congestion warning, and traffic operation management decision support.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, and more specifically, to a traffic situation assessment method based on traffic flow parameter prediction. Background Technology

[0002] With the continuous advancement of urbanization and the sustained growth of motor vehicle ownership, the operational status of urban road traffic systems is becoming increasingly complex, and problems such as traffic congestion, declining traffic efficiency, and unstable traffic operation are gradually becoming prominent. Traffic management departments need to grasp the road traffic operation status in a timely and accurate manner and make reasonable assessments of the traffic situation to support traffic guidance, signal control, congestion warning, and operation management. Therefore, traffic situation assessment technology has become a key component of intelligent transportation systems.

[0003] Most existing traffic situation assessment methods are based on the analysis of real-time or historical traffic flow data. Common assessment indicators include traffic volume, average speed, road occupancy, or traffic density. These methods typically determine whether a road is in a smooth or congested state by comparing the traffic flow parameters collected at the current moment with preset thresholds or by classifying road operating conditions according to empirical rules. These methods are simple to implement and have low computational costs, and have been widely used in early traffic operation monitoring.

[0004] However, in actual traffic operations, traffic conditions exhibit significant time-varying and lagging characteristics. Assessments relying solely on current or historical traffic flow parameters are insufficient to reflect impending changes in traffic conditions. For instance, in the critical phase of traffic flow transitioning from smooth to congested, assessments based solely on current observation data often only provide results after congestion has already occurred, lacking an effective depiction of future traffic trend changes and failing to meet the needs of traffic management departments for proactive and forward-looking interventions.

[0005] Furthermore, some existing technologies in traffic situation assessment often focus on a single traffic flow parameter or simply overlay multiple traffic flow parameters, failing to fully consider the inherent relationships between parameters such as traffic flow, speed, density, and occupancy. Under conditions of significant traffic fluctuations or substantial external disturbances, such assessment methods are prone to unstable results and insufficient accuracy. Simultaneously, existing traffic situation assessment methods lack adaptive adjustment mechanisms to address deviations between predicted results and actual operating conditions, resulting in limited adaptability of the assessment models to changes in the traffic operating environment.

[0006] Therefore, how to fully utilize existing traffic detection facilities, introduce predictions of future trends in traffic flow parameters, and combine multiple traffic flow parameters for comprehensive analysis to achieve dynamic and forward-looking assessment of road traffic conditions has become one of the urgent technical problems to be solved in the fields of traffic engineering and intelligent transportation.

[0007] Therefore, there is an urgent need for a traffic situation assessment method based on traffic flow parameter prediction to solve these problems. Summary of the Invention

[0008] The purpose of this invention is to solve the technical problems mentioned in the background section and to provide a traffic situation assessment method based on traffic flow parameter prediction, comprising the following steps: Collect raw traffic flow data for the target road or road segment, including at least traffic volume data, average driving speed data, and road occupancy data; The raw traffic flow data is preprocessed to obtain continuous traffic flow time series data; A traffic flow parameter feature set is constructed based on the traffic flow time series data; A traffic flow parameter prediction model is established using the traffic flow parameter feature set to predict the traffic flow parameters within a preset time window in the future, thereby obtaining the predicted traffic flow parameters. Based on the predicted traffic flow parameters, a traffic situation assessment model is constructed to assess the traffic operation status of the target road or road segment within the preset time window, and the traffic situation assessment result is obtained. The corresponding traffic situation level is output based on the traffic situation assessment results.

[0009] As a preferred technical solution of the present invention, the original traffic flow data is collected by a geomagnetic detector, a video detection device, a radar detection device, or a combination thereof, and the collection period is a fixed time interval.

[0010] As a preferred technical solution of the present invention, the preprocessing of the raw traffic flow data includes: Time synchronization processing is performed on raw traffic flow data from different data sources; Identify and remove abnormal data; Compensate and repair missing data; The processed data is smoothed and filtered to reduce the impact of random fluctuations on subsequent analysis.

[0011] As a preferred technical solution of the present invention, the traffic flow parameter feature set includes at least traffic flow parameters, traffic speed parameters and traffic density parameters, wherein the traffic density parameter is calculated based on the traffic flow parameters and traffic speed parameters.

[0012] As a preferred technical solution of the present invention, the traffic flow parameter prediction model is a short-term prediction model, which is used to predict the trend of traffic flow parameter changes in the next few minutes to tens of minutes.

[0013] As a preferred technical solution of the present invention, the traffic flow parameter prediction model is trained based on historical traffic flow time series data and can dynamically update the prediction results according to the latest collected traffic flow parameters.

[0014] As a preferred technical solution of the present invention, the traffic situation assessment model calculates the traffic situation assessment value by weighting and fusing the predicted traffic flow parameters, predicted traffic speed parameters, and predicted traffic density parameters.

[0015] As a preferred technical solution of the present invention, the weights of each traffic flow parameter in the weighted fusion process are set or adaptively adjusted according to road type, traffic operation characteristics or historical operation data.

[0016] As a preferred technical solution of the present invention, the traffic situation level includes at least smooth traffic, light congestion, moderate congestion and severe congestion, and the traffic situation level is obtained by matching the traffic situation assessment results with a preset threshold range.

[0017] As a preferred technical solution of the present invention, the traffic situation assessment results are used in at least one application scenario among traffic guidance release, traffic signal control optimization, traffic congestion early warning, or traffic operation management decision support.

[0018] Beneficial effects: Compared with the prior art, the traffic situation assessment method based on traffic flow parameter prediction provided by this invention introduces the predictive analysis of the future changing trend of traffic flow parameters in the traffic situation assessment process. This makes the traffic situation assessment no longer limited to the passive judgment of the current or historical traffic state, but can make a forward-looking assessment of the road traffic operation state within a future preset time window. This provides advance notice for traffic management and control, and significantly improves the timeliness and initiative of traffic situation perception.

[0019] This invention achieves a more comprehensive reflection of road traffic conditions by unifying and integrating various traffic flow parameters such as traffic flow, traffic speed, traffic density, and road occupancy through modeling and analysis. This avoids the limitations of relying solely on single traffic indicators or simple threshold judgments. When traffic conditions are fluctuating or at the point of near-congestion, the traffic situation assessment results obtained by this invention are more stable, helping to reduce misjudgments and frequent changes in situation levels, thus improving the reliability of the assessment results.

[0020] Furthermore, this invention introduces a dynamic update and adaptive adjustment mechanism in the process of traffic flow parameter prediction and traffic situation assessment, enabling the prediction model and assessment model to continuously correct their parameters based on the latest collected traffic flow data. This enhances the system's adaptability to different road types, different traffic operation characteristics, and changes in traffic patterns at different times, thereby improving the overall assessment accuracy and robustness.

[0021] From an engineering application perspective, the method of this invention can directly acquire traffic flow data based on existing traffic detection facilities without requiring large-scale modifications to the hardware system. The algorithm structure is clear, and the calculation process is standardized, facilitating deployment and implementation on traffic management platforms, edge computing nodes, or cloud systems. By applying the evaluation results to scenarios such as traffic guidance, signal control optimization, congestion warning, and traffic operation management decision support, this invention can effectively improve the decision-making efficiency of traffic management departments and the overall level of road traffic operation, demonstrating significant practical value and promising prospects for wider application. Attached Figure Description

[0022] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the following description is provided in conjunction with embodiments and appendices. Figure 1 The present invention will be further described in detail below. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0024] Example 1: In this example, a main road section in a city is used as the object of traffic situation assessment. The road has a two-way, multi-lane structure, and traffic flow detection devices are deployed upstream, midstream, and downstream along the road's travel direction. The traffic flow detection devices include geomagnetic detectors and video detection devices. The geomagnetic detectors are used to collect information on the number of vehicles passing through, while the video detection devices are used to collect information on vehicle speed and road occupancy. Each detection device operates according to a uniform sampling period, which can be set to 30 seconds or 60 seconds. The collected raw traffic flow data is uploaded to the traffic management platform via wired or wireless communication.

[0025] After receiving raw traffic flow data, the traffic management platform first performs time synchronization processing on data from different detection devices to eliminate time deviations caused by device clock errors. Subsequently, it performs quality checks on the collected raw traffic flow data. When abnormal data or data significantly deviating from the normal range is detected, it is automatically marked as abnormal and removed. In cases of short-term data gaps, historical traffic flow data from the same time period is used to compensate for and repair the missing data, thereby obtaining continuous and stable traffic flow time series data. Simultaneously, to reduce the impact of random fluctuations on the analysis results, the processed traffic flow data undergoes smoothing processing.

[0026] After data preprocessing, the traffic management platform constructs a traffic flow parameter feature set based on the preprocessed traffic flow time series data. This feature set includes at least traffic flow parameters, traffic speed parameters, traffic density parameters, and road occupancy parameters, where the traffic density parameter is calculated from the relationship between the traffic flow and traffic speed parameters. By combining these various traffic flow parameters, a multidimensional traffic flow parameter feature vector is formed for subsequent predictive analysis.

[0027] During the traffic flow parameter prediction phase, the traffic management platform establishes a short-term traffic flow parameter prediction model based on historical traffic flow parameter feature vectors. This model predicts the changing trends of traffic flow, traffic speed, traffic density, and road occupancy over the next few minutes. During operation, the prediction model dynamically updates its parameters using the latest collected traffic flow parameters, ensuring that the prediction results promptly reflect changes in road traffic conditions and thus improving the accuracy and adaptability of the predictions.

[0028] After obtaining the predicted traffic flow parameters for the future time window, the traffic management platform further constructs a traffic situation assessment model based on these parameters. The traffic situation assessment model integrates and analyzes various predicted traffic flow parameters, assigns corresponding weights according to the degree of influence of different parameters on traffic operation status, and comprehensively calculates the traffic situation assessment results for the future time window. These weights can be set or dynamically adjusted based on road type, traffic operation characteristics, and historical prediction results to make the traffic situation assessment results more consistent with actual traffic conditions.

[0029] Based on the traffic situation assessment results and preset traffic situation level determination rules, the traffic management platform classifies the traffic operation status of the target road within a future time window. The traffic situation level is divided into at least four categories: smooth traffic, light congestion, moderate congestion, and severe congestion, and the determined traffic situation level is output as the final assessment result.

[0030] In practical applications, the traffic management platform uses the traffic situation assessment results for various traffic management scenarios. When the prediction results indicate that moderate or severe congestion may occur within a future time window, the system automatically sends congestion warning information to the travel guidance system and publishes detour suggestions to the public through variable message signs or navigation services. At the same time, the traffic signal control system can adjust the signal timing of adjacent intersections in advance based on the predicted traffic flow trends. In addition, traffic management personnel can also conduct a comprehensive assessment of road operation based on the traffic situation assessment results, providing decision support for traffic scheduling and emergency management.

[0031] Example 2: In this embodiment, a section of urban main road is selected as the target for traffic situation assessment. Geomagnetic detectors and video detection equipment are deployed at upstream, midstream, and downstream locations of the road section for use in fixed sampling periods. Real-time acquisition of raw traffic flow data. This raw traffic flow data includes at least the number of vehicles passing through the detection section per unit time, vehicle speed information, and detector occupancy time information. Within any sampling period, traffic flow is expressed by the formula:

[0032] The calculation yielded, where Indicates time Traffic flow, expressed in vehicles per second or vehicles per hour; This represents the number of vehicles passing through the detection section within the sampling period Δt; Indicates the sampling period length.

[0033] Average driving speed:

[0034] Calculation, where Indicates time average driving speed Indicates the first The vehicle's speed during the sampling period.

[0035] Road occupancy rate approved:

[0036] Obtain, among which Indicates time Road occupancy rate This indicates the cumulative time the detector is occupied by vehicles within the sampling period.

[0037] This constitutes the original traffic flow data vector at each moment:

[0038] in This indicates transpose.

[0039] Due to time discrepancies, noise, and data gaps inherent in different detection devices, the raw traffic flow data requires preprocessing before subsequent analysis. First, the original timestamps are corrected. pass:

[0040] Obtain, among which This is the original timestamp. This is the time offset estimated by the system.

[0041] Subsequently, anomaly detection was performed on the traffic flow parameters, for any parameter , in length of Calculate the local mean within the sliding time window:

[0042] and standard deviation

[0043] in This represents the mean within the window. This represents the standard deviation within the window.

[0044] when When established, the parameter value is determined to be an outlier. The anomaly detection coefficient is used. Outliers are replaced by linear interpolation between adjacent valid time points, and the replacement value is represented as... .

[0045] When a parameter is detected to be missing at a certain moment, compensation is made using the average value of the parameters at the same historical moment. The compensation value is obtained through:

[0046] Calculation, where Indicates the first The observed values ​​at the corresponding time points of each historical cycle. This represents the number of historical samples.

[0047] To reduce the impact of random fluctuations, the compensated data is subjected to exponential smoothing. The smoothed parameters... Depend on

[0048] Received, among which This is the smoothing coefficient. After the above processing, a continuous and stable traffic flow time series is obtained.

[0049] The meanings of the symbols are the same as those described above.

[0050] Based on this, a traffic flow parameter feature set is further constructed. Traffic density is calculated through basic traffic flow relationships and expressed as...

[0051] in Traffic density, measured in vehicles per meter or vehicles per kilometer, is the basis of this formula.

[0052] This results in a traffic flow parameter feature vector at each moment:

[0053] And construct the most recent The historical input sequence composed of sampling times:

[0054] Used for traffic flow parameter prediction.

[0055] To perform short-term predictions of traffic flow parameters within a preset time window, this implementation method employs a multivariate autoregressive prediction model. The predicted characteristics of traffic flow parameters for the next time moment are expressed as follows:

[0056] in The predicted feature vector, A vector of constant terms. The model order is... For the first The order regression coefficient matrix.

[0057] When making multi-step predictions for the future, recursive calculations are performed.

[0058] in This indicates the prediction step size.

[0059] When newly collected traffic flow parameters arrive, the model parameters are updated online using the recursive least squares method to ensure that the prediction model is dynamically adjusted according to changes in traffic conditions, thereby improving prediction accuracy.

[0060] After obtaining the predicted traffic flow parameters for the future time window, a traffic situation assessment is performed. First, the predicted parameters are normalized. For any given predicted parameter... The normalized form is

[0061] in and These are the minimum and maximum values ​​obtained from historical statistics.

[0062] Since speed and congestion level are inversely correlated, the speed term is subjected to inverse normalization. Subsequently, a weighted fusion method is used to calculate the traffic situation assessment value.

[0063] in The traffic situation assessment value at the predicted time. The weight coefficients satisfy the condition that the sum of the weights is 1.

[0064] To obtain the overall situation over the entire forecast time window, the window-averaged situation value is calculated.

[0065] in To predict the number of steps.

[0066] The weighting coefficients are adaptively adjusted based on the prediction error. First, the weighting coefficients of each parameter are calculated in the most recent... The average absolute error over a given time period

[0067] Then, the weights are determined inversely to the magnitude of the error and normalized so that parameters with higher prediction accuracy have a greater weight in the situation assessment.

[0068] Finally, based on the comprehensive traffic situation assessment value The system matches the target road against a preset threshold range, classifying the traffic status of the target road within a future time window as smooth, lightly congested, moderately congested, or severely congested. The evaluation results are then output to the traffic management platform for applications such as traffic guidance dissemination, signal control optimization, congestion warning, and traffic operation management decision support.

[0069] 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, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A traffic situation assessment method based on traffic flow parameter prediction, characterized in that, Includes the following steps: Collect raw traffic flow data for the target road or road segment, including at least traffic volume data, average driving speed data, and road occupancy data; The raw traffic flow data is preprocessed to obtain continuous traffic flow time series data; A traffic flow parameter feature set is constructed based on the traffic flow time series data; A traffic flow parameter prediction model is established using the traffic flow parameter feature set to predict the traffic flow parameters within a preset time window in the future, thereby obtaining the predicted traffic flow parameters. Based on the predicted traffic flow parameters, a traffic situation assessment model is constructed to assess the traffic operation status of the target road or road segment within the preset time window, and the traffic situation assessment result is obtained. The corresponding traffic situation level is output based on the traffic situation assessment results.

2. The traffic situation assessment method based on traffic flow parameter prediction according to claim 1, characterized in that, The raw traffic flow data is collected using geomagnetic detectors, video detection equipment, radar detection equipment, or a combination thereof, with a fixed collection period.

3. The traffic situation assessment method based on traffic flow parameter prediction according to claim 1, characterized in that, The preprocessing of the raw traffic flow data includes: Time synchronization processing is performed on raw traffic flow data from different data sources; Identify and remove abnormal data; Compensate and repair missing data; The processed data is smoothed and filtered to reduce the impact of random fluctuations on subsequent analysis.

4. The traffic situation assessment method based on traffic flow parameter prediction according to claim 1, characterized in that, The traffic flow parameter feature set includes at least traffic flow parameters, traffic speed parameters, and traffic density parameters, wherein the traffic density parameter is calculated based on the traffic flow parameters and traffic speed parameters.

5. The traffic situation assessment method based on traffic flow parameter prediction according to claim 1, characterized in that, The traffic flow parameter prediction model is a short-term prediction model, used to predict the trend of traffic flow parameter changes in the next few minutes to tens of minutes.

6. The traffic situation assessment method based on traffic flow parameter prediction according to claim 5, characterized in that, The traffic flow parameter prediction model is trained based on historical traffic flow time series data and can dynamically update the prediction results according to the latest collected traffic flow parameters.

7. The traffic situation assessment method based on traffic flow parameter prediction according to claim 1, characterized in that, The traffic situation assessment model calculates the traffic situation assessment value by weighting and fusing the predicted traffic flow parameters, predicted traffic speed parameters, and predicted traffic density parameters.

8. The traffic situation assessment method based on traffic flow parameter prediction according to claim 7, characterized in that, The weights of each traffic flow parameter in the weighted fusion process are set or adaptively adjusted based on road type, traffic operation characteristics, or historical operation data.

9. The traffic situation assessment method based on traffic flow parameter prediction according to claim 1, characterized in that, The traffic situation level includes at least smooth traffic, light congestion, moderate congestion, and severe congestion. The traffic situation level is obtained by matching the traffic situation assessment results with a preset threshold range.

10. The traffic situation assessment method based on traffic flow parameter prediction according to claim 1, characterized in that, The traffic situation assessment results are used in at least one of the following application scenarios: traffic guidance release, traffic signal control optimization, traffic congestion early warning, or traffic operation management decision support.

Images