A traffic congestion prediction method based on GPS data

Abstract

The application discloses a traffic congestion prediction method based on GPS data in the technical field of traffic prediction, and comprises the following steps: S1, establishing a basic reference model; S2, importing historical congestion information, marking a pre-congestion area on the basic reference model based on the historical congestion information; S3, correcting the pre-congestion area, updating the pre-congestion area into a congestion area, a smooth area, a first slow-down area and a second slow-down area based on the distribution of the pre-congestion area on the basic reference model and the traffic conditions of the downstream place block corresponding to the place block after the interval time period, and obtaining a prediction model of the interval time period; and S4, uploading vehicle GPS data to the prediction model, obtaining the average driving speed corresponding to the place block based on the place block where the vehicle GPS data is located in the prediction model; the prediction model is more accurate through the correction of the pre-congestion area, the traffic road congestion condition can be predicted through the GPS data of the current time, and the estimated arrival time of the vehicle is obtained.

Description

Technical Field

[0001] This invention belongs to the field of traffic prediction technology, specifically a traffic congestion prediction method based on GPS data. Background Technology

[0002] Traffic congestion refers to a situation where, within a certain period of time, due to increased traffic demand, the total traffic volume passing through a certain road segment or intersection exceeds the road's traffic capacity (the passage capacity of the road segment or intersection), resulting in the traffic flow being unable to proceed smoothly and the excess traffic flow remaining on the road (road segment or intersection).

[0003] Identifying and predicting traffic conditions has become an important means of alleviating traffic congestion. For example, Chinese Patent Publication No. CN104778834B discloses a method for judging urban road traffic congestion based on vehicle GPS data. This method addresses the problem that existing urban road traffic congestion judgment methods, which rely on traditional traffic information detection equipment, have limited application scope. The method constructs an urban road segment travel time prediction model based on an artificial neural network model. Using this model, the travel time data for the current road segment is calculated based on the vehicle's GPS location vector, road segment number vector, timestamp vector, and speed vector. Based on this travel time data, the road segment traffic flow speed and density are further calculated. Using the traffic flow speed and density data as input, the road traffic congestion status is determined. The availability of current GPS data allows for rapid and accurate judgment of traffic congestion status.

[0004] However, judging traffic congestion status based on GPS data at the current moment is inaccurate because traffic congestion is related to overall traffic and ignores the impact of traffic conditions outside the current road segment on the current road segment. This makes it difficult for drivers or passengers to estimate the time to reach their destination. Summary of the Invention

[0005] To address the aforementioned problem of drivers or passengers finding it inconvenient to estimate arrival times, the present invention aims to provide a traffic congestion prediction method based on GPS data. This method uses current GPS data to predict traffic congestion, making it easier to anticipate the arrival time of vehicles at their destinations.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows:

[0007] A traffic congestion prediction method based on GPS data includes the following steps: S1, establishing a basic reference model, using the current urban road map as the basic reference model, and importing historical vehicle driving data from different time periods into the corresponding basic reference model;

[0008] S2. Import historical congestion information. Within a certain time interval, divide the basic reference model into several location blocks. Based on the historical driving data of vehicles corresponding to the location block, calculate the vehicle traffic volume and driving time of several vehicles corresponding to the location block. Based on the driving time of several vehicles, obtain the average driving speed corresponding to the location block.

[0009] Then, the historical congestion information is associated with the corresponding location blocks in the basic reference model. The historical congestion information includes congestion and free flow. The average driving speed corresponding to the congestion of different location blocks is obtained, and the location block corresponding to the congestion is marked as a pre-congestion area in the basic reference model.

[0010] S3. Pre-congestion zone correction: Based on all the basic reference models during this time interval, obtain the probability A of the location block being marked as a pre-congestion zone in all the basic reference models. If the probability A is greater than 60%, then update the pre-congestion zone to a congestion zone.

[0011] If probability A is less than 60%, after the interval, the traffic conditions of the downstream location block corresponding to the location block are obtained. The traffic conditions of the downstream location block are obtained based on the historical congestion information corresponding to the same location block in all basic reference models. If the proportion of historical congestion information that is congested in the basic value is greater than the proportion of historical congestion information that is unobstructed in the basic value, then the traffic conditions of the downstream location block are congested. If the proportion of historical congestion information that is congested in the basic value is less than the proportion of historical congestion information that is unobstructed in the basic value, then the traffic conditions of the downstream location block are unobstructed.

[0012] If the traffic conditions in the downstream location block are smooth, the pre-congestion area will be updated to a smooth area;

[0013] If the traffic condition of a downstream location block is congested, repeat the above process of obtaining the traffic condition of the downstream location block. During this time interval, obtain the traffic conditions of all upstream location blocks associated with the downstream location block. If the probability that the traffic condition of an upstream location block is congested in the total upstream location blocks is greater than 50%, then update the pre-congestion area to the first deceleration and slow-moving area. If the probability that the traffic condition of an upstream location block is congested in the total upstream location blocks is less than 50%, then update the pre-congestion area to the second deceleration and slow-moving area.

[0014] Based on the distribution of congested areas, unobstructed areas, the first deceleration zone, and the second deceleration zone in the basic reference model, a prediction model for this time interval is obtained.

[0015] S4. Upload the vehicle GPS data to the prediction model. Based on the vehicle GPS data at the current time, obtain the prediction model for the corresponding time interval. Based on the location block where the vehicle GPS data is located in the prediction model, obtain the average driving speed corresponding to that location block. Based on the average driving speed corresponding to each location block, obtain the estimated time.

[0016] The above scheme achieves the following beneficial effects: by importing vehicle historical driving data from different time periods into the corresponding basic reference model, it is easy to link the vehicle historical driving data with the basic reference model, thereby obtaining the distribution of the pre-congestion area on the basic reference model.

[0017] By calculating the probability that a pre-congestion zone is a pre-congestion zone in all base reference models within a certain time interval, it is easier to confirm whether the pre-congestion zone will become a congestion zone.

[0018] When the probability A of a location block being marked as a pre-congestion area in all base reference models is less than 60%, the traffic conditions of the downstream location blocks corresponding to that location block are obtained after the specified time interval. By obtaining the traffic conditions of the downstream location blocks and all upstream location blocks associated with them, it is easier to determine whether the pre-congestion area is a smooth flow area or a slow-moving area. This allows the resulting prediction model to link the location blocks at the current time interval with the downstream location blocks at the next time interval, thereby improving the accuracy of traffic congestion prediction.

[0019] By obtaining the average driving speed of each location block based on the current GPS data, the estimated time can be obtained. At the same time, it helps drivers to understand and predict traffic congestion in advance, and to predict when the vehicle will reach its destination, reducing the uncertainty of waiting for traffic to resume.

[0020] Furthermore, in S4, the geographical location information corresponding to the current location of the vehicle will be obtained. The geographical location information includes intersections, construction sections, and speed-limited sections. The estimated time will be extended and adjusted based on the geographical location information.

[0021] Beneficial effects: By identifying geographical location information, more room is provided for predicting congestion time, making the predicted time more accurate and facilitating the judgment of traffic congestion.

[0022] Furthermore, in S1, the basic reference model includes a special prediction model and a regular prediction model. The special prediction model is established based on vehicle driving data corresponding to different special time periods in a year, and the time periods in a year other than the special time periods are regular time periods. The regular prediction model is established based on vehicle driving data corresponding to different time periods in a day within the regular time period.

[0023] Beneficial effects: By distinguishing between special time periods and regular time periods and establishing different basic reference models, the system can select between special and regular prediction models based on different time periods throughout the year, thereby improving the accuracy of traffic congestion prediction.

[0024] Furthermore, in S3, the establishment of special prediction models only uses the historical driving data and historical congestion information of vehicles corresponding to specific time periods as references, while the establishment of regular prediction models only uses the historical driving data and historical congestion information of vehicles corresponding to regular time periods as references.

[0025] Beneficial effects: Different base reference models are established using corresponding historical vehicle data, which makes special or conventional prediction models more closely match actual traffic conditions, thus improving the accuracy of traffic congestion prediction.

[0026] Furthermore, in S3, the average driving speed in the first deceleration zone is less than the average driving speed in the second deceleration zone.

[0027] Beneficial effects: Average driving speed reflects traffic congestion. By setting up first or second slow-down zones with different average driving speeds, the accuracy of subsequent traffic congestion detection can be improved.

[0028] Furthermore, in S4, when the vehicle's GPS data is located in the first deceleration zone or the second deceleration zone, the average driving speed corresponding to the first deceleration zone or the second deceleration zone is displayed.

[0029] Beneficial effects: It allows drivers to adjust vehicle speed in a timely manner, reducing the occurrence of phantom traffic jams and thus reducing the likelihood of traffic congestion.

[0030] Furthermore, the special time period is based on holidays as reference time.

[0031] Beneficial effects: Traffic conditions during holidays differ from those during regular times. Using the baseline reference model during holidays as a special prediction model makes it easier to distinguish between holidays and regular times, thus improving the accuracy of traffic congestion prediction.

[0032] Furthermore, the regular time cycle includes the normal time cycle and the weekend time cycle, wherein the time periods of the regular time cycle include daily time periods, working time periods, off-get off work time periods, and special time periods.

[0033] Beneficial effects: Traffic conditions during normal time periods differ from those during weekend time periods. Traffic conditions during regular time periods exhibit distinct time-segmentation characteristics. By establishing predictive models for different time periods, the accuracy of traffic congestion prediction can be improved. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the traffic congestion prediction method according to an embodiment of the present invention. Detailed Implementation

[0035] The following detailed description illustrates the specific implementation method:

[0036] The basic implementation examples are as follows: Figure 1 As shown: A traffic congestion prediction method based on GPS data includes the following steps: S1, establishing a basic reference model, using the current urban road map as the basic reference model, and importing historical vehicle driving data from different time periods into the corresponding basic reference model;

[0037] The basic reference model includes a special prediction model and a regular prediction model. The special prediction model is established based on vehicle driving data corresponding to different special time periods in a year, and the time periods other than the special time periods in a year are the regular time periods. The regular prediction model is established based on vehicle driving data corresponding to different time periods in a day within the regular time period.

[0038] The special time period uses holidays as a reference time; the regular time period includes normal time period and weekend time period, wherein the time period of the regular time period includes daily time period, working time period, off-get off work time period and special time period.

[0039] For example, in a typical city, during holidays such as the Mid-Autumn Festival, traffic flow shows a significant upward trend before or on the day before the holiday, making traffic congestion more likely. However, the day on a holiday differs from regular time. During regular time, traffic flow shows a significant upward trend during rush hour, making traffic congestion more likely, while outside of rush hour, traffic flow shows a downward trend. Therefore, there are certain differences between special prediction models and conventional prediction models. By distinguishing between the two, it is easier to improve the accuracy of subsequent traffic congestion predictions.

[0040] S2. Import historical congestion information. Within a certain time interval, divide the basic reference model into several location blocks. Based on the historical driving data of vehicles corresponding to the location block, calculate the vehicle traffic volume and driving time of several vehicles corresponding to the location block. Based on the driving time of several vehicles, obtain the average driving speed corresponding to the location block.

[0041] Then, the historical congestion information is associated with the corresponding location blocks in the basic reference model. The historical congestion information includes congestion and free flow. The average driving speed corresponding to the congestion of different location blocks is obtained, and the location block corresponding to the congestion is marked as a pre-congestion area in the basic reference model.

[0042] For example, in a city's roads, different locations may have different road conditions. One location may experience ghost traffic jams due to delays caused by multiple vehicles caused by surrounding connecting roads, even though the roads themselves may be clear. By comparing these phenomena, the accuracy of subsequent traffic congestion predictions can be improved.

[0043] S3. Pre-congestion zone correction: Based on all basic reference models within this time interval, the establishment of special prediction models only uses the historical driving data and historical congestion information of vehicles corresponding to the special time period as reference, while the establishment of regular prediction models only uses the historical driving data and historical congestion information of vehicles corresponding to the regular time period as reference.

[0044] Obtain the probability A of marking the location block as a pre-congestion area in all base reference models. If the probability A is greater than 60%, update the pre-congestion area to a congestion area.

[0045] If probability A is less than 60%, after the interval, the traffic conditions of the downstream location block corresponding to the location block are obtained. The traffic conditions of the downstream location block are obtained based on the historical congestion information corresponding to the same location block in all basic reference models. If the proportion of historical congestion information that is congested in the basic value is greater than the proportion of historical congestion information that is unobstructed in the basic value, then the traffic conditions of the downstream location block are congested. If the proportion of historical congestion information that is congested in the basic value is less than the proportion of historical congestion information that is unobstructed in the basic value, then the traffic conditions of the downstream location block are unobstructed.

[0046] If the traffic conditions in the downstream location block are smooth, the pre-congestion area will be updated to a smooth area;

[0047] For example, a certain road in a city may experience a temporary slowdown in vehicle speed due to factors such as merging or speed limits, but after vehicles pass through this section of road, the traffic flow in the upstream areas surrounding the downstream locations is low, and vehicle traffic flows smoothly.

[0048] If the traffic condition of a downstream location block is congested, repeat the above process of obtaining the traffic condition of the downstream location block. During this time interval, obtain the traffic conditions of all upstream location blocks associated with the downstream location block. If the probability that the traffic condition of an upstream location block is congested in the total upstream location blocks is greater than 50%, then update the pre-congestion area to the first deceleration and slow-moving area. If the probability that the traffic condition of an upstream location block is congested in the total upstream location blocks is less than 50%, then update the pre-congestion area to the second deceleration and slow-moving area.

[0049] For example, a road in a city may experience a temporary slowdown in vehicle speed due to merging or speed limits, depending on the volume of traffic. However, after passing through this section of road, the upstream and downstream locations of the downstream location have high traffic volume. Due to the convergence of traffic volume, vehicles are already congested when they reach the downstream location.

[0050] The average driving speed in the first deceleration zone is less than the average driving speed in the second deceleration zone.

[0051] Based on the distribution of congested areas, unobstructed areas, the first deceleration zone, and the second deceleration zone in the basic reference model, a prediction model for this time interval is obtained.

[0052] S4. Upload the vehicle GPS data to the prediction model. Based on the vehicle GPS data at the current time, obtain the prediction model for the corresponding time interval. Based on the location block where the vehicle GPS data is located in the prediction model, obtain the average driving speed corresponding to that location block. Based on the average driving speed corresponding to each location block, obtain the estimated time.

[0053] The prediction model obtained through S4 links the location blocks at the current time interval with the downstream location blocks at the next time interval, thereby improving the accuracy of traffic congestion prediction. It is convenient to predict traffic congestion based on GPS data at the current moment, thereby predicting the time for vehicles to reach their destination, reducing the uncertainty of waiting for traffic to resume, alleviating drivers' driving tension, and making it easier for drivers to make rational judgments, thus reducing the occurrence of phantom traffic jams.

[0054] Example 2

[0055] The difference from the above embodiments is that in S4, the geographical location information corresponding to the current location block of the vehicle is also obtained. The geographical location information includes intersections, construction sections, and speed-limited sections. The estimated time is extended and adjusted based on the geographical location information. The extension adjustment time is determined by the road conditions of the corresponding location block.

[0056] The specific implementation process is as follows: Location blocks based on geographic location information have complex road traffic or road occupancy situations, which are prone to traffic accidents or congestion and have uncertainties. By identifying geographic location information, more room is reserved for predicting congestion time, making the predicted time more accurate and facilitating the judgment of traffic congestion.

[0057] Example 3

[0058] The difference from the above embodiments is that, in S4, when the vehicle GPS data is located in the first deceleration zone or the second deceleration zone, the average driving speed corresponding to the first deceleration zone or the second deceleration zone will also be displayed to the vehicle driver or passengers.

[0059] The specific implementation process is as follows: By reminding drivers of the average driving speed, the time for vehicles to restart after stopping is reduced, making it easier for drivers to adjust the vehicle speed in a timely manner, reducing the occurrence of ghost traffic jams, and thus reducing the possibility of traffic congestion.

[0060] The above descriptions are merely embodiments of the present invention, and common knowledge such as specific structures and / or characteristics in the solutions are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the structure of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A traffic congestion prediction method based on GPS data, characterized in that, The steps include: S1. Establishing a basic reference model: using the current city road map as the basic reference model, importing historical vehicle driving data from different time periods into the corresponding basic reference model; S2. Import historical congestion information. Within a certain time interval, divide the basic reference model into several location blocks. Based on the historical driving data of vehicles corresponding to the location block, calculate the vehicle traffic volume and the driving time of several vehicles corresponding to the location block. Based on the vehicle traffic volume and the driving time of several vehicles, obtain the average driving speed corresponding to the location block. Then, the historical congestion information is associated with the corresponding location blocks in the basic reference model. The historical congestion information includes congestion and free flow. The average driving speed corresponding to the congestion of different location blocks is obtained, and the location block corresponding to the congestion is marked as a pre-congestion area in the basic reference model. S3. Pre-congestion zone correction: Based on all the basic reference models during this time interval, obtain the probability A of the location block being marked as a pre-congestion zone in all the basic reference models. If the probability A is greater than 60%, then update the pre-congestion zone to a congestion zone. If probability A is less than 60%, after the interval, the traffic conditions of the downstream location block corresponding to the location block are obtained. The traffic conditions of the downstream location block are obtained based on the historical congestion information corresponding to the same location block in all basic reference models. If the proportion of historical congestion information that is congested in the basic value is greater than the proportion of historical congestion information that is unobstructed in the basic value, then the traffic conditions of the downstream location block are congested. If the proportion of historical congestion information that is congested in the basic value is less than the proportion of historical congestion information that is unobstructed in the basic value, then the traffic conditions of the downstream location block are unobstructed. If the traffic conditions in the downstream location block are smooth, the pre-congestion area will be updated to a smooth area; If the traffic condition of a downstream location block is congested, repeat the above process of obtaining the traffic condition of the downstream location block. During this time interval, obtain the traffic conditions of all upstream location blocks associated with the downstream location block. If the probability that the traffic condition of an upstream location block is congested in the total upstream location blocks is greater than 50%, then update the pre-congestion area to the first deceleration and slow-moving area. If the probability that the traffic condition of an upstream location block is congested in the total upstream location blocks is less than 50%, then update the pre-congestion area to the second deceleration and slow-moving area. Based on the distribution of congested areas, unobstructed areas, the first deceleration zone, and the second deceleration zone in the basic reference model, a prediction model for this time interval is obtained. S4. Upload the vehicle GPS data to the prediction model. Based on the vehicle GPS data at the current time, obtain the prediction model for the corresponding time interval. Based on the location block where the vehicle GPS data is located in the prediction model, obtain the average driving speed corresponding to that location block. Based on the average driving speed corresponding to each location block, obtain the estimated time.

2. The traffic congestion prediction method based on GPS data according to claim 1, characterized in that: In S4, the geographical location information corresponding to the current location of the vehicle will also be obtained. The geographical location information includes intersections, construction sections, and speed-limited sections. The estimated time will be extended and adjusted based on the geographical location information.

3. The traffic congestion prediction method based on GPS data according to claim 1, characterized in that: In S1, the basic reference model includes a special prediction model and a regular prediction model. The special prediction model is established based on vehicle driving data corresponding to different special time periods in a year, and the time periods other than the special time periods in a year are regular time periods. The regular prediction model is established based on vehicle driving data corresponding to different time periods in a day within the regular time period.

4. The traffic congestion prediction method based on GPS data according to any one of claims 1 or 3, characterized in that: In S3, the establishment of special prediction models only uses the historical driving data and historical congestion information of vehicles corresponding to specific time periods as references, while the establishment of regular prediction models only uses the historical driving data and historical congestion information of vehicles corresponding to regular time periods as references.

5. The traffic congestion prediction method based on GPS data according to claim 1, characterized in that: In S3, the average driving speed in the first deceleration zone is less than the average driving speed in the second deceleration zone.

6. The traffic congestion prediction method based on GPS data according to claim 1, characterized in that: In S4, when the vehicle's GPS data is located in the first or second deceleration zone, the average driving speed corresponding to the first or second deceleration zone is displayed.

7. The traffic congestion prediction method based on GPS data according to claim 3, characterized in that: The special time period is based on holidays as reference time.

8. The traffic congestion prediction method based on GPS data according to claim 3, characterized in that: The regular time cycle includes the normal time cycle and the weekend time cycle. The time periods of the regular time cycle include weekday time periods, working hours, off-get off work hours, and special time periods.

Images