A method for evaluating urban road traffic congestion alarm based on intersection flow
The intersection flow data collected by traffic sensors evaluates the stage and causes of road congestion, solves the problem of difficulty in assessing road congestion in existing technologies, and realizes the rapid response strategy formulation and management efficiency improvement of traffic management departments.
Patent Information
- Application Number
- CN202211000227.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-08-19
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2042-08-19
AI Technical Summary
It is difficult for existing technology to effectively evaluate the development stage and causes of road congestion, which makes it difficult for traffic management departments to quickly formulate response strategies.
The intersection flow data collected by traffic sensors are used to determine the stage and cause of road congestion through historical traffic statistics and predictive type estimates, including alarm types such as imminent congestion, existing congestion, traffic influx and spontaneous large flow.
Real-time assessment of the road congestion stage and causes is achieved, the error of prediction results is reduced, the management efficiency of the traffic management department is improved, and the response strategies can be formulated quickly.
Smart Images

Figure CN115359660B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of intelligent transportation and relates to an urban road traffic congestion alarm evaluation method based on intersection flow. Background Art
[0002] The traffic structure of cities in my country is becoming increasingly complex, and problems such as poor road network, insufficient facilities, and difficulty in smooth travel are becoming more and more prominent, which puts higher demands on urban traffic management departments. Traffic management departments need a more efficient automated congestion alarm assessment method to estimate the road congestion situation and the stage of congestion development in real time, reduce workload and improve management efficiency. At present, various indicators such as road section speed, road traffic density, and traffic flow are used at home and abroad to evaluate road congestion, but these indicators are difficult to reflect the development stage of congestion and the causes of congestion, and cannot help traffic management departments quickly form congestion solutions and measures. The present invention proposes an urban road traffic congestion alarm assessment method based on intersection flow, which can evaluate the stage of road congestion and the causes of congestion in real time. The congestion stage includes three stages: no congestion, about to be congested, and already congested. The causes of congestion include spontaneous large flow, traffic congestion, etc.
[0003] The data source of the present invention is simple, and only the traffic flow data of the intersection collected by the traffic sensors is needed. Traffic sensors such as bayonet cameras, ground sensing coils, infrared, etc. can be used as data collection devices. The method is highly efficient and can be well integrated into the existing intelligent transportation system. The alarm types such as equipment abnormality, impending congestion, existing congestion, traffic influx, spontaneous large traffic, etc. are intuitive and clear, which is more conducive to the traffic management department to quickly formulate response strategies.
[0004] The traffic volume at an intersection refers to the number of vehicles passing through the exit of a road section within a unit time period. It can be detected by various methods such as camera, inductor, microwave radar, etc. The device is offline when the system cannot accurately detect the vehicles passing through the intersection due to the device itself, communication network failure, power failure, etc. Usually, the traffic volume data at the intersection is zero. The current section will be congested after a short period of time, which is manifested as the section not being congested in the current time period, but being congested after several time periods. The current section is already congested when the current section is congested in the current time period. Traffic influx refers to a large amount of traffic flowing into the current section from the upstream section. Spontaneous high traffic refers to the situation where the traffic volume of the current section increases to the endurance limit of the section and is not caused by the influx of vehicles. Summary of the invention
[0005] The purpose of the present invention is to provide an urban road traffic congestion alarm assessment method based on intersection flow, which uses traffic flow data collected by traffic sensors to estimate the possibility of equipment abnormalities and road section congestion abnormalities. At the same time, it estimates road section congestion abnormalities such as impending congestion alarms, existing congestion alarms, traffic influx, and traffic bursts. Based on the different manifestations of traffic flow abnormalities in statistics, from a long-term perspective, the changes in the quantity of traffic flow in various dimensions obey the normal distribution, and abnormal situations often appear at the end of the normal distribution.
[0006] The present invention specifically comprises the following steps:
[0007] Step 1: Historical traffic statistics:
[0008] For the observed road section r, assume that the traffic volume collected by the equipment in the kth time period on the dth day is Calculate various statistics of historical traffic and estimate the reasonable range of traffic distribution;
[0009] Specifically:
[0010] Step 1.1: Collect the traffic flow data of road section r in the kth time period in the past n days and output Calculate the mean μ of the traffic volume for these n days L and variance σ L .
[0011] Step 1.2: Calculate the maximum historical flow rate of road section r in the past n days.
[0012] Step 1.3: Calculate the traffic volume on the dth day The value range of F(μ L , σ L )=[μ L -mσ L , μ L +mσ L ], where m is the confidence range. For example, when the conventional confidence level is 99%, the value m = 2.58. In this way, the value range of the normal traffic flow of the road section in the current period is obtained.
[0013] Step 1.4: Calculate the flow difference satisfy At the same time, calculate the mean μ of the flow difference over the past n days DL and variance σ DL And the range of the flow difference value on the dth day F(μ DL , σ DL )=[μ DL -mσ DL , μ DL +mσ DL ].
[0014] Step 1.5: Collect the flow of the upstream section r' of section r at time k Calculate the upstream flow ratio satisfy At the same time, calculate the mean μ of the upstream flow ratio over the past n days RL and variance σ RL And the range of upstream flow ratio on day d (μ RL , σ RL )=[μ RL -mσ RL , μ RL +mσ RL ].
[0015] Step 2: Prediction type estimation:
[0016] Determine the prediction type of road section r in the kth time period on the d′th day, which is divided into 6 types: zero flow prediction WZ r , large traffic indicates WM r , the upward derivative flow indicates WDU r , the downward derivative flow indicates WDD r , Upstream and downstream traffic indicates WRU r And the upstream and downstream traffic indicates WRD downward r The prediction type is an estimate of the abnormal situation and is the basis for evaluating the alarm type. Multiple types of predictions may occur on a road section within a period of time.
[0017] Specifically:
[0018] Step 2.1: When When the traffic flow of road section r at time k is zero, the output of zero traffic indicates WZ r .
[0019] Step 2.2: When When the traffic volume of road section r at time k exceeds the historical maximum, the output of large traffic volume indicates WM r .
[0020] Step 2.3: When When the flow rate growth rate range greatly exceeds the credible range of the historical growth rate, the output derivative flow rate indicates that WDU r ;when When the flow rate deceleration range greatly exceeds the credible range of historical deceleration, the output derivative flow rate predicts WDD downward. r .
[0021] Step 2.4: When When the current section flow is much larger than the upstream section exit flow, the output upstream and downstream flow indicates WRU r ;when When the current section flow is much smaller than the upstream section flow, the output upstream and downstream flow indicates WRD r .
[0022] Step 3: Alarm type assessment:
[0023] The alarm type is obtained according to different predicted states and combinations of predicted states. There may be multiple alarms on a road. The alarm types of the present invention are divided into equipment abnormality alarm and road abnormality alarm. The road abnormality alarm is further divided into impending congestion alarm, congestion alarm, traffic influx alarm and spontaneous high traffic alarm. The specific steps are as follows:
[0024] Step 3.1: Evaluate the abnormal alarm of the equipment. When the equipment is offline, the system cannot obtain the traffic flow of the road section through the equipment, and the corresponding road section traffic flow is zero for a long time. For the current road section equipment, the existence of zero traffic for x consecutive cycles during the peak period indicates WZ r , output device abnormal alarm; zero flow for 2*x consecutive cycles during the off-peak period indicates WZ r , and there is no zero flow indication WZ in the downstream section during this period r , output device abnormal alarm. Where x = 30 minutes / period time length.
[0025] Step 3.2: Evaluate the road anomaly warning, specifically:
[0026] Step 3.2.1: Evaluate the impending congestion alarm. The impending congestion alarm is issued when the traffic flow of road section r does not reach the maximum value in the k-1 time period, but reaches the maximum value in the k-th time period. Therefore, if there is no heavy traffic on road section r in the k-1th time period, it indicates WM r , and the presence of large flow in the kth time period indicates that WM r , output an alarm of impending congestion.
[0027] Step 3.2.2: Evaluate the congestion alarm. The congestion alarm is issued when the traffic volume of the road segment r has been very large in the previous {kx|x=1,2,3,...,n} time periods, and the traffic volume in the kth time period is much larger than the traffic volume of the upstream road segment. Therefore, if the road segment r has an impending congestion alarm or an already congested alarm in the previous x periods, and there is an upward flow of upstream and downstream traffic in the kth time period, it indicates WRU r , the output is already congested and an alarm is issued.
[0028] Step 3.2.3: Evaluate the traffic influx alarm. The traffic influx alarm is issued when the traffic volume of section r suddenly increases in the kth time period, and the traffic volume of the upstream section r′ of section r also suddenly increases in the kth time period. Therefore, if there is an upward derivative flow in the kth time period of section r, it indicates that WDU r, and there is also a derivative flow in the upstream section r′ in the kth time period, indicating WDU r′ , output traffic influx alarm.
[0029] Step 3.2.4: Evaluate spontaneous high-flow alarm. Spontaneous high-flow alarm is issued when the traffic volume of section r suddenly increases in the kth time period, the traffic volume of downstream section r″ also suddenly increases in the kth time period, but the traffic volume of upstream section r′ does not change significantly in the kth time period. Therefore, if there is an upward derivative flow in section r in the kth time period, it indicates that WDU r , and the downstream section r″ also has a derivative flow in the kth time period, indicating that WDU r″ , but there is no derivative flow in the upstream section r′ in the kth time period, indicating WDU upward r′ And the derivative flow downward indicates WDD r′ , output spontaneous high flow alarm.
[0030] The present invention is applicable to various road models and can reduce the error of prediction results; it can realize alarm type evaluation only by using traffic sensors to collect intersection flow data; it can evaluate various traffic congestion alarm types in real time and provide feedback to traffic management departments to help change solution strategies in real time. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION
[0032] The present invention is further described below in conjunction with the accompanying drawings, but the protection scope of the present invention is not limited thereto.
[0033] In this embodiment, in the 72nd time period from March to June 2017, a total of 10 roads in a certain area of Hangzhou were collected by using camera to estimate the type of urban road traffic alarms. A day is divided into 144 time periods, each of which is 10 minutes. For example, the road section numbered 5713301101 is analyzed. Figure 1 As shown, a method for evaluating urban road traffic congestion alarm based on intersection flow specifically includes the following steps:
[0034] Step 1. Historical traffic statistics: For the observed road section numbered 5713301101, the vehicle flow collected by the equipment in the 72nd time period on the 45th day was 653 vehicles.
[0035] Step 1.1. Collect the historical traffic flow data of road section 5713301101 in the 72nd time period. The data comes from 10 days, namely {197, 309, 278, 321, 428, 547, 493, 159, 166, 214}. The mean of the traffic flow in these 10 days is 311.2, and the variance is 130.47.
[0036] Step 1.2: Calculate the maximum historical flow rate of road section 5713301101 in the 72nd time period, which is 547 vehicles.
[0037] Step 1.3: Calculate the value range of the traffic flow on the 45th day [0, 647.8], where when the conventional confidence level is 99%, the value m=2.58.
[0038] Step 1.4: Collect the traffic flow data of 497 vehicles in the 71st time period of the 45th day of section 5713301101, and calculate the traffic flow difference between the 72nd and 71st time periods to be 156 vehicles; at the same time, calculate the traffic flow difference of the past 10 days to be {4, 11, 15, 23, 46, -12, -15, 19, 7, -1}, with a mean of 9.7 and a variance of 16.9. The range of historical traffic flow difference is [-33.9, 53.3].
[0039] Step 1.5, collect the vehicle volume of the upstream section 5713301102 of section 5713301101 in the 72nd time period of the 45th day, which is 331 vehicles. The upstream flow ratio is 1.97. At the same time, the mean upstream flow ratio of the historical 10 days is calculated to be {1.13, 1.56, 0.9, 1.15, 1.33, 1.28, 0.9, 0.32, 1.04, 1.29}, so the mean is 1.09, the variance is 0.32, and the historical upstream flow ratio range is [0.26, 1.92].
[0040] Step 2: Prediction type estimation, specifically:
[0041] Step 2.1: It is found that the traffic volume of road section 5713301101 in the 72nd time period of the 45th day is 653 vehicles, which is greater than the historical maximum value of 547 vehicles, so a high traffic volume is predicted.
[0042] Step 2.2: It is found that the traffic volume in the 72nd time period of the 45th day of section 5713301101 is 156 more than that in the 71st time period, which exceeds the historical range of 53 vehicles. Therefore, there is an upward prediction of the derivative flow.
[0043] Step 2.3, it is found that the upstream flow ratio of section 5713301101 and upstream section 5713301102 in the 72nd time period of the 45th day is 1.97>1.92, so there is an upward prediction of upstream and downstream flows.
[0044] 3. Alarm type evaluation: According to the traffic flow data of section 5713301101 in the 71st time period on the 54th day, it was found that the traffic flow did not reach the maximum value in the 72nd time period, but exceeded the maximum value. Therefore, the section outputs the "road section is about to be congested" alarm; according to calculations, it was found that the upstream section of the section did not have any derivative flow upward prediction or derivative flow downward prediction, while the downstream section also had a derivative flow upward prediction, so the "spontaneous high traffic" alarm was output.
[0045] In addition, this project also evaluates and counts the alarm types of the 10 road sections where data was collected, compares the calculated results with the actual data, and calculates the detection rate and error rate of the design method of this project. The detection rate calculation method is (detection times - error times) / actual times, and the error rate is error times / detection times. The calculation results are as follows:
[0046] After the project team members conducted field observations on 10 roads, in the 72nd time period from March to June 2017, there were 5 real equipment anomalies and 49 real congestions on 10 roads in a certain area of Hangzhou City, of which 21 were about to be congested, 13 were already congested, 14 were due to traffic influx, and 9 were spontaneous large traffic. According to the design method of this project, the number of equipment anomaly alarms detected was 5, of which 1 was a false detection; the road section was already congested 12 times, of which 1 was a false detection; the road section had traffic influx 12 times, of which 2 were false detections. Therefore, the equipment anomaly detection rate of the design method of this project is 80%, and the error rate is 20%; the detection rate of the road section that is already congested is 84.6%, and the error rate is 8.3%; the detection rate of the road section with traffic influx is 71.4%, and the error rate is 16.7%.
[0047] The above is a specific implementation case of the present invention. If the changes made due to the concept of the present invention do not exceed the spirit covered by the specification, they should still fall within the protection scope of the present invention.
Claims
1. A method for evaluating urban road traffic congestion alarm based on intersection flow, Features: The specific steps include: Step 1: Historical traffic statistics: For the observed road section r, assume that the traffic volume collected by the equipment in the kth time period on the dth day is Calculate various statistics of historical traffic and estimate the reasonable range of traffic distribution; Specifically: Step 1.1: Collect the traffic flow data of road section r in the kth time period in the past n days and output Calculate the mean traffic volume μ for these n days L and variance σ L ; Step 1.2: Calculate the maximum historical flow rate of road section r in the past n days. Step 1.3: Calculate the traffic volume on the dth day The value range of F(μ L ,σ L )=[μ L -mσ L ,μ L +mσ L ], where m is the confidence range; the value range of the normal traffic flow on the road section in the current period is obtained; Step 1.4: Calculate the flow difference satisfy At the same time, calculate the mean μ of the flow difference over the past n days DL and variance σ DL And the range of the flow difference value on the dth day F(μ DL ,σ DL )=[μ DL -mσ DL ,μ DL +mσ DL ]; Step 1.5: Collect the flow of the upstream section r' of section r at time k Calculate the upstream flow ratio satisfy At the same time, calculate the mean μ of the upstream flow ratio over the past n days RL and variance σ RL And the range of upstream flow ratio on day d (μ RL ,σ RL )=[μ RL -mσ RL ,μ RL +mσ RL ]; Step 2: Prediction type estimation: Determine the prediction type of road section r in the kth time period on the d′th day, which is divided into 6 types: zero flow prediction WZ r , large traffic indicates WM r , the upward derivative flow indicates WDU r , the downward derivative flow indicates WDD r , Upstream and downstream traffic indicates WRU r And the upstream and downstream traffic indicates WRD downward r ; The prediction type is an estimate of the abnormal situation and is the basis for the evaluation of the alarm type; multiple types of predictions may occur on a road section within a period of time; Specifically: Step 2.1: When When the traffic flow of road section r at time k is zero, the output of zero traffic indicates WZ r ; Step 2.2: When When the traffic volume of road section r at time k exceeds the historical maximum, the output of large traffic volume indicates WM r ; Step 2.3: When When the flow rate growth rate range greatly exceeds the credible range of the historical growth rate, the output derivative flow rate indicates that WDU r ;when When the flow rate deceleration range greatly exceeds the credible range of historical deceleration, the output derivative flow rate predicts WDD downward. r ; Step 2.4: When When the current section flow is much larger than the upstream section exit flow, the output upstream and downstream flow indicates WRU r ;when When the current section flow is much smaller than the upstream section flow, the output upstream and downstream flow indicates WRD r ; Step 3: Alarm type assessment: obtain the alarm type according to different predicted states and combinations of predicted states; a road may have multiple alarms; the alarm types are divided into equipment abnormality alarm and road abnormality alarm, and the road abnormality alarm is further divided into impending congestion alarm, congestion alarm, traffic influx alarm and spontaneous high traffic alarm; The specific steps are as follows: Step 3.1: Evaluate the abnormal alarm of the equipment; when the equipment is offline, the system cannot obtain the traffic flow of the road section through the equipment, and the corresponding traffic flow of the road section is zero for a long time; for the current road section equipment, the existence of zero traffic for x consecutive cycles during the peak period indicates WZ r , output device abnormal alarm; zero flow for 2*x consecutive cycles during the off-peak period indicates WZ r , and there is no zero flow indication WZ in the downstream section during this period r , output device abnormal alarm; where x = 30 minutes / period time length; Step 3.2: Evaluate the road anomaly warning, specifically: Step 3.2.1: Evaluate the impending congestion alarm; the impending congestion alarm is issued when the traffic flow of road section r does not reach the maximum value in the k-1th time period, but reaches the maximum value in the kth time period; therefore, if there is no heavy traffic on road section r in the k-1th time period, it indicates WM r , and the presence of large flow in the kth time period indicates that WM r , output the impending congestion alarm; Step 3.2.2: Evaluate the congestion alarm; the congestion alarm is issued when the traffic volume of the road section r has been very large in the previous {kx|x=1,2,3,…,n} time periods, and the traffic volume in the kth time period is much larger than the traffic volume of the upstream road section; therefore, if the road section r has an impending congestion alarm or an existing congestion alarm in the previous x periods, and there is an upward flow of upstream and downstream traffic in the kth time period, it indicates WRU r , the output is already congested alarm; Step 3.2.3: Evaluate the traffic influx alarm; the traffic influx alarm is issued when the traffic volume of section r suddenly increases in the kth time period, and the traffic volume of the upstream section r′ of section r also suddenly increases in the kth time period; therefore, if there is an upward derivative flow in the kth time period of section r, it indicates that WDU r , and there is also a derivative flow in the upstream section r′ in the kth time period, indicating WDU r' , output traffic influx alarm; Step 3.2.4: Evaluate spontaneous high-flow alarms; Spontaneous high-flow alarms are issued when the traffic volume of section r suddenly increases in the kth time period, the traffic volume of the downstream section r″ also suddenly increases in the kth time period, but the traffic volume of the upstream section r′ does not change significantly in the kth time period; therefore, if there is an upward derivative flow in the kth time period of section r, it indicates that WDU r , and the downstream section r″ also has a derivative flow in the kth time period, indicating that WDU r'' , but there is no derivative flow in the upstream section r′ in the kth time period, indicating WDU upward r' And the derivative flow downward indicates WDD r' , output spontaneous high flow alarm.
