A highway abnormal traffic assessment method and system based on early warning equipment counting

Abstract

This invention discloses a method and system for assessing abnormal highway traffic based on early warning device counts. First, the early warning device generates a time series of highway cross-section traffic and vehicle operating status based on vehicle driving conditions. This information is then sent to a central server for processing. Second, the average flow rate of the current road section is calculated based on the vehicle operating status, thereby calculating the traffic volume of the current road section. Based on the traffic volume, the highway's Level of Service (ε) rating is calculated to determine the degree of road congestion, i.e., different traffic flow states. Finally, traffic anomalies in adjacent sections of the early warning device on the highway are separately inspected and determined for each traffic flow state, and corresponding warnings are processed for each anomaly. This invention fully considers the importance of traffic flow to vehicle status verification, improves the accuracy of highway traffic anomaly assessment, and ensures highway traffic safety.

Description

Technical Field

[0001] The present invention belongs to the technical field of road traffic, and in particular relates to a method and system for evaluating abnormal highway traffic based on early warning device counting. Background Art

[0002] With the continuous increase in the number of cars on the road, the incidence of traffic accidents on highways has also increased year by year. The high speeds of cars on highways create significant momentum and impact, so once a traffic accident occurs, it often causes significant harm. Furthermore, some drivers, due to inexperience or fatigue, are prone to secondary accidents, further exacerbating the consequences and causing serious losses. Road traffic safety has become a hot topic in society, attracting increasing attention and research from all sectors of society.

[0003] While the number of vehicles and total highway mileage are rapidly increasing, the development of Intelligent Traffic Systems (ITS) has lagged behind. There is a lack of intelligent traffic management systems to effectively detect and prevent traffic accidents, as well as a lack of corresponding traffic accident warning mechanisms. When traffic accidents do occur, accident handling and rescue response are also slow.

[0004] This invention utilizes data analysis and mining to collect and analyze vehicle travel data, uncovering unusual patterns and behavioral patterns. By analyzing metrics such as average vehicle speed, travel distance, and dwell time, it can determine whether abnormal conditions such as traffic congestion or accidents exist. This method, in conjunction with related background technologies, effectively determines whether vehicle travel on highways is abnormal, providing real-time traffic monitoring and management, which is crucial for ensuring safe and smooth road traffic. Therefore, based on the principle of automatic blocking in railway transport, this method aims to divide highways into different zones, obtain vehicle operating status and location information, and derive the traffic status of each zone through counting and analysis of early warning equipment. Summary of the Invention

[0005] Purpose of the invention: The present invention provides a method and system for evaluating abnormal traffic on highways based on counting of early warning devices. According to the principle of automatic blocking of railway transportation, the highway is divided into different blocks. The abnormal status of vehicles in the area is judged according to the operating status and position information of vehicles on the highway, and a warning device is used to warn the following vehicles to slow down or change lanes in advance to reduce the occurrence rate of secondary accidents.

[0006] Technical solution: The invention provides a method for evaluating abnormal highway traffic based on early warning device counting, comprising the following steps:

[0007] (1) Data collection: The warning device generates a time series of highway cross-section flow and vehicle operation status based on vehicle driving conditions, and sends the information to the central server for processing;

[0008] (2) Calculate the average flow rate of the current road section based on the vehicle operation status, and then calculate the traffic volume of the current road section;

[0009] (3) Based on traffic volume, the degree of road congestion is determined by calculating the highway service level ε level, i.e. different traffic flow states;

[0010] (4) Inspect and determine traffic anomalies in adjacent warning equipment sections of the expressway according to different traffic flow conditions, and perform corresponding warning processing work according to different abnormal conditions.

[0011] Furthermore, the early warning device includes an information processor, an ultrasonic sensor, a memory, a solar power supply device, a warning device, a receiving device and a transmitting device; the information processor collects the time flow series of the highway section and the vehicle operation status data through the ultrasonic sensor, and saves the information in the memory; the transmitting device transmits the location information and vehicle information in the memory to the central server for information processing. If an abnormal condition occurs in the vehicle, the central server transmits the warning information to the rear early warning device. After receiving the information, the receiving device transmits the information to the early warning device information processor. The early warning device information processor will turn on the warning device to warn the rear vehicle. If the vehicle returns to normal, the warning device will be turned off.

[0012] Furthermore, the implementation process of step (2) is as follows:

[0013] The data is normalized, and the maximum capacity and actual traffic volume of the three lanes are unified to the standard of the leftmost lane. The headway is the difference between the time it takes for two vehicles in the same lane to pass through the projection section of the warning device, that is, the time difference between the two vehicles in the same lane passing through the section. The average of the flow rates obtained by the two warning devices is taken as the average flow rate of the current section, and the traffic volume of the section can be calculated:

[0014]

[0015]

[0016]

[0017] in, Representative The average flow rate recorded by the early warning device; F r 、F r+1represent the average flow rates recorded by the rth warning device and the r+1th warning device respectively; F is the average flow rate between the two warning devices; F r1 、F r2 、F r3 represent the average flow rates of the three lanes recorded by the rth warning device; F (r+1)1 、F (r+1)2 、F (r+1)3 represents the average flow rate of the three lanes recorded by the r+1th warning device; L is the distance between the two warning devices; u is the average speed between the two warning devices preliminarily estimated based on previous data; K n is the reduction factor between lanes; is the average headway, i.e. The average value of the time difference of vehicles passing the warning device within the time period; Q is the traffic volume in this time interval.

[0018] Furthermore, the implementation process of step (3) is as follows:

[0019] The capacity of the leftmost road is C1, the capacity of the second road and the third road is C2 and C3 respectively; then according to the reduction coefficient K between the lanes n , normalizing the standards of all lanes to the leftmost road, then the maximum traffic capacity of the interval is:

[0020]

[0021] The degree of road congestion can be determined by calculating the highway service level ε:

[0022]

[0023] Furthermore, the different traffic flow states in step (3) include a free flow state, a stable flow state, and a congested flow state;

[0024] When ε≤0.35, the vehicle congestion between adjacent warning device nodes is in a free flow state;

[0025] When 0.35<ε≤0.9, the vehicle congestion between adjacent warning device nodes is in a stable flow state;

[0026] When u>0.9, the vehicle congestion level between adjacent warning device nodes is in a congested flow state.

[0027] Furthermore, the implementation process of step (4) is as follows:

[0028] S1: Free-flow state vehicle anomaly assessment method: When the number of vehicles entering the section is greater than the number of vehicles leaving, the section is considered abnormal. This means that the counting results of the rth warning device node and the r+1th warning device are analyzed, and the cumulative count fs is set to determine whether the abnormality is detected by the warning device vehicle data.

[0029] S2: When the vehicle status in the interval is in a stable flow: the central server obtains the highway cross-section time flow series of two adjacent warning devices, calculates the average arrival time of vehicles between adjacent warning devices on the highway using the least squares method, and thus obtains the average vehicle speed in the interval; the stable flow traffic abnormality state is determined based on the relationship between the speed-density linear relationship model with moderate traffic density in the GreenShields model and the calculated interval vehicle speed;

[0030] S3: When the traffic flow in the section is congested: the range of the monitoring angle is recorded as L 监 , the required length of the road section is L 路 The difference between the total number of vehicles passing through the two warning devices before and after, that is, the number of vehicles on the road section in question is C ω After conversion, the number of vehicles entering this lane from other lanes is τ1, and the number of vehicles leaving this lane is τ2. The actual traffic volume of this lane is:

[0031]

[0032]

[0033] Among them, f SW f is the correction factor for hard shoulder width; W is the lane width correction factor; f LG is the truck correction factor; P T and P R E is the proportion of trucks, buses and tourist buses on the expressway; T and E R The number of trucks, buses and tourist buses on the highway lane is converted into equivalent values ​​of passenger cars; by calculating the difference between the actual traffic volume of the road ahead and the maximum traffic volume of the road, the appropriate diversion direction is compared to enable lane change and diversion in advance.

[0034] Furthermore, the implementation process of step S2 is as follows:

[0035] Assume that the average arrival time of highway traffic flow T represents the average time taken by vehicles in the interval from the rth warning device to the r+1th warning device, the average speed of highway traffic flow V represents the average speed between the rth warning device and the r+1th warning device, and the vehicle speed on the highway is 16.7~33.4m / s; the distance between two adjacent warning device nodes is L, then the average arrival time range of the interval vehicles is The average speed of vehicles in the interval ranges from [16.7, 33.4];

[0036] The average arrival time of vehicles in the interval is calculated using the recorded time flow series number n of the highway section. The total number of vehicles in the counting interval is A. Ideally, all vehicles in the counting interval have the same speed. The formula for the distance traveled is:

[0037] AL=AVT

[0038] In fact, the driving speed of each vehicle in the counting interval is different, and each group has vehicles that arrive at a time different from the average arrival time T. n Represents the number of vehicles that the nth group of vehicles passes through at the average arrival time T, so the total distance traveled by the n groups of vehicles is:

[0039] AL=B1VT+σ1+B2VT+σ2+B3VT+σ3+…+B n VT+σ n

[0040] =(B1+B2+B3+…B n )VT+(σ1+σ2+σ3+…+σ n )

[0041] σ is the random error, i.e., the distance traveled by vehicles in the group that differ from the average arrival time T:

[0042] σ n =v1t1+v2t2+…v j t j

[0043] σ n The smaller the value, the closer the average arrival time of vehicles in the interval is to T, and the more accurate the average arrival time T and the average speed V are. The distance error is converted into whether the vehicle arrives at the r+1 warning device from the rth warning device at the average arrival time T. The more vehicles arrive at this time, the more accurate the average arrival time is. The average arrival time T is calculated by sliding the window. As we move up, the error in the traffic counting results of the highway sections between adjacent warning devices is constantly changing, but there is a minimum value:

[0044]

[0045] Where ξ represents S r+1,t and S r,t-T The error caused by the difference between the average arrival time T and the arrival time of vehicles in the sequence, S r+1,t It is expressed as the number of vehicles arriving at the r+1 warning device per unit time at time t, S r,t-T It is expressed as the number of vehicles passing through the section of the r-th warning device at time tT, where T is Sliding within the time period is used to calculate the minimum error in the time series, and the minimum value ξ is selected min , the corresponding time T at this time is the average arrival time of highway traffic flow;

[0046] The average speed V of the interval is calculated by the average arrival time T of traffic flow at adjacent warning devices on the highway:

[0047]

[0048]

[0049]

[0050] Where V is the average speed of the section; V i is the speed of the i-th vehicle; A is the number of vehicles traveling on the road section within the counting interval; t i is the travel time of the i-th vehicle;

[0051] For the GreenShields model, the traffic flow speed and traffic flow density are linearly related, satisfying:

[0052]

[0053] Where V G Traffic flow speed; V f is the road speed limit; k j is the congestion density; K represents the traffic flow density of the interval, that is, the number of vehicles N in the area is obtained by the difference in vehicle counts recorded by adjacent warning device nodes. Obtain the traffic flow density of the interval at this moment, substitute it into the GreenShields model to obtain the average speed of the interval at this moment; take d as the upper limit of the interval average speed sequence, and calculate the difference U of the interval average speed sequence obtained by the GreenShields model and the least squares method at the corresponding moment d =|V Gt -V t |Whether it is smooth or not: Check whether the vehicle operation in this section is abnormal.

[0054] Furthermore, the process of checking whether the vehicle operation in the section is abnormal by checking whether the difference of the average speed sequence in the section is stable is as follows:

[0055] The interval average speed difference sequence is [U1,U2,U3,U4…U d ], the interval average speed difference sequence is transformed into an increasing sequence by the accumulation generation operator

[0056]

[0057]

[0058]

[0059]

[0060] Set the average speed difference sequence between adjacent warning equipment nodes on the highway It can be connected by a d-order linear regression process; by judging whether the multivariate linear regression equation has a linear relationship, it can be determined whether the vehicle in the section has an abnormality:

[0061]

[0062] in, represents the autoregressive coefficient, ε represents the error term, which has independence and normality, ε ij ~N(0,σ 2 );

[0063] Least squares method to solve for autoregressive coefficients:

[0064]

[0065] Test whether the d-order linear regression equation conforms to the linear regression: Model Assumptions The original hypothesis is H0: all autoregressive coefficients are 0, Alternative hypothesis H1: There is at least one autoregressive coefficient Not 0;

[0066] If the null hypothesis is rejected in the model hypothesis, it means that the average speed difference sequence of adjacent warning equipment nodes on the highway can form a d-order linear regression equation; by derivation, the d-order linear regression equation is obtained. The slope at the point, the difference between the interval average speed series obtained by the cumulative generation operator should be stable when it is stable, so the slope obtained by the d-order linear regression equation at point The relationship between the slope at and the overall slope determines whether the sequence difference is smooth; the overall slope of the d-order linear regression equation is used Instead of the slope x at all points i The average value is used to calculate the variance, and the threshold value & is set. If the variance value S 2 ≤&, it means that the vehicle operation status in this section is normal. If S 2 >&Then the vehicle operation status in this section is abnormal;

[0067] If the null hypothesis is not rejected in the model assumptions, it means that the average speed difference sequence between adjacent warning device nodes on the highway cannot form a d-order linear regression equation, the average speed difference sequence in this interval is not stable, and the vehicle operation is abnormal.

[0068] The present invention provides a highway abnormal traffic assessment system based on early warning device counting, comprising:

[0069] Data collection module: The warning device generates a time series of highway cross-section flow and vehicle operation status based on vehicle driving conditions; and sends the information to the central server for processing;

[0070] The central processing unit analyzes and processes the data transmitted by the early warning device, resends the corresponding instructions back to the designated early warning device and executes the corresponding instructions;

[0071] The traffic flow state classification module calculates the average flow rate of the current road section based on the vehicle operation status, and then calculates the traffic volume of the current road section. It also uses the traffic volume to calculate the highway service level ε level to determine the road congestion level, that is, different traffic flow states;

[0072] Abnormal state determination and warning module: inspect and determine traffic abnormalities in adjacent warning equipment sections of the highway according to different traffic conditions, and perform corresponding warning processing for different abnormal conditions.

[0073] Beneficial effects: Compared with the existing technology, the beneficial effects of the present invention are: based on the principle of automatic blocking of railway transportation, the present invention divides the highway into different blocks to obtain the operating status and location information of vehicles on the highway; through this information, the abnormal status of vehicles in each block of the highway can be calculated and the block location can be determined to facilitate the response to traffic accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 Flowchart of the highway abnormal traffic assessment method based on early warning device counting;

[0075] Figure 2 This is a schematic diagram of the data adjustment process of the early warning device under the free flow state of the present invention. DETAILED DESCRIPTION

[0076] The present invention will be described in further detail below with reference to the accompanying drawings.

[0077] Based on the principle of automatic blocking of railway transportation, the highway is divided into different blocks to obtain the running status and location information of the vehicles on the highway. With this information, the abnormal status of the vehicles in each block of the highway can be calculated and the location of the block can be determined to facilitate the response. Figure 1 As shown, the present invention provides a highway abnormal traffic assessment method based on early warning device counting, which determines the vehicle running status in the area and warns the following vehicles through the warning device to slow down in advance to avoid the occurrence of secondary accidents. The method specifically includes the following steps:

[0078] Step 1: The early warning device generates a highway cross-sectional time flow sequence based on the vehicle operating conditions, and sends the cross-sectional time flow sequence and vehicle operating conditions and other information to the central server for processing.

[0079] Warning device counting involves calculating the time series of traffic passing through the warning device's cross-section per unit time, using warning devices installed at a specific distance from the highway edge. Warning device features include a warning device information processor, ultrasonic sensor, memory, solar power supply, warning device, receiver, and transmitter.

[0080] The early warning device information processor uses ultrasonic sensors to collect highway cross-sectional time flow series and vehicle operating status data. The highway cross-sectional time flow series is the number of vehicles passing through a one-way section of the highway per unit time. For example, the cumulative number of vehicles passing per unit time recorded within a time period ns is the cross-sectional time flow series, and the number of such series is n. This information is stored in memory, and the transmitter transmits the location and vehicle information stored in the memory to a central server for processing. If a vehicle experiences an abnormality, the central server transmits a warning message to the following early warning device. The receiver then transmits the message to the early warning device information processor, which activates the warning device to warn the following vehicle. If the vehicle returns to normal, the warning device is deactivated. For example, one early warning device is placed in the middle of the highway every kilometer.

[0081] A central server, located in a highway service area or at a highway worker's workstation, collects and analyzes data from warning devices. Warning devices within a specific area transmit the collected data via a transmitter to the central server. After analysis, the server sends the corresponding instructions back to the designated warning device, which then executes the instructions.

[0082] Step 2: Divide the congestion degree of the interval according to the highway service level ε level, and divide the vehicle operation status of the interval into free flow, stable flow, and congested flow.

[0083] Due to the continuity and stability of the driving speed on the highway, the operating status of the vehicles passing through the front and rear warning devices will not change much in a short period of time. Therefore, in order to ensure the accuracy, the data recorded by the front (rth warning device) and rear (r+1th warning device) two warning devices are used to preliminarily estimate the average flow rate F of the interval.

[0084] Assuming that the distance between the two warning devices is L, and the estimated average speed of the interval is u, the calculated flow rate time of the interval is The two early warning devices are calculated separately Time flow rate. Because the three lanes in the expressway have different capacities, the reduction coefficient between the lanes is K n According to relevant data, the capacity of the first lane is 1 (i.e., 100%, k1=1), the capacity of the second lane is 0.8-0.9 of the first lane (i.e., k2=0.8-0.9), and the capacity of the third lane is 0.65-0.8 (i.e., k3=0.65-0.8). By normalizing the data, the maximum capacity and actual traffic volume of the three lanes are unified to the standard of the leftmost lane (i.e., the first lane). The average of the flow rates calculated by the two warning devices before and after is selected as the average flow rate of the current section, and the interval traffic volume is obtained from this:

[0085]

[0086]

[0087]

[0088] in, Representative The average flow rate recorded by the early warning device; F r 、F r+1 represent the average flow rates recorded by the rth warning device and the r+1th warning device respectively; F is the average flow rate between the two warning devices; F r1 、F r2 、F r3 represent the average flow rates of the three lanes recorded by the rth warning device; F (r+1)1 、F (r+1)2 、F (r+1)3 represents the average flow rate of the three lanes recorded by the r+1th warning device; L is the distance between the two warning devices; u is the average speed between the two warning devices preliminarily estimated based on previous data; K n is the reduction factor between lanes; is the average headway, i.e. The average value of the time difference of vehicles passing the warning device within the time period; Q is the traffic volume in this time interval.

[0089] Assume that the capacity of the leftmost road (the first road) is C1, and the capacity of the second and third roads are C2 and C3. Then according to the reduction coefficient K between the lanes n , normalizing the standards of all lanes to the leftmost road (the first road), then the maximum traffic capacity of the interval is:

[0090]

[0091] The degree of road congestion is determined by calculating the highway service level ε. The road congestion degree is calculated by calculating the ratio of the road traffic volume in the section to the maximum traffic volume. The formula is as follows:

[0092]

[0093] Based on the highway service level, the value of ε can be used to determine when the vehicle traffic status in the section is free flow, stable flow, or congested flow. When ε ≤ 0.35, the vehicle congestion between adjacent warning device nodes is free flow; when 0.35 < ε ≤ 0.9, the vehicle congestion between adjacent warning device nodes is stable flow; and when ε > 0.9, the vehicle congestion between adjacent warning device nodes is congested flow.

[0094] Step 3: For different traffic conditions, inspect and determine the traffic anomalies in the adjacent warning equipment sections of the highway. Then, issue corresponding warnings based on the abnormal conditions. The specific method is as follows:

[0095] Free-flow and congested-flow vehicle states are relatively random; stable-flow vehicles travel more smoothly, with minimal interaction between vehicles. By dividing free-flow and stable-flow traffic into sections of highway congestion, traffic anomalies in adjacent sections of highway warning equipment can be examined and determined, and warnings can be issued for these anomalies. If a congested-flow section exists for only one or two consecutive sections and does not spread over time, this indicates an anomaly within that congested-flow section, and a simple warning can be sufficient. However, over long distances, where multiple consecutive sections exhibit congested-flow conditions, a simple warning to slow down vehicles is ineffective. By combining warning equipment with existing monitoring equipment, the remaining capacity of the road ahead can be estimated, and warning lights can be used to divert vehicles behind, thus reducing congestion duration.

[0096] (3.1) Free-flow state vehicle anomaly assessment method: When the traffic flow is in the free-flow state, the vehicles are subject to few constraints. When more vehicles enter this section than leave, this section of the road is in an abnormal state. It is judged according to the cross-sectional time flow sequence recorded by adjacent warning devices on the highway. It is set to judge whether it is abnormal based on the cumulative vehicle data passing through the warning device, that is, analyze the counting results of the r-th warning device node and the (r + 1)-th warning device. Use the counting data at the time boundary to judge. For the convenience of expressing time it is denoted as qs, denoted as zs; fs is the set test interval group, and a 5s interval is used for regular counting tests.

[0097] Suppose the number of vehicles passing through the r-th warning device within 0 to fs in the first group is denoted as m1, and the number of vehicles passing through the (r + 1)-th warning device within z to f + qs is denoted as w1. The number of vehicles passing through the r-th warning device within f to 2fs in the second group is m2, and the number of vehicles passing through the (r + 1)-th warning device within f + z to 2f + qs is w2, and so on.

[0098] Transmit the data to the central server for data analysis of m and w:

[0099] If m1 = w1, then the vehicles in the first group just pass through completely, and proceed to process the next group of data.

[0100] If m1 < w1, then there is an abnormal number of vehicles in the first group of vehicles, and mark this section as a key section.

[0101] If m1 > w1, then there are abnormal vehicles in the first group of vehicles, and mark this section as an abnormal section.

[0102] Perform a process analysis on the key section. The possible situation is that the data recorded when the vehicles passing through the r-th warning device within f to 2fs in the second group is m2, but within this section, the vehicle speed exceeds the predetermined speed, resulting in the number of vehicles leaving exceeding the number of vehicles entering, so there is no abnormality, but data analysis adjustment is required.

[0103] As Figure 2 shown, the adjustment process is as follows: The passing data of the second group is changed from m2 to m2 - (w1 - m1), and judge the relationship between m2 - (w1 - m1) and w2. If they are equal, it means that the first group of data can be adjusted normally in the second group of data. If not, adjust the third group of data and perform the same steps as the previous group (the steps of the first group and the second group) with the second group, and so on.

[0104] (3.2) When the vehicle flow in the interval is stable: The central server obtains the time-varying traffic flow series of the highway sections between two adjacent warning devices and calculates the average arrival time of vehicles between adjacent warning devices using the least squares method. This calculation yields the average vehicle speed in the interval. The GreenShields model is suitable for roads with relatively stable capacity and traffic conditions with relatively stable vehicle density and flow. Abnormal traffic conditions in the interval can be determined based on the relationship between the speed-density linear relationship model with moderate traffic density and the desired interval speed.

[0105] The central server obtains the highway cross-sectional time flow sequence from the early warning device, that is, the number of vehicles passing through the one-way section of the highway per unit time recorded by the early warning device. The cumulative record of the number of vehicles passing per unit time recorded in time ns is the cross-sectional time flow sequence. The number of sequences is n, and the number of all vehicles in the calculation interval is A. Assuming that the distance traveled by each vehicle in the interval is L, the distance is exchanged for speed, and the difference between the vehicles and the relative average arrival time is taken as the error. Based on the least squares method, the average arrival time of vehicles between adjacent early warning devices on the highway is calculated, and the average speed of vehicles in the interval is obtained.

[0106] Assume that the average arrival time of highway traffic flow T represents the average time taken by vehicles in the interval from the rth warning device to the r+1th warning device, the average speed of highway traffic flow V represents the average speed between the rth warning device and the r+1th warning device, the vehicle speed on the highway is 60~120km / h, that is, 16.7~33.4m / s, and the distance between two adjacent warning device nodes is L. Then the average arrival time of traffic flow ranges from The average speed range of traffic flow interval is [16.7, 33.4].

[0107] The average arrival time of vehicles in the interval is calculated using the recorded highway section time flow series number n. The value range of n is: Ensure that all vehicles in this section are within the counting interval; assuming L = 1000m, the sequence value of n is 60.

[0108] The average arrival time of vehicles in a certain interval is calculated using data from n sets of highway traffic flow sequences. The total number of vehicles in this interval is A. Ideally, all vehicles in this interval have the same speed. The distance traveled is given by:

[0109] AL=AVT (6)

[0110] In fact, the driving speed of each vehicle in the counting interval is different, and each group has vehicles that arrive at a time different from the average arrival time T, where B nRepresents the number of vehicles that the nth group of vehicles passes through at the average arrival time T, and the total distance traveled by n groups of vehicles:

[0111] AL=B1VT+σ1+B2VT+σ2+B3VT+σ3+…+B n VT+σ n

[0112] =(B1+B2+B3+…B n )VT+(σ1+σ2+σ3+…+σ n ) (7)

[0113] σ is the random error, that is, the distance traveled by vehicles in the group that are different from the average arrival time T:

[0114] σ n =v1t1+v2t2+…v j t j (8)

[0115] It can be seen from the above formula that within this interval σ n The smaller it is, the closer the average arrival time of vehicles in the actual interval is to T, and the more accurate the average arrival time T and average speed V are. The range of the average arrival time of the interval is That is, a sliding window is used within the time range, and the distance error σ generated by the traffic flow counting results of adjacent warning devices is calculated each time it moves, and the minimum error is found by the least squares method. Because all vehicles travel the same distance within the interval, the distance error is converted into whether the vehicle arrives from the rth warning device to the r+1th warning device at the average arrival time T. The more vehicles arrive at this time, the more accurate the average arrival time is. The average arrival time T is calculated by sliding the window. In China Mobile, the error in the traffic flow counting results of highway sections between adjacent warning devices is constantly changing, but there is a clear downward and upward trend over a period of time, and there is a minimum value:

[0116]

[0117] Where ξ represents S r+1,t and S r,t-T The error caused by the difference between the average arrival time T and the arrival time of vehicles in the sequence, S r+1,t It is expressed as the number of vehicles arriving at the r+1 warning device per unit time at time t, S r,t-T It is represented by the number of vehicles passing through the section of the r-th warning device at time tT, and T is represented by the number of vehicles passing through the section of the r-th warning device at time tT. Sliding within the time period is used to calculate the minimum error in the time series, and the minimum value ξ is selected min, the corresponding time T is the average arrival time of highway traffic flow.

[0118] The average arrival time T of traffic flow in adjacent warning equipment sections of the highway can be used to calculate the average vehicle speed V in the interval:

[0119]

[0120]

[0121]

[0122] Where V is the average speed of the section; V i is the speed of the i-th vehicle; A is the number of vehicles traveling on the road section within the counting interval; t i is the travel time of the i-th vehicle.

[0123] For the GreenShields model, the traffic flow speed and traffic flow density are linearly related, satisfying:

[0124]

[0125] Where: V G Traffic flow speed, V f K is the road speed limit; j is the congestion density; K represents the traffic flow density of the interval. The number of vehicles N in the area is obtained by the difference in vehicle counts recorded by adjacent warning device nodes. The traffic flow density of the interval at this moment is obtained and substituted into the GreenShields model to obtain the average speed of the interval at this moment. Taking d as the upper limit of the interval average speed sequence, the difference U of the interval average speed sequence obtained by calculating the GreenShields model at the corresponding moment and the least squares method is obtained. d =|V Gt -V t |Whether it is smooth or not: Check whether the vehicle operation in this section is abnormal.

[0126] The interval average speed difference sequence is [U1,U2,U3,U4…U d The average speed difference sequence of the interval is transformed into an increasing sequence through the accumulation generation operator (AGO).

[0127]

[0128]

[0129]

[0130]

[0131] Set the average speed difference sequence between adjacent warning equipment nodes on the highway The d-order linear regression equation can be used to connect the two. By determining whether the multivariate linear regression equation has a linear relationship, it is determined whether the vehicle in the section has an abnormality.

[0132]

[0133] in, represents the autoregressive coefficient, ε represents the error term, which has independence and normality, ε ij ~N(0,σ 2 ).

[0134] Least squares method to solve for autoregressive coefficients:

[0135]

[0136] Test whether the d-order linear regression equation conforms to the linear regression: Model Assumptions The original hypothesis is H0: all autoregressive coefficients are 0, Alternative hypothesis H1: There is at least one autoregressive coefficient Not 0.

[0137] If the null hypothesis is rejected in the model hypothesis, it means that the average speed difference sequence of adjacent warning equipment nodes on the highway can form a d-order linear regression equation. By derivation, the d-order linear regression equation is obtained. The slope at point , the difference between the interval average velocity series obtained by the two methods should be stable when the slope obtained by the cumulative generation operator (AGO) is stable, so the d-order linear regression equation is used at point The relationship between the slope at and the overall slope determines whether the sequence difference is smooth.

[0138] Use the overall slope of the d-order linear regression equation Instead of the slope x at all points i The variance of these values ​​is obtained by calculating the mean value S 2 The test is a direct relationship between the autoregressive coefficients.

[0139]

[0140]

[0141] Set the threshold &, if the variance value S 2 ≤&, it means that the vehicle operation status in this section is normal; if S 2 >&Then the vehicle operation status in this section is abnormal.

[0142] If the null hypothesis is not rejected in the model assumptions, it means that the average speed difference sequence between adjacent warning device nodes on the highway cannot form a d-order linear regression equation, the average speed difference sequence in this interval is not stable, and the vehicle operation in this area is abnormal.

[0143] (3.3) When traffic flow in a section is congested, vehicles are in an extremely unstable state, approaching or reaching maximum traffic volume. Even a small increase in traffic volume or a small disturbance within the traffic flow can cause major operational problems, even traffic disruptions. Driving freedom, comfort, and convenience are extremely low, significantly hindering drivers. If this condition persists for only one or two consecutive sections and does not spread over time, it indicates an abnormal condition within the congested section, and a simple warning should be issued.

[0144] If congestion persists for multiple consecutive sections over a long distance, warning vehicles to slow down will have little effect. By integrating early warning devices with existing monitoring equipment, the remaining capacity of the road ahead can be estimated. Warning lights can then be used to divert vehicles behind, reducing congestion. Based on existing early warning devices, an additional monitoring device is deployed every other early warning device to analyze the status of vehicles around it, such as vehicle type ratio and lane change probability. This information, combined with the data from the early warning devices, is used to analyze current road conditions and provide information to vehicles behind, thus reducing congestion.

[0145] Assume that the range of the monitoring angle is recorded as L 监 , the required length of the road section is L 路 Although the monitoring field of view is far less than the length of the road section, the proportion of vehicle types entering each lane can be roughly understood. The detailed process of lane changes between vehicles during highway driving cannot be fully understood. For this probabilistic problem, a sampling method is used for calculation, replacing the whole with the part. By analyzing the proportion of vehicle types on each road on the highway and the probability of lane change, the number of vehicles on the road obtained by the early warning device is adjusted. The number of vehicles in this section obtained by the early warning device can be obtained by the difference in the total number of vehicles passed by the two early warning devices before and after. Assume that the difference in the total number of vehicles passed by the two early warning devices before and after, that is, the number of vehicles on the required section of the lane is C ω Assuming that the number of vehicles entering this lane from other lanes is τ1 and the number of vehicles leaving this lane is τ2, the actual traffic volume of this lane is:

[0146]

[0147]

[0148] Where, fSW f is the correction factor for hard shoulder width; W is the lane width correction factor; f LG is the truck correction factor; P T and P R E is the proportion of trucks, buses and tourist buses on the expressway; T and E R It is the equivalent value of trucks, buses and tourist buses on the expressway lane converted into passenger cars.

[0149] By calculating the difference between the actual traffic volume of the road ahead and the maximum traffic volume of the road, the appropriate diversion direction can be compared, allowing lane change and diversion to be carried out in advance, reducing congestion time and ensuring smooth road traffic.

[0150] The present invention also provides a highway abnormal traffic assessment system based on early warning device counting, comprising:

[0151] In the data collection module, the early warning device generates a highway cross-section time flow sequence and vehicle operation status according to the vehicle driving conditions; and sends the information to the central server for processing.

[0152] The central processing unit analyzes and processes the data transmitted by the early warning device, resends the corresponding instructions back to the designated early warning device and executes the corresponding instructions.

[0153] The traffic flow state classification module calculates the average flow rate of the current road section based on the vehicle operation status, and then calculates the traffic volume of the current road section; and calculates the highway service level ε level through traffic volume to judge the road congestion level, that is, different traffic flow states.

[0154] Abnormal state determination and warning module: inspect and determine traffic abnormalities in adjacent warning equipment sections of the highway according to different traffic conditions, and perform corresponding warning processing for different abnormal conditions.

[0155] The above embodiments are intended only to illustrate the technical concepts and features of the present invention. Their purpose is to enable those skilled in the art to understand the contents of the present invention and implement them accordingly. They are not intended to limit the scope of protection of the present invention. Any equivalent changes or modifications made in accordance with the spirit of the present invention are intended to be covered by the scope of protection of the present invention.

Claims

1. A highway abnormal traffic assessment method based on early warning device counting, characterized in that: The following steps are involved: (1) Data collection: The early warning equipment generates a time flow series of highway sections and vehicle operation status based on vehicle driving conditions; and sends the information to a central server for processing; (2) Calculate the average flow rate of the current road section based on the vehicle operation status, and then calculate the traffic volume of the current road section; (3) Based on the traffic volume, the road congestion level is determined by calculating the highway service level ε level, i.e., different traffic flow states; the different traffic flow states include free flow state, stable flow state and congested flow state; (4) Inspect and determine traffic anomalies in adjacent warning equipment sections of the expressway according to different traffic flow conditions, and perform corresponding warning processing work according to different abnormal conditions; The implementation process of step (4) is as follows: S1: Free-flow state vehicle anomaly assessment method: When more vehicles enter the section than leave, the section is considered abnormal. That is, the counting results of the rth warning device node and the r+1th warning device are analyzed, and the cumulative count fs is set to determine whether it is abnormal based on the warning device vehicle data; S2: When the vehicle status in the interval is in a stable flow: the central server obtains the highway cross-section time flow series of two adjacent warning devices, calculates the average arrival time of vehicles between adjacent warning devices on the highway using the least squares method, and thus obtains the average vehicle speed in the interval; the stable flow traffic abnormality state is determined based on the relationship between the speed-density linear relationship model with moderate traffic density in the GreenShields model and the calculated interval vehicle speed; S3: When the traffic flow in the section is congested: the range of the monitoring angle is recorded as L 监 , the required length of the road section is L 路 The difference between the total number of vehicles passing through the two warning devices before and after, that is, the number of vehicles on the road section in question is C ω After conversion, the number of vehicles entering this lane from other lanes is τ1, and the number of vehicles leaving this lane is τ2. The actual traffic volume of this lane is: Among them, f SW f is the correction factor for hard shoulder width; W is the lane width correction factor; f LG is the truck correction factor; P T and P R E is the proportion of trucks, buses and tourist buses on the expressway; T and E R The equivalent number of trucks, buses, and coaches on the highway is converted into passenger cars. By calculating the difference between the actual traffic volume on the road ahead and the maximum traffic volume on that road, the appropriate diversion direction is determined, allowing traffic to change lanes in advance. The implementation process of step S2 is as follows: Assume that the average arrival time of highway traffic flow T represents the average time taken by vehicles in the interval from the rth warning device to the r+1th warning device, the average speed of highway traffic flow V represents the average speed between the rth warning device and the r+1th warning device, and the vehicle speed on the highway is 16.7~33.4m / s; the distance between two adjacent warning device nodes is L, then the average arrival time range of the interval vehicles is The average speed of vehicles in the interval ranges from [16.7, 33.4]; The average arrival time of vehicles in the interval is calculated using the recorded time flow series number n of the highway section. The total number of vehicles in the counting interval is A. Ideally, all vehicles in the counting interval have the same speed. The formula for the distance traveled is: AL=AVT In fact, the driving speed of each vehicle in the counting interval is different, and each group has vehicles that arrive at a time different from the average arrival time T. n Represents the number of vehicles that the nth group of vehicles passes through at the average arrival time T, so the total distance traveled by the n groups of vehicles is: AL=B1VT+σ1+B2VT+σ2+B3VT+σ3+…+B n VT+s n =(B1+B2+B3+…B n )VT+(σ1+σ2+σ3+…+σ n ) σ is the random error, i.e., the distance traveled by vehicles in the group that differ from the average arrival time T: σ n =v1t1+v2t2+…v j t j σ n The smaller the value, the closer the average arrival time of vehicles in the interval is to T, and the more accurate the average arrival time T and the average speed V are. The distance error is converted into whether the vehicle arrives at the r+1 warning device from the rth warning device at the average arrival time T. The more vehicles arrive at this time, the more accurate the average arrival time is. The average arrival time T is calculated by sliding the window. As we move up, the error in the traffic counting results of the highway sections between adjacent warning devices is constantly changing, but there is a minimum value: Where ξ represents S r+1,t and S r,t-T The error caused by the difference between the average arrival time T and the arrival time of vehicles in the sequence, S r+1,t It is expressed as the number of vehicles arriving at the r+1 warning device per unit time at time t, S r,t-T It is expressed as the number of vehicles passing through the section of the r-th warning device at time tT, where T is Sliding within the time period is used to calculate the minimum error in the time series, and the minimum value ξ is selected min , the corresponding time T at this time is the average arrival time of highway traffic flow; The average speed V of the interval is calculated by the average arrival time T of traffic flow at adjacent warning devices on the highway: Where V is the average speed of the section; V i is the speed of the i-th vehicle; A is the number of vehicles traveling on the road section within the counting interval; t i is the travel time of the i-th vehicle; For the GreenShields model, the traffic flow speed and traffic flow density are linearly related, satisfying: Where V G Traffic flow speed; V f is the road speed limit; k j is the congestion density; K represents the traffic flow density of the interval, that is, the number of vehicles N in the area is obtained by the difference in vehicle counts recorded by adjacent warning device nodes. Obtain the traffic flow density of the interval at this moment, substitute it into the GreenShields model to obtain the average speed of the interval at this moment; take d as the upper limit of the interval average speed sequence, and calculate the difference U of the interval average speed sequence obtained by the GreenShields model and the least squares method at the corresponding moment d =|V Gt -V t |Whether it is smooth or not: Check whether the vehicle operation in this section is abnormal.

2. The method for evaluating abnormal highway traffic based on early warning device counting according to claim 1, characterized in that: The warning device includes an information processor, an ultrasonic sensor, a memory, a solar power supply device, a warning device, a receiving device and a transmitting device; the information processor collects the time flow sequence of the highway section and the vehicle operation status data through the ultrasonic sensor and saves the information into the memory; The transmitting device transmits the location information and vehicle information in the memory to the central server for information processing. If an abnormal condition occurs in the vehicle, the central server will transmit the warning information to the rear warning device. After receiving the information, the receiving device will transmit the information to the warning device information processor. The warning device information processor will turn on the warning device to warn the rear vehicle. If the vehicle returns to normal, the warning device will be turned off.

3. The method for evaluating abnormal highway traffic based on early warning device counting according to claim 1, characterized in that: The implementation process of step (2) is as follows: The data was normalized, and the maximum capacity and actual traffic volume of the three lanes were unified to the standard of the leftmost lane. The headway was the difference in time between the two vehicles in the same lane passing through the projection section of the warning device, that is, the time difference between the two vehicles in the same lane passing through the section. The average flow rate of the two warning devices before and after is the average flow rate of the current section, and the traffic volume of the section can be calculated: in, Representative The average flow rate recorded by each warning device; F r 、F r+1 represent the average flow rates recorded by the rth warning device and the r+1th warning device respectively; F is the average flow rate between the two warning devices; F r1 、F r2 、F r3 represent the average flow rates of the three lanes recorded by the rth warning device; F (r+1)1 、F (r+1)2 、F (r+1)3 represents the average flow rate of the three lanes recorded by the r+1th warning device; L is the distance between the two warning devices; u is the average speed between the two warning devices preliminarily estimated based on previous data; K n is the reduction factor between lanes; is the average headway, i.e. The average value of the time difference of vehicles passing the warning device within the time period; Q is the traffic volume in this time interval.

4. The method for evaluating abnormal highway traffic based on early warning device counting according to claim 3 is characterized in that: The implementation process of step (3) is as follows: The capacity of the leftmost road is C1, the capacity of the second road and the third road is C2 and C3 respectively; then according to the reduction coefficient K between the lanes n , normalizing the standards of all lanes to the leftmost road, then the maximum traffic capacity of the interval is: The degree of road congestion can be determined by calculating the highway service level ε:

5. The method for evaluating abnormal highway traffic based on early warning device counting according to claim 1, characterized in that: The basis for judging different traffic flow states in step (3) is: When ε≤0.35, the vehicle congestion between adjacent warning device nodes is in a free flow state; When 0.35<ε≤0.9, the vehicle congestion between adjacent warning device nodes is in a stable flow state; When ε>0.9, the vehicle congestion level between adjacent warning device nodes is in a congested flow state.

6. The method for evaluating abnormal highway traffic based on early warning device counting according to claim 1, characterized in that: The process of checking whether the vehicle operation in the interval is abnormal by checking whether the interval average speed sequence difference is stable is as follows: The interval average speed difference sequence is [U1,U2,U3,U4…U d ], the interval average speed difference sequence is transformed into an increasing sequence by the accumulation generation operator Set the average speed difference sequence between adjacent warning equipment nodes on the highway It can be connected by a d-order linear regression process; by judging whether the multivariate linear regression equation has a linear relationship, it can be determined whether the vehicle in the section has an abnormality: in, Represents the autoregressive coefficient, Φ represents the error term, with independence and normality, Φ ij ~N(0,σ 2 ); Least squares method to solve for autoregressive coefficients: Test whether the d-order linear regression equation conforms to the linear regression: Model Assumptions The original hypothesis is H0: all autoregressive coefficients are 0, Alternative hypothesis H1: There is at least one autoregressive coefficient Not 0; If the null hypothesis is rejected in the model hypothesis, it means that the average speed difference sequence of adjacent warning equipment nodes on the highway can form a d-order linear regression equation; by derivation, the d-order linear regression equation is obtained. The slope at the point, the difference between the interval average speed series obtained by the cumulative generation operator should be stable when it is stable, so the slope obtained by the d-order linear regression equation at point The relationship between the slope at and the overall slope determines whether the sequence difference is smooth; the overall slope of the d-order linear regression equation is used Instead of the slope x at all points i The average value is used to calculate the variance, and the threshold value & is set. If the variance value S 2 ≤&, it means that the vehicle operation status in this section is normal. If S 2 >&Then the vehicle operation status in this section is abnormal; If the null hypothesis is not rejected in the model assumptions, it means that the average speed difference sequence between adjacent warning device nodes on the highway cannot form a d-order linear regression equation, the average speed difference sequence in this interval is not stable, and the vehicle operation is abnormal.

7. A highway abnormal traffic assessment system based on early warning device counting using the method according to any one of claims 1 to 6, characterized in that: include: Data collection module, the early warning device generates highway cross-section time flow series and vehicle operation status according to vehicle driving conditions; and sends the information to a central server for processing; The central processing unit analyzes and processes the data transmitted by the early warning device, resends the corresponding instructions back to the designated early warning device and executes the corresponding instructions; Traffic flow state classification module calculates the average flow rate of the current road section based on the vehicle operation status, and then calculates the traffic volume of the current road section; The highway service level ε level is calculated by traffic volume to judge the road congestion level, that is, different traffic flow states; Abnormal state determination and warning module: inspect and determine traffic abnormalities in adjacent warning equipment sections of the highway according to different traffic conditions, and perform corresponding warning processing for different abnormal conditions.

Images