A method for predicting highway traffic flow speed considering inter-vehicle motion state difference and hysteresis effect

Abstract

The application discloses a kind of freeway traffic flow speed prediction methods considering the difference between inter-vehicle motion state and hysteresis effect.The specific prediction method is divided into six steps, namely step one, calculate the disturbance of single vehicle to traffic flow;Step two, the disturbance of single vehicle to traffic flow is distributed to road unit;Step three, calculate the total disturbance of all road units;Step four, construct hysteresis effect model;Step five, determine the hysteresis period number and coefficient of hysteresis effect model;Step six, predict traffic flow speed.The method collects vehicle real-time trajectory information according to the intelligent degree of freeway and predicts traffic flow speed with the aid of induction coil, vehicle-mounted GPS and other detection equipment, which is beneficial to timely take traffic control measures to prevent and alleviate congestion, thereby improving the operation efficiency and safety of traffic system, and providing theoretical basis and model support for freeway traffic flow speed prediction and traffic flow organization control.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of highway traffic flow prediction, and particularly relates to a highway traffic flow speed prediction method considering the motion state difference and hysteresis effect between vehicles. BACKGROUND

[0002] In recent years, with the continuous development of perception technology and communication technology, intelligent monitoring devices are increasingly popular, providing support for accurate and real-time acquisition of vehicle operation information. On the basis of comprehensive collection of highway traffic operation data, traffic flow prediction is carried out according to historical traffic information and real-time conditions, so as to determine the future trend of traffic operation state, which is of great significance for precise control of highway traffic conditions, timely prevention and rapid relief of congestion, and improvement of highway service level and operation efficiency.

[0003] In previous traffic flow prediction research, domestic and foreign researches used time series analysis, probability statistics method and machine learning algorithm to capture the uncertainty and complex characteristics of the time series change of traffic flow parameters, and carried out traffic state deduction modeling and traffic flow prediction in specific scenarios. For example, patent 202210277813.X obtains the GPS trajectory data of all vehicles in the research area within one month, calculates the traffic flow speed approximation value at an interval of 15 minutes, constructs a traffic pattern similarity graph and a multi-graph convolution network, and outputs the time series of traffic flow speed prediction value; patent 201811264868.7 collects traffic flow speed historical data within 180 days, establishes and calibrates a prediction model in a vehicle networking environment, and realizes short-term traffic flow speed prediction; "Short-Term Traffic Speed Forecasting Using a Deep Learning Method Based on Multitemporal Traffic Flow Volume" proposes a deep learning method based on traffic flow data to predict traffic speed based on the relationship between flow and speed in traffic flow theory; "AMulti-step Traffic Speed Forecasting Model Based on Graph Convolutional LSTM" proposes a deep learning model based on historical traffic flow speed data and network topology, uses graph convolution network (GCN) to process the spatial relationship features of traffic flow speed, and uses LSTM to process the time relationship features, realizing multi-step traffic flow speed prediction.

[0004] However, the existing research based on the historical data of macro traffic flow parameters such as flow and density or the traffic flow speed time series itself has the following three problems: first, the existing research ignores the influence of the micro vehicle motion difference on the traffic state. In actual traffic scenes, the vehicle motion state will be affected by the nearby other vehicles, and the change of the local vehicle motion state will also act on the overall traffic flow, and the result is that the average speed of the traffic flow changes after a period of time. Therefore, the change of the traffic flow speed is not only affected by the change of other macro parameters such as flow and density, but also the result of the micro vehicle motion difference; second, the existing research needs a large amount of historical data for traffic flow speed feature extraction and model training; third, the existing research based on historical traffic data for traffic flow speed prediction modeling can better reflect the change rule of the traffic flow speed time series in a specific scene and a specific period of time, but the prediction effect of these models will decrease significantly in the face of special situations and occasional events. Therefore, it is necessary to further analyze the relationship between the macro traffic flow speed change and the micro vehicle motion state, solve the problem of insufficient historical data leading to difficulty in model training and calibration, and ensure the precision and universal applicability of the traffic flow speed prediction method in various scenes.

[0005] Based on the above background, if the traffic flow speed is regarded as the dependent variable and the micro vehicle motion difference is regarded as the independent variable, and considering that the traffic flow speed change has relative continuity, the value of the dependent variable will be affected by the change of the independent variable and the value of the dependent variable in the previous period. This phenomenon that the dependent variable is affected by the lag value of itself and the lag value of the independent variable is usually described by using the lag effect model in economics. Therefore, it is urgent to design a highway traffic flow speed prediction method considering the motion state difference between vehicles and the lag effect, establish a lag effect model to express the influence of the micro vehicle motion state on the macro traffic flow speed, and calculate and output the time series of the traffic flow speed prediction value by taking the collected real-time traffic flow data as the model input, so as to provide theoretical basis and model support for highway traffic control. SUMMARY

[0006] In view of the deficiencies of the existing research, the technical problem to be solved by the present application is to provide a highway traffic flow speed prediction method considering the motion state difference between vehicles and the lag effect, so as to timely take traffic control measures, achieve the purpose of preventing and relieving congestion, thereby improving the operation efficiency and safety of the traffic system, and providing theoretical basis and model support for highway traffic flow speed prediction and traffic flow organization and control.

[0007] The present application is realized by adopting the following technical scheme, and the description is as follows:

[0008] 1. A highway traffic flow speed prediction method considering the motion state difference between vehicles and the lag effect, characterized in that it comprises the following steps:

[0009] Step 1: Calculate the turbulence of a single vehicle on traffic flow.

[0010] First, obtain the traffic flow density ρ, the speed and displacement of each vehicle within the area where the highway traffic flow speed needs to be predicted. The length of the area is no less than 150 meters, and the width includes all lanes in one direction of traffic. The traffic flow density ρ is the average number of vehicles per kilometer within the area. Calculate the traffic flow speed:

[0011]

[0012] The traffic flow velocity is the average speed of all vehicles within the area where the highway traffic flow velocity is to be predicted; N is the total number of vehicles within the area where the highway traffic flow velocity is to be predicted; i is the vehicle number; i = 1, 2, ..., N; v i Let i be the speed of vehicle i;

[0013] Second, establish a Cartesian coordinate system, with the intersection of the lane edge line near the central median and the upstream edge line of the area where the highway traffic flow speed is to be predicted as the origin. Set the direction of traffic flow speed as the positive x-axis and the direction perpendicular to the direction of traffic flow speed as the y-axis. The positive y-axis is defined by rotating 90 degrees counterclockwise along the positive x-axis.

[0014] Third, decompose the velocity of each vehicle along the positive x-axis and positive y-axis to obtain the velocity component v along the traffic flow velocity direction. i cosθ i The component of velocity v perpendicular to the direction of traffic flow velocity i sinθ i , where θ i Let θ be the angle between the direction of vehicle i's velocity and the direction of traffic flow velocity, 0 ≤ θ i ≤π;

[0015] Fourth, calculate the speed difference between each vehicle:

[0016] △v xij =v i cosθ i -v j cosθ j

[0017] △v xij v represents the velocity difference between vehicle i and vehicle j along the x-direction. j cosθ j Let i be the velocity component of vehicle j along the positive x-axis, and i be the vehicle number; j = 1, 2, ..., N;

[0018] △v yij =vi sinθ i -v j sinθ j

[0019] △v yij is the relative speed difference between vehicle i and vehicle j in the y direction; v j sinθ j is the relative speed of vehicle j in the positive y direction;

[0020] Fifth, calculate the relative displacement difference between vehicles:

[0021] △x ij = x i -x j

[0022] △x ij is the displacement x-coordinate difference between vehicle i and vehicle j; x i is the displacement x-coordinate of vehicle i; x j is the displacement x-coordinate of vehicle j;

[0023] △y ij = y i -y j

[0024] △y ij is the displacement y-coordinate difference between vehicle i and vehicle j; y i is the displacement y-coordinate of vehicle i; y j is the displacement y-coordinate of vehicle j;

[0025] Sixth, calculate the average vehicle spacing

[0026]

[0027] L is the length of the region to be predicted in the highway traffic flow speed; M is the number of lanes;

[0028] Seventh, use the turbulence to express the difference in the motion state of vehicles in the traffic flow, and calculate the turbulence E generated by a single vehicle to the traffic flow i :

[0029]

[0030] wherein, when the speed and direction of all vehicles in the traffic flow are the same, the turbulence generated by a single vehicle to the traffic flow is 0;

[0031] Step two, distribute the turbulence generated by a single vehicle to the traffic flow to the road unit

[0032] The highway traffic flow speed to be predicted region is divided into several equal rectangular road units, and the disturbance E i The single vehicle disturbance distribution value E i,p,q :

[0033]

[0034] p is the horizontal coordinate of the geometric center of the road unit; q is the vertical coordinate of the geometric center of the road unit; L i,p,q is the distance from the geometric center of the vehicle i to the geometric center (p, q) of the road unit;

[0035] The single vehicle disturbance distribution value of each vehicle to the road unit is summed up to calculate the total disturbance distribution value E p,q of the road unit.

[0036]

[0037] Step three, calculate the total disturbance of all road units

[0038] The total disturbance distribution value of all road units is summed up to calculate the total disturbance E a of the highway traffic flow speed to be predicted region.

[0039]

[0040] g is the road unit number, g = 1, 2, …, G; G is the total number of road units;

[0041] Step four, construct a hysteresis effect model

[0042] The hysteresis effect theory is used to express the functional relationship between the traffic flow speed and the total disturbance of the highway traffic flow speed to be predicted region, and a hysteresis effect model is established:

[0043]

[0044] wherein and are the traffic flow speeds at time t, time t-1, time t-2 and time t-m, respectively, wherein the time interval between time t and time t-1 is the time length of one hysteresis period, and the time length of one hysteresis period is not greater than 2 seconds and not less than 0.1 second; E a (t-l), E a (t-l-1), E a (t-l-2) and E a(t-l-2) are the total disturbances of the freeway traffic flow speed to be predicted at the time t-l, t-l-1, t-l-2 and t-l-s respectively; l is the initial lag number of the total disturbance of the freeway traffic flow speed to be predicted, l ∈ N; s is the maximum lag number of the total disturbance of the freeway traffic flow speed to be predicted, s ∈ N + ; m is the maximum lag number of the traffic flow speed, m ∈ N + ; α is the constant term of the lag effect model; β c is the coefficient, c = 0, 1, 2, …, s; γ d is the coefficient, d = 1, 2, …, m; u(t) is the random error;

[0045] Step five, determine the lag number and the coefficient of the lag effect model

[0046] The Akaike information criterion is used to construct the objective function:

[0047]

[0048] Wherein, AIC is the objective function; SSR is the residual sum of squares; n is the sample size, that is, the total number of time points contained in the time period input into the model;

[0049] Calculate the objective function value under different combinations of s, l and m values, and get the best s, l and m values corresponding to the minimum value of the objective function AIC;

[0050] The constant term and the coefficient of the lag effect model are estimated by using the stepwise regression method, and α, β c , γ d and u(t) are obtained;

[0051] Step six, predict the traffic flow speed

[0052] 6-1) The total disturbance of the freeway traffic flow speed to be predicted and the traffic flow speed are used as the input of the lag effect model, wherein the corresponding time points of the first and last terms of the total disturbance of the freeway traffic flow speed to be predicted are t-l and t-l-s, and the corresponding time points of the first and last terms of the traffic flow speed are t-1 and t-m; the predicted value of the traffic flow speed at the time t+1 is calculated as the output of the model, that is,

[0053] 6-2) the As a new model input, the first and last items of the total turbulence time series of the expressway traffic flow speed area to be predicted and the traffic flow speed time series are updated, and the corresponding time of the first and last items of the total turbulence time series of the expressway traffic flow speed area to be predicted is t-l+1 period and t-l-s+1 period respectively, and the corresponding time of the first and last items of the traffic flow speed time series is t and t-m+1 period respectively; the model output is calculated to obtain the traffic flow speed prediction value at t+2 period

[0054] 6-3) According to the steps of 6-1 and 6-2, the traffic flow speed at each future time is calculated to obtain the time series of traffic flow speed prediction values.

[0055] Compared with the prior art, the beneficial effects of the expressway traffic flow speed prediction method considering the motion state difference between vehicles and the hysteresis effect are:

[0056] (1) The method defines the motion state difference between micro vehicles as turbulence, realizes the quantitative expression of the turbulence of a single vehicle to the traffic flow, the single vehicle turbulence distribution value of the road unit, the total turbulence distribution value of the road unit and the total turbulence of the expressway traffic flow speed area to be predicted, and determines the influence of the change of the motion state of the micro vehicle on the macro traffic flow speed;

[0057] (2) The method establishes a hysteresis effect model with traffic flow speed as the dependent variable and the total turbulence of the expressway traffic flow speed area to be predicted as the independent variable, without the need for a large amount of historical traffic data for model training or calibration, and the model input is the collected real-time trajectory information of the vehicle, and the output is the time series of traffic flow speed prediction values, which reduces the limitation conditions of data source and data amount;

[0058] (3) The method is suitable for expressway traffic flow speed prediction under various traffic conditions, and compared with the prediction method based on historical data, the scene migration is good and the precision is higher. BRIEF DESCRIPTION OF DRAWINGS

[0059] Figure 1 is the overall flow chart of the expressway traffic flow speed prediction method considering the motion state difference between vehicles and the hysteresis effect according to the application;

[0060] Figure 2 is a schematic diagram of the turbulence of a single vehicle to the traffic flow according to the application;

[0061] Figure 3 is a schematic diagram of the single vehicle turbulence distribution value of the road unit according to the application;

[0062] Figure 4 is a schematic diagram of the total turbulence distribution value of the road unit according to the application;

[0063] Figure 5 is a schematic diagram of predicting traffic flow speed. DETAILED DESCRIPTION

[0064] The detailed content of the present application and the technical solutions in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0065] The present application provides a highway traffic flow speed prediction method considering the motion state difference and hysteresis effect between vehicles. In actual application, the real-time trajectory information of vehicles is collected by sensing coils, vehicle-mounted GPS and other detection devices according to the intelligent degree of the highway. The change trend of traffic flow speed is estimated, and traffic control measures are taken in time to prevent and alleviate congestion, thereby improving the operation efficiency and safety of the traffic system, and providing a theoretical basis and model support for highway speed prediction and traffic flow organization and control.

[0066] Referring to Figure 1 The highway traffic flow speed prediction method considering the motion state difference and hysteresis effect between vehicles provided by the present application is composed of the following steps, which will be described in detail below.

[0067] Step one, calculate the disturbance of a single vehicle to traffic flow

[0068] First, obtain the traffic flow density ρ, the speed and displacement of each vehicle in the highway traffic flow speed prediction area. The length of the highway traffic flow speed prediction area is not less than 150 meters, and the width is all the lanes in the one-way driving direction. The traffic flow density ρ is the average number of vehicles per kilometer in the highway traffic flow speed prediction area. The traffic flow speed is calculated as follows:

[0069]

[0070] is the traffic flow speed, that is, the average speed of all vehicles in the highway traffic flow speed prediction area; N is the total number of vehicles in the highway traffic flow speed prediction area; i is the vehicle number; i = 1, 2, …, N; v i is the speed of vehicle i;

[0071] Second, a plane rectangular coordinate system is established, with the intersection point of the lane edge line close to the central reservation and the upstream edge line of the region to be predicted on the speed of the traffic flow of the expressway as the origin, the direction of the speed of the traffic flow as the positive direction of the x-axis, and the direction perpendicular to the direction of the speed of the traffic flow as the direction of the y-axis, wherein the positive direction of the y-axis is obtained by rotating the positive direction of the x-axis by 90 degrees counterclockwise;

[0072] Third, the speed of each vehicle is decomposed along the positive direction of the x-axis and the positive direction of the y-axis to obtain the component speed v i cosθ i of the speed of the vehicle along the direction of the speed of the traffic flow and the component speed v i sinθ i of the speed of the vehicle perpendicular to the direction of the speed of the traffic flow, wherein θ i is the included angle between the direction of the speed of the vehicle i and the direction of the speed of the traffic flow, and 0≤θ i ≤π;

[0073] Fourth, the component speed difference between each vehicle is calculated:

[0074] △v xij =v i cosθ i -v j cosθ j

[0075] △v xij is the component speed difference between the vehicle i and the vehicle j along the x direction; v j cosθ j is the component speed of the vehicle j along the positive direction of the x-axis, i is the vehicle number; j=1, 2, …, N;

[0076] △v yij =v i sinθ i -v j sinθ j

[0077] △v yij is the component speed difference between the vehicle i and the vehicle j along the y direction; v j sinθ j is the component speed of the vehicle j along the positive direction of the y-axis;

[0078] Fifth, the component displacement difference between each vehicle is calculated:

[0079] △x ij =x i -x j

[0080] △x ij is the displacement horizontal coordinate difference between the vehicle i and the vehicle j; x i is the displacement horizontal coordinate of the vehicle i; xj is the displacement horizontal coordinate of vehicle j;

[0081] △y ij is the displacement horizontal coordinate of vehicle i; i -y j

[0082] △y ij is the displacement vertical coordinate difference between vehicle i and vehicle j; y i is the displacement vertical coordinate of vehicle i; y j is the displacement vertical coordinate of vehicle j;

[0083] Sixth, calculate the average vehicle spacing

[0084]

[0085] L is the length of the region to be predicted in the highway traffic flow speed; M is the number of lanes;

[0086] Seventh, use the turbulence to express the difference in the motion state of the vehicles in the traffic flow, and calculate the turbulence E generated by a single vehicle to the traffic flow i

[0087]

[0088] wherein, when the speed and direction of all vehicles in the traffic flow are the same, the turbulence generated by a single vehicle to the traffic flow is 0;

[0089] Step two, distribute the turbulence generated by a single vehicle to the traffic flow to the road unit

[0090] Divide the region to be predicted in the highway traffic flow speed into several equal rectangular road units, and distribute the turbulence E generated by vehicle i to the traffic flow to each road unit, and calculate the single vehicle turbulence distribution value E of the road unit i i,p,q

[0091]

[0092] p is the horizontal coordinate of the geometric center of the road unit; q is the vertical coordinate of the geometric center of the road unit; L i,p,q is the distance from the geometric center of vehicle i to the geometric center (p, q) of the road unit;

[0093] Sum the single vehicle turbulence distribution values of each vehicle to the road unit, and calculate the total turbulence distribution value E of the road unit p,q

[0094]

[0095] Step three, calculate the total turbulence of all road units​​​​

[0096] The total disturbance E of the highway traffic flow speed to be predicted region is calculated by summing up the disturbance assignment total value of all road units a :

[0097]

[0098] g is the road unit number, g = 1, 2, …, G; G is the total number of road units;

[0099] Step four, constructing a hysteresis effect model

[0100] The hysteresis effect theory is used to express the functional relationship between the traffic flow speed and the total disturbance of the highway traffic flow speed to be predicted region, and a hysteresis effect model is established:

[0101]

[0102] Wherein and are the traffic flow speeds at time t, time t-1, time t-2 and time t-m, respectively, wherein the time interval between time t and time t-1 is the time length of one hysteresis period, and the time length of one hysteresis period is not greater than 2 seconds and not less than 0.1 second; E a (t-l), E a (t-l-1), E a (t-l-2) and E a (t-l-2) are the total disturbances of the highway traffic flow speed to be predicted region at time t-l, time t-l-1, time t-l-2 and time t-l-s, respectively; l is the initial hysteresis period number of the total disturbance of the highway traffic flow speed to be predicted region, l ∈ N; s+l is the maximum hysteresis period number of the total disturbance of the highway traffic flow speed to be predicted region, s ∈ N + ; m is the maximum hysteresis period number of the traffic flow speed, m ∈ N + ; α is the constant term of the hysteresis effect model; β c is the coefficient, c = 0, 1, 2, …, s; γ d is the coefficient, d = 1, 2, …, m; u(t) is a random error;

[0103] Step five, determining the hysteresis period number and the coefficient of the hysteresis effect model

[0104] The Akaike information criterion is used to construct an objective function:

[0105]

[0106] Wherein, AIC is the objective function; SSR is the residual sum of squares; n is the sample size, that is, the total number of time points contained in the time period input into the model;

[0107] Calculate the target function value under different combinations of s, l and m values, and obtain the optimal s, l and m values corresponding to the minimum value of the target function AIC;

[0108] The constant term and coefficient of the lag effect model are estimated by using the stepwise regression method to obtain α, β c , γ d and u(t);

[0109] Step six, predicting traffic flow speed

[0110] 6-1) The total turbulence of the highway traffic flow speed to be predicted region and the traffic flow speed are taken as the input of the lag effect model, wherein the corresponding time of the first and last terms of the total turbulence time series of the highway traffic flow speed to be predicted region is t-l period and t-l-s period, and the corresponding time of the first and last terms of the traffic flow speed time series is t-1 period and t-m period; the predicted value of the traffic flow speed at t+1 period is calculated as the output of the model, that is

[0111] 6-2) Take as the new model input, update the first and last terms of the total turbulence time series of the highway traffic flow speed to be predicted region and the traffic flow speed time series, then the corresponding time of the first and last terms of the total turbulence time series of the highway traffic flow speed to be predicted region is t-l+1 period and t-l-s+1 period, and the corresponding time of the first and last terms of the traffic flow speed time series is t and t-m+1 period; calculate the output of the model to obtain the predicted value of the traffic flow speed at t+2 period

[0112] 6-3) Repeat the iteration according to the steps of 6-1 and 6-2 to calculate the traffic flow speed at each future time to obtain the time series of the predicted value of the traffic flow speed.

[0113] Embodiment

[0114] The following describes the implementation process and test results of the highway traffic flow speed prediction method considering the motion state difference and lag effect of vehicles according to the present application, but the protection scope of the present application is not limited to the following embodiments.

[0115] The present application selects the Shenyang-Siping section of the Jingha Expressway in Liaoning Province for example analysis. The main line design speed of the expressway is 120 km / h, with 8 lanes in both directions, and the conditions for obtaining vehicle running information are met. The hourly distribution of daily average actual traffic volume is taken as the traffic running condition to construct a simulation scenario.

[0116] The length of the region to be predicted of the expressway traffic flow speed is set to 300 meters, the driving direction is upward, and the time length of one lag period is 1.0 second. According to steps one to three of the method, the traffic flow speed at each time, the disturbance caused by a single vehicle to the traffic flow, and the total disturbance of all road units are calculated with a step length of 1.0 second, and the time series of the traffic flow speed and the total disturbance with a time span of 3600 seconds and a length of 7200 are obtained.

[0117] According to steps four to five of the method, the lag effect model is constructed and the number of lag periods and the coefficients are calculated, and the results are shown in Table 1.

[0118] Table 1 Number of lag periods and coefficients of lag effect model

[0119]

[0120]

[0121] From Table 1, the initial number of lag periods and the maximum number of lag periods of the total disturbance of the region to be predicted of the expressway traffic flow speed are 6 and 11 respectively, the maximum number of lag periods of the traffic flow speed is 2, and the time length of one lag period is 1.0 second, indicating that the traffic flow speed value at the current time is not only affected by the traffic flow speed values 1.0 seconds and 2.0 seconds ago, but also affected by the total disturbances 6.0 seconds, 7.0 seconds, 8.0 seconds, 9.0 seconds, 10.0 seconds, and 11.0 seconds ago. Moreover, the coefficients of each parameter in Table 1 are the influence degrees of the total disturbances at different times and the traffic flow speed on the traffic flow speed at the current time.

[0122] According to step six of the method, the first 20 minutes of the traffic flow speed time series and the first 20 minutes of the total disturbance time series are used as the input of the lag effect model, and the traffic flow speeds at 20 minutes 01 second, 20 minutes 02 second, …, 59 minutes 59 second are iteratively calculated with a step length of 1.0 second, and the time series 1 of the traffic flow speed prediction value with a total length of 2399 seconds is output.

[0123] The time series 1 of the traffic flow speed prediction value is compared with the time series (the last 40 minutes) of the traffic flow speed true value of the simulation output, and the prediction root mean square percentage error is 3.47%, which proves the effectiveness of the method. In addition, in order to verify that the method has good prediction effect without a large amount of historical data, an autoregressive integrated moving average model (ARIMA) is used to predict the traffic flow speed in this scenario. Taking time as the independent variable and traffic flow speed as the dependent variable, after passing through the stationarity and ADF test, the traffic flow speed time series of the first 20 minutes is input for model training with a step of 1.0 second, and the traffic flow speed prediction value time series 2 of the last 40 minutes is output. The two traffic flow speed prediction value time series are compared with the true value, and the error absolute value distribution result is shown in Table 2.

[0124] Table 2 Error absolute value distribution of two traffic flow speed prediction value time series

[0125]

[0126] From Table 2, for the traffic flow speed prediction value time series 1 obtained by using the method, the proportion of time points with error absolute value of [0, 0.5) m / s is 26.7% of the time series 1, and the proportion of time points with error absolute value of [0, 1) m / s is 73.2% of the whole time series; and for the traffic flow speed prediction value time series 2 obtained by using ARIMA, the proportion of time points with error absolute value of [0, 0.5) m / s is 9.2% of the time series 2, and the proportion of time points with error absolute value of [0, 1) m / s is 46.4% of the whole time series. It is shown that the prediction accuracy of the method is higher than that of the traffic flow speed prediction method based on historical data.

Claims

1.A method for freeway traffic flow speed prediction considering inter-vehicle motion state difference and hysteresis effect, characterized in that, The method comprises the following steps: Step one, calculating the disturbance of a single vehicle to the traffic flow First, obtaining the traffic flow density ρ, the speed and displacement of each vehicle in the highway traffic flow speed prediction area, wherein the length of the highway traffic flow speed prediction area is not less than 150 meters, the width is all lanes in the one-way driving direction, the traffic flow density ρ is the average number of vehicles per kilometer in the highway traffic flow speed prediction area; calculating the traffic flow speed: is the traffic flow speed, i.e. the average speed of all vehicles in the region for which the traffic flow speed is to be predicted; N is the total number of vehicles in the region for which the traffic flow speed is to be predicted; i is the vehicle number; i = 1, 2,..., N; v i is the speed of vehicle i; Second, establishing a plane rectangular coordinate system, taking the intersection of the lane edge line close to the central partition and the upstream edge line of the highway traffic flow speed prediction area as the origin, setting the traffic flow speed direction as the positive direction of the x-axis, and the direction perpendicular to the traffic flow speed direction as the y-axis direction, wherein the counterclockwise rotation of 90 degrees along the positive direction of the x-axis is the positive direction of the y-axis; Third, the speed of each vehicle is decomposed along the positive direction of the x-axis and the positive direction of the y-axis to obtain the component speed v along the traffic flow speed direction i cosθ i and the component speed v i sinθ i , where θ i is the included angle between the speed direction of the vehicle i and the traffic flow speed direction, 0≤θ i ≤π Fourth, calculating the speed difference between each vehicle: Δv xij = v i cos θ i - v j cos θ j Δv xij is the relative speed difference between vehicle i and vehicle j in the x direction; v j cos θ j is the component speed of vehicle j in the positive x direction, i is the vehicle number; j = 1, 2, …, N; Δv yij = v i sin θ i - v j sin θ j Δv yij is the relative speed difference between vehicle i and vehicle j in the y direction; v j sin θ j is the component speed of vehicle j in the positive y direction; Fifth, calculating the displacement difference between each vehicle: Δx ij = x i - x j Δx ij is the displacement x-coordinate difference between vehicle i and vehicle j; x i is the displacement x-coordinate of vehicle i; x j is the displacement x-coordinate of vehicle j; Δy ij = y i - y j Ay ij is the difference in longitudinal displacement coordinate between vehicle i and vehicle j; y i is the longitudinal displacement coordinate of vehicle i; y j is the longitudinal displacement coordinate of vehicle j; Sixth, calculate average vehicle spacing L is the length of the highway traffic flow speed prediction area; M is the number of lanes; Seventh, the difference of motion state between vehicles in traffic flow is used to express the turbulence, and the turbulence E generated by a single vehicle to traffic flow is calculated i : Wherein, it is set that when the speed and direction of all vehicles in the traffic flow are the same, the disturbance of a single vehicle to the traffic flow is 0; Step two, distributing the disturbance of a single vehicle to the traffic flow to road units The highway traffic flow speed region to be predicted is divided into several equal rectangular road units, and the disturbance E generated by vehicle i to the traffic flow is allocated to each road unit, and the single vehicle disturbance allocation value E of the road unit is calculated i : E = 1 + 0.5 * (V - V i,p,q : p is the horizontal coordinate of the road element geometric center; q is the vertical coordinate of the road element geometric center; L i,p,q is the distance from the vehicle i geometric center to the road element geometric center (p, q); The single vehicle disorder assignment value of each vehicle to the road unit is summed up to calculate the total disorder assignment value E of the road unit p,q ; Step three, calculating the total disturbance of all road units The total disturbance E of the region to be predicted of the speed of the traffic flow on the motorway is calculated by summing the disturbance assigned to all the road elements a : g is the road unit number, g = 1, 2, …, G; G is the total number of road units; Step four, constructing a hysteresis effect model The hysteresis effect theory is used to represent the functional relationship between the traffic flow speed and the total disturbance of the highway traffic flow speed prediction area, and a hysteresis effect model is established: wherein and are the traffic flow speeds at time t, time t-1, time t-2 and time t-m, respectively, wherein the time interval between time t and time t-1 is the time length of one lag period, and the time length of one lag period is not greater than 2 seconds and not less than 0.1 second; E a (t-l), E a (t-l-1), E a (t-l-2) and E a (t-l-2) are the total turbulences of the expressway traffic flow speed region to be predicted at time t-l, time t-l-1, time t-l-2 and time t-l-s, respectively; l is the initial lag period number of the total turbulence of the expressway traffic flow speed region to be predicted, l ∈ N; s is the maximum lag period number of the total turbulence of the expressway traffic flow speed region to be predicted, s ∈ N + ; m is the maximum lag period number of the traffic flow speed, m ∈ N + ; α is a constant term of the lag effect model; β c is a coefficient, c = 0, 1, 2, …, s; γ d is a coefficient, d = 1, 2, …, m; and u(t) is a random error. Step five, determining the hysteresis period number and coefficient of the hysteresis effect model The Akaike information criterion is used to construct an objective function: Wherein, AIC is the objective function; SSR is the residual sum of squares; n is the sample size, that is, the total number of time points contained in the time period input into the model; Calculate the objective function value under different s, l and m value combinations to obtain the best s, l and m values corresponding to the minimum value of the objective function AIC; The constant term and the coefficients of the lag effect model are estimated by using the stepwise regression method to obtain α, β c , γ d and u(t); Step six, predicting the traffic flow speed 6-1) the total turbulence of the highway traffic flow speed in the region to be predicted is input into the traffic flow speed as a lag effect model, wherein the corresponding time of the first and last items of the time series of the total turbulence of the highway traffic flow speed in the region to be predicted is t-l and t-l-s, respectively, and the corresponding time of the first and last items of the time series of the traffic flow speed is t-1 and t-m, respectively; the predicted value of the traffic flow speed at the time of t+1 is calculated as the output of the model, that is 6-2) As new model inputs, the first and last terms of the total turbulence time series and the traffic flow speed time series of the area to be predicted for highway traffic flow speed are updated. The corresponding times for the first and last terms of the total turbulence time series of the area to be predicted for highway traffic flow speed are then t-l+1 and tl-s+1, respectively, and the corresponding times for the first and last terms of the traffic flow speed time series are t and t-m+1, respectively. The model output is then calculated to obtain the predicted traffic flow speed at time t+2. 6-3) According to the steps of 6-1 and 6-2, the traffic flow speed at each future time is calculated to obtain the time series of traffic flow speed prediction values.

Images