Lane-level congestion prediction method for urban intelligent network connection mixed traffic flow

Abstract

The invention discloses a lane-level congestion prediction method for urban intelligent network connection mixed traffic flow. The method comprises six steps. The method comprises the following steps: acquiring the instantaneous displacement of an intelligent network connection vehicle and the instantaneous speed of the intelligent network connection vehicle, measuring and calculating the instantaneous displacement of a non-intelligent network connection vehicle and the instantaneous speed of the non-intelligent network connection vehicle, calculating the average speed of the mixed traffic flow on a target lane, searching the time sequence feature points of the average speed of the mixed traffic flow on the target lane, and calculating the average speed of the mixed traffic flow on the target lane. Calculating the total disorder of the mixed traffic flow on the target lane, searching the time sequence feature points of the total disorder of the mixed traffic flow, and predicting the end moment of the single congestion change of the target lane and the average speed of the mixed traffic flow at the moment. According to the method, the influence of microscopic vehicle motion state changes of all lanes of a road section on the speed and congestion evolution of a target lane is expressed through modeling, collected real-time traffic flow data serves as model input, prediction results of the congestion trend and the traffic flow speed are output, and therefore the purpose of preventing and relieving congestion is achieved.

Description

Technical Field

[0001] The present invention relates to the field of lane-level prediction of urban traffic flow, and in particular to a lane-level congestion prediction method for urban intelligent networked mixed traffic flow. Background Art

[0002] With the rapid development of intelligent networking and autonomous driving technologies, mixed traffic flows consisting of conventional vehicles and intelligent networked vehicles will exist for a long time in the future. The multi-source perception fusion of C-V2X vehicle-mounted terminal equipment and intelligent roadside equipment provides support for accurate and real-time acquisition of vehicle operation information. On the basis of comprehensive collection of urban mixed traffic flow operation data, traffic flow speed prediction based on historical traffic information and real-time conditions and determination of future changes in traffic operation status are of great significance for achieving accurate control of urban traffic conditions, timely prevention and rapid relief of congestion, and improvement of road service levels and operating efficiency.

[0003] Most of the existing congestion trend analysis and lane-level traffic flow prediction technologies use time series analysis, probability statistics methods and machine learning algorithms to capture the uncertainty and complex characteristics of the time series changes of traffic flow parameters to perform traffic state deduction modeling, thereby predicting traffic flow in specific scenarios. For example, Patent 202010262372.7 uses LSTM to predict the correlation between different lane sections on the same road section; Patent 202110114317.8 improves the accuracy of traffic flow prediction to the lane level by building a neural network with an embedded spatiotemporal attention module; Patent 202410397699.3 uses machine learning methods to model and predict the conflict risks of each lane in the future; Patent 201710065260.0 introduces vehicle risk thresholds and considers their impact on driving behavior, realizing the prediction of lane-level queue length and average speed of traffic flow. However, the existing research technology ignores the hysteresis effect between the spatiotemporal distribution of microscopic vehicle motion differences and traffic state changes, and the machine learning-based technology requires a large amount of historical data for traffic flow parameter feature extraction and model training. In addition, in practical applications, in addition to predicting the average speed of traffic flow, it is also indispensable to predict the key time nodes of changes in congestion trends.

[0004] Based on the above background, there is an urgent need to design a lane-level congestion prediction method for intelligent connected mixed traffic flow, to model the impact of changes in the motion state of microscopic vehicles in each lane of the road section on the speed and congestion evolution process of the target lane, to use the collected real-time traffic flow data as the model input, and to output the prediction results of congestion trend and average traffic flow speed, so as to achieve the purpose of preventing and alleviating congestion. Summary of the invention

[0005] The technical problem to be solved by the present invention is to propose a lane-level congestion prediction method for urban intelligent networked mixed traffic flow in view of the shortcomings of the existing technology, so as to achieve accurate prediction of the start time of congestion change and the average speed of the mixed traffic flow at that time, and provide theoretical and technical support for lane-level prediction of urban traffic flow and traffic flow organization and control.

[0006] The present invention is implemented by adopting the following technical solutions, which are described as follows:

[0007] Step 1: Obtain the instantaneous displacement and instantaneous speed of the intelligent networked vehicle;

[0008] A coordinate system is established with the intersection of the stop line of the downstream intersection adjacent to the data collection area and the center line of the road as the origin, the driving direction for congestion prediction as the positive direction of the X-axis, and the positive direction of the Y-axis rotated 90 degrees counterclockwise along the positive direction of the X-axis;

[0009] Collected by vehicle or road test equipment and t is the data acquisition time, a is the number of the intelligent connected vehicle passing through the data collection area at time t, a = 1, 2, ..., m(t), m(t) is the number of intelligent connected vehicles passing through the data collection area at time t, is the instantaneous displacement of the intelligent connected vehicle a at time t, which includes two parts: is the instantaneous displacement abscissa of the intelligent connected vehicle a at time t, is the instantaneous displacement ordinate of the intelligent connected vehicle a at time t, is the instantaneous speed of the intelligent connected vehicle a at time t;

[0010] Step 2: Calculate the instantaneous displacement and instantaneous speed of the non-intelligent networked vehicle;

[0011] b is the number of non-intelligent connected vehicles passing through the data collection area at time t, b = 1, 2, ..., n(t), n(t) is the number of non-intelligent connected vehicles passing through the data collection area at time t, is the instantaneous displacement of the non-intelligent connected vehicle b at time t, which includes two parts: is the instantaneous displacement abscissa of the non-intelligent connected vehicle b at time t, is the instantaneous displacement ordinate of the non-intelligent connected vehicle b at time t, is the instantaneous speed of the non-intelligent connected vehicle b at time t;

[0012] For non-intelligent connected vehicles within the sensing range of intelligent connected vehicles, the data collected by vehicle-mounted or road test equipment and For non-intelligent connected vehicles that are not within the perception range of intelligent connected vehicles, the instantaneous displacement of the intelligent connected vehicles in front and behind and the instantaneous speed of the intelligent connected vehicles in front and behind are estimated. and

[0013]

[0014] in, For Functions related to q(t) and Q(t), For Functions related to q(t) and Q(t), for q(t), Q(t), q′(t) and Q′(t); at time t, the number of the adjacent intelligent connected vehicles before and after the non-intelligent connected vehicle numbered b is a f with a r The numbers of the non-intelligent connected vehicle numbered b and the adjacent intelligent connected vehicles on the left and right are a and f ′ and a r ′; At time t, along the X-axis, the intelligent connected vehicle a f With intelligent connected vehicles r The number of non-intelligent connected vehicles between is Q(t), and non-intelligent connected vehicle b is the number of intelligent connected vehicles a. r The q(t)th non-intelligent connected vehicle in the neighborhood; at time t, along the Y-axis, the intelligent connected vehicle a f The number of non-intelligent connected vehicles between ′ and the intelligent connected vehicle ar′ is Q′(t), and the non-intelligent connected vehicle b is the q′(t)th non-intelligent connected vehicle that is the closest neighbor of the intelligent connected vehicle ar′; is the intelligent connected vehicle a at time t r The instantaneous speed, is the intelligent connected vehicle a at time t r The instantaneous displacement abscissa, is the intelligent connected vehicle a at time t r The instantaneous displacement ordinate of is the intelligent connected vehicle a at time t f The instantaneous speed, is the intelligent connected vehicle a at time t f The instantaneous displacement abscissa, is the intelligent connected vehicle a at time t f The instantaneous displacement ordinate, is the intelligent connected vehicle a at time t r ''s instantaneous speed, is the intelligent connected vehicle a at time t f ′’s instantaneous speed;

[0015] Step 3: Calculate the average speed of mixed traffic flow on the target lane

[0016] is the average of the instantaneous speeds of all intelligent connected vehicles and all non-intelligent connected vehicles on the target lane at time t:

[0017]

[0018] A(t) is the number of mixed traffic flow vehicles on the target lane at time t, that is, the sum of all intelligent connected vehicles m′(t) and all non-intelligent connected vehicles n′(t) on the target lane, A(t) = m′(t) + n′(t);

[0019] Step 4: Search for the time series characteristic points of the average speed of the mixed traffic flow on the target lane;

[0020] There are 4 types of time series feature points of average speed of mixed traffic flow, which are defined as the starting point of congestion formation, T 1 ′(t), congestion formation end point T 2 ′(t), congestion dissipation starting point T 3 ′(t) and the congestion dissipation endpoint T 4 ′(t); When searching, the characteristic points of the average speed time series of these four types of mixed traffic flows meet the following mathematical characteristics:

[0021] (1) Congestion starting point T 1 ′(t) is The moment when the high-speed fluctuation turns to the maximum point of monotonous decline, where the range of high-speed fluctuation is Not less than 40km / h;

[0022] (2) Congestion formation end point T 2 ′(t) is The moment corresponding to the minimum point from monotonically decreasing to slow fluctuation, where the range of slow fluctuation is No more than 30 km / h;

[0023] (3) Congestion elimination starting point T 3 ′(t) is The moment corresponding to the minimum point where the slow fluctuation turns into a monotonically rising one;

[0024] (4) Congestion relief endpoint T 4 ′(t) is The moment corresponding to the maximum point when the monotonic rise turns to high-speed fluctuation;

[0025] Step 5, calculating the total disorder of the mixed traffic flow on the target lane and searching for the time series characteristic points of the total disorder of the mixed traffic flow;

[0026] The length of the data collection area is L, the number of lanes in the driving direction for congestion prediction is Z, and the lane number to be predicted is z 0 ; Divide the data collection area into Z times K road grids of the same size, and calculate the total disorder E of the mixed traffic flow on the target lane at time t S (t):

[0027]

[0028] Where z is the number of the road grid along the Y axis, z = 1, 2, ..., Z, k is the number of the road grid along the X axis, k = 1, 2, ..., K, K is the total number of road grids on a single lane, E C (z, k, t) is the mixed traffic flow disorder on the road grid (z, k) at time t, E S (z 0 , t) is the total disorder of mixed traffic flow on the target lane at time t, i is the vehicle number, d z,k,i (t) is the distance from the geometric center of vehicle i to the geometric center of the road grid numbered (z, k) at time t, is the average inter-vehicle distance of mixed traffic flow in the data collection area at time t, v i (t)cosδ i With v j (t)cosδ j They are respectively the vehicle i and vehicle j at time t. Component velocity in direction, v i (t)sinδ i With v j (t)sinδ j At time t, vehicle i and vehicle j are perpendicular to The component velocity of the direction, is the time when vehicle i and vehicle j are moving along The displacement difference in the direction, G represents the vehicle type, G∈{I,N}, At time t, vehicle i is perpendicular to vehicle j. Directional displacement difference;

[0029] There are 4 types of time series characteristic points of total disorder of mixed traffic flow, which are the starting point T of total disorder of mixed traffic flow caused by congestion, 1 (t), corresponding to the total disorder of mixed traffic flow caused by congestion, the end point T 2 (t), corresponding to the dissipation of congestion, the total disorder of mixed traffic flow increases starting point T 3 (t) The total disorder increase of mixed traffic flow corresponding to the dissipation of congestion T4 (t); During the search, the characteristic points of the total disorder time series of these four types of mixed traffic flows meet the following mathematical characteristics:

[0030] (1) The total disorder of mixed traffic flow caused by congestion increases starting point T 1 (t) is the congestion formation stage E S (z 0 , t) corresponds to the turning point from the beginning of low value fluctuation to significant increase, where the range of low value fluctuation is Not more than 300;

[0031] (2) The total disorder of mixed traffic flow caused by congestion increases the terminal point T 2 (t) is the congestion formation stage E S (z 0 , t) the moment corresponding to the turning point when the significant increase ends and turns to low-value fluctuation;

[0032] (3) The starting point T of the total disorder increase of mixed traffic flow corresponding to the dissipation of congestion 3 (t) is the congestion dissipation stage E S (z 0 , t) the moment corresponding to the turning point from the beginning of low-value fluctuation to significant increase;

[0033] (4) The total disorder of mixed traffic flow corresponding to the dissipation of congestion increases the end point T 4 (t) is the congestion dissipation stage E S (z 0 , t) the moment corresponding to the turning point when the significant increase ends and turns to low-value fluctuation;

[0034] Step 6: Predict the end time T of a single congestion change in the target lane a ′ and T a Average speed of mixed traffic flow at time ′

[0035]

[0036] E S (z 0 , t) The change start time Tp includes T 1 (t) and T 3 (t); E S (z 0 , t) Change end time T a Including T 2 (t) and T 4 (t); The time when the target lane congestion changes and starts T p 'Including T 1 ′(t) and T 3 ′(t); the target lane congestion change end time Ta 'Including T 2 ′(t) and T 4 ′(t);

[0037] Average speed of mixed traffic flow corresponding to the start time of congestion change Including the average speed of mixed traffic flow corresponding to the starting point of congestion Average speed of mixed traffic flow corresponding to the starting point of congestion dissipation Average speed of mixed traffic flow corresponding to the end of congestion change Including the average speed of mixed traffic flow corresponding to the congestion formation end point Average speed of mixed traffic flow corresponding to the end point of congestion dissipation

[0038] By observing T p , T p ′、T a , And calculate each time E S (t) Solution T 2 ′(t) and T 4 ′(t):

[0039]

[0040]

[0041] where c w With c w ′ is the polynomial coefficient, w=1, 2, 3, 4; △E S (z 0 , t) is E at time t S (z 0 , t) compared to the change at the previous moment, △E S (z 0 , t) = E S (z 0 , t)-E S (z 0 , t-1); s is the average speed of mixed traffic flow in the monotonic interval [Tp′, T a ′].

[0042] Furthermore, the data collection area in step 1 is not less than 50 meters in length, and is as wide as all Z lanes in the driving direction for which congestion prediction is to be performed. The time interval between the data acquisition time t and the previous time t-1 does not exceed 2 seconds.

[0043] Further, in the step 2, according to Calculation of q(t), Q(t), q′(t) and Q′(t) and The specific method is:

[0044]

[0045] Furthermore, s in step 6 is approximately equal to the average speed of the mixed traffic flow in the monotonic interval [T p ′,T a ], that is,

[0046] Compared with the prior art, the lane-level congestion prediction method for urban intelligent networked mixed traffic flow described in the present invention has the following beneficial effects:

[0047] (1) The method takes into account the impact of the spatiotemporal transmission of the differences in the motion states of microscopic vehicles on the development process of macroscopic traffic congestion, and establishes a polynomial relationship model between the average speed time series of single-lane mixed traffic flow and the total disorder time series of single-lane mixed traffic flow. The model input is the real-time trajectory information of intelligent networked vehicles and intelligent networked vehicles obtained by collection and measurement. No large amount of historical traffic data is required for model training or calibration. The prediction of the end time of congestion formation and the end time of congestion dissipation is achieved, and the predicted value of the average speed of traffic flow at the end time of congestion change is output, which improves the prediction accuracy to the lane level, providing decision-making basis and technical support for vehicle path induction, traffic flow organization and control, and congestion prevention and evacuation;

[0048] (2) The method collects the instantaneous speed and displacement of intelligent connected vehicles and non-intelligent connected vehicles within the perception range of intelligent connected vehicles through road test equipment and on-board equipment of intelligent connected vehicles (including but not limited to cameras, millimeter wave radars, laser radars, ultrasonic sensors, etc.); for non-intelligent connected vehicles that are not within the perception range of intelligent connected vehicles, the instantaneous displacement and instantaneous speed of the non-intelligent connected vehicles are estimated based on their relative position relationship with the intelligent connected vehicles in front and behind, thereby realizing real-time collection of mixed traffic flow operation parameters without meeting the requirement of high road detector coverage, thereby reducing data acquisition costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is an overall flow chart of the lane-level congestion prediction method for urban intelligent networked mixed traffic flow described in the present invention;

[0050] Figure 2 is a schematic diagram of the coordinate system of the present invention;

[0051] Figure 3 It is a schematic diagram of measuring the instantaneous displacement and the instantaneous speed of a non-intelligent networked vehicle according to the present invention;

[0052] Figure 4 is a schematic diagram of calculating the total disorder of mixed traffic flow on the target lane according to the present invention;

[0053] Figure 5 It is a schematic diagram of the time series characteristic points of the average speed of mixed traffic flow on the target lane and the time series characteristic points of the average speed of mixed traffic flow on the target lane according to the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the details of the present invention and the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0055] See also Figure 1 The lane-level congestion prediction method for urban intelligent networked mixed traffic flow described in the present invention comprises the following steps:

[0056] Step 1: Obtain the instantaneous displacement and instantaneous speed of the intelligent networked vehicle;

[0057] Step 2: Calculate the instantaneous displacement and instantaneous speed of the non-intelligent networked vehicle;

[0058] Step 3: Calculate the average speed v(t) of the mixed traffic flow on the target lane;

[0059] Step 4: Search for the time series characteristic points of the average speed of the mixed traffic flow on the target lane;

[0060] Step 5, calculating the total disorder of the mixed traffic flow on the target lane and searching for the time series characteristic points of the total disorder of the mixed traffic flow;

[0061] Step 6: Predict the end time T of a single congestion change in the target lane a ′ and T a Average speed of mixed traffic flow at time ′

[0062]

[0063] The above steps are discussed in conjunction with an embodiment of the present invention:

[0064] According to step 1, on the urban roads (main roads, secondary roads and branch roads) where congestion prediction is to be carried out, a one-way driving direction is set as the driving direction for which congestion prediction is to be carried out, and a coordinate system is established with the intersection of the stop line of the downstream intersection adjacent to the data collection area and the center line of the road as the origin, the driving direction for which congestion prediction is to be carried out as the positive direction of the X-axis, and the positive direction of the Y-axis rotated 90 degrees counterclockwise along the positive direction of the X-axis, such as Figure 2 As shown;

[0065] Collected by on-board sensors of intelligent connected vehicles or road test equipment (road induction coils, bayonet cameras, ultrasonic monitors) and t is the data acquisition time, a is the number of the intelligent connected vehicle passing through the data collection area at time t, a = 1, 2, ..., m(t), m(t) is the number of intelligent connected vehicles passing through the data collection area at time t, is the instantaneous displacement of the intelligent connected vehicle a at time t, which includes two parts: is the instantaneous displacement abscissa of the intelligent connected vehicle a at time t, is the instantaneous displacement ordinate of the intelligent connected vehicle a at time t, is the instantaneous speed of the intelligent connected vehicle a at time t;

[0066] The data collection area in step 1 is not less than 50 meters in length and all Z lanes in the driving direction for which congestion prediction is to be performed. The time interval between the data acquisition time t and the previous time t-1 is not more than 2 seconds. In one embodiment of the present invention, the data collection area is 150 meters in length and 3 lanes in width in one direction. The road type is a secondary trunk road. The time interval between the data acquisition time t and the previous time t-1 is 1 second. Each subsequent data acquisition time is recorded with 16:30 as the starting point of the data acquisition time and 18:30 as the end point of the data acquisition time (16:30:01 corresponds to the first data acquisition time, 17:00:00 corresponds to the 1800th data acquisition time, 17:30:00 corresponds to 3600s, 18:30:00 corresponds to 5400s... and so on).

[0067] According to step 2, the instantaneous displacement and instantaneous speed of the non-intelligent connected vehicle are calculated; for non-intelligent connected vehicles within the sensor and network sensing range of the intelligent connected vehicle, the intelligent connected vehicle on-board sensor or road test equipment is used to collect and b is the number of non-intelligent connected vehicles passing through the data collection area at time t, b = 1, 2, ..., n(t), n(t) is the number of non-intelligent connected vehicles passing through the data collection area at time t, is the instantaneous displacement of the non-intelligent connected vehicle b at time t, which includes two parts: is the instantaneous displacement abscissa of the non-intelligent connected vehicle b at time t, is the instantaneous displacement ordinate of the non-intelligent connected vehicle b at time t, is the instantaneous speed of the non-intelligent connected vehicle b at time t;

[0068] See also Figure 3 For non-intelligent connected vehicles that are not within the sensing range of the intelligent connected vehicles, the instantaneous displacement of the intelligent connected vehicles in front and behind and the instantaneous speed of the intelligent connected vehicles in front and behind are estimated. and

[0069]

[0070] in, For Functions related to q(t) and Q(t), For Functions related to q(t) and Q(t), for q(t), Q(t), q′(t) and Q′(t); at time t, the number of the adjacent intelligent connected vehicles before and after the non-intelligent connected vehicle numbered b is a f with a r The numbers of the non-intelligent connected vehicle numbered b and the adjacent intelligent connected vehicles on the left and right are a and f ′ and a r ′; At time t, along the X-axis, the intelligent connected vehicle a f With intelligent connected vehicles r The number of non-intelligent connected vehicles between is Q(t), and non-intelligent connected vehicle b is the number of intelligent connected vehicles a. r The q(t)th non-intelligent connected vehicle in the neighborhood; at time t, along the Y-axis direction, the intelligent connected vehicle af′ and the intelligent connected vehicle a r ′ is Q′(t), and non-intelligent connected vehicle b is the number of connected vehicles between intelligent connected vehicle a and r q′(t)th non-intelligent connected vehicle in the ′ neighbor; and They are the intelligent connected vehicles a at time t. r with a r ''s instantaneous speed, is the intelligent connected vehicle a at time t r The instantaneous displacement abscissa, is the intelligent connected vehicle a at time t r The instantaneous displacement ordinate of and They are the intelligent connected vehicles a at time t. f with af ''s instantaneous speed, is the intelligent connected vehicle a at time t f The instantaneous displacement abscissa, is the intelligent connected vehicle a at time t f The instantaneous displacement ordinate of

[0071] In this embodiment, further, according to the step 2 q(t), Q(t), q′(t) and Q′(t), calculation and

[0072]

[0073]

[0074] Then, in step 3, the average speed of mixed traffic flow on the target lane is calculated. is the average of the instantaneous speeds of all intelligent connected vehicles and all non-intelligent connected vehicles on the target lane at time t:

[0075]

[0076] A(t) is the number of mixed traffic flow vehicles on the target lane at time t, that is, the sum of all intelligent connected vehicles m′(t) and all non-intelligent connected vehicles n′(t) on the target lane, A(t)=m′(t)+n′(t); the target lane in this embodiment is the second lane from the center line of the road to the roadside, that is, the middle lane of the three lanes in the driving direction for which congestion prediction is to be performed.

[0077] From step 4, search for the time series feature points of the average speed of mixed traffic flow on the target lane: there are 4 types of time series feature points of the average speed of mixed traffic flow, which are respectively defined as the starting point of congestion formation T 1 ′(t), congestion formation end point T 2 ′(t), congestion dissipation starting point T 3 ′(t) and the congestion dissipation endpoint T 4 ′(t);

[0078] When searching, the characteristic points of the average speed time series of these four types of mixed traffic flows meet the following mathematical characteristics:

[0079] (1) Congestion starting point T 1 ′(t) is The moment when the high-speed fluctuation turns to the maximum point of monotonous decline, where the range of high-speed fluctuation is Not less than 40km / h;

[0080] (2) Congestion formation end point T 2 ′(t) is The moment corresponding to the minimum point from monotonically decreasing to slow fluctuation, where the range of slow fluctuation is No more than 30 km / h;

[0081] (3) Congestion elimination starting point T 3 ′(t) is The moment corresponding to the minimum point where the slow fluctuation turns into a monotonically rising one;

[0082] (4) Congestion relief endpoint T 4 ′(t) is The moment corresponding to the maximum point when the monotonic rise turns to high-speed fluctuation;

[0083] The average speed time series of mixed traffic flow on a target lane covers the number of congestion times and various characteristic points. The complete process of a single congestion starts to form → the congestion ends → the congestion continues → the congestion begins to dissipate → the congestion completely dissipates covers one characteristic point of each of the four types, and the order of appearance of the characteristic points of the average speed time series of mixed traffic flow is T 1 ′(t), T 2 ′(t), T 3 ′(t) and T 4 ′(t), in this embodiment, if Figure 5 As shown;

[0084] According to step 5, the total disorder of the mixed traffic flow on the target lane is calculated and the time series characteristic points of the total disorder of the mixed traffic flow are searched;

[0085] See also Figure 4 , the length of the data collection area is L, the number of lanes in the driving direction for congestion prediction is Z, and the lane number to be predicted is z 0 ; Divide the data collection area into Z times K road grids of the same size, and calculate the total disorder E of the mixed traffic flow on the target lane at time t S (t):

[0086]

[0087] Where z is the number of the road grid along the Y axis, z = 1, 2, ..., Z, k is the number of the road grid along the X axis, k = 1, 2, ..., K, K is the total number of road grids on a single lane, E C (z, k, t) is the mixed traffic flow disorder on the road grid (z, k) at time t, E S (z 0 , t) is the total disorder of mixed traffic flow on the target lane at time t, i is the vehicle number, d z,k,i(t) is the distance from the geometric center of vehicle i to the geometric center of the road grid numbered (z, k) at time t, is the average inter-vehicle distance of mixed traffic flow in the data collection area at time t, v i (t)cosδ i With v j (t)cosδ j They are respectively the vehicle i and vehicle j at time t. Component velocity in direction, v i (t)sinδ i With v j (t)sinδ j At time t, vehicle i and vehicle j are perpendicular to The component velocity of the direction, is the time when vehicle i and vehicle j are moving along The displacement difference in the direction, G represents the vehicle type, G∈{I,N}, At time t, vehicle i is perpendicular to vehicle j. Directional displacement difference;

[0088] In this embodiment, Z is 3, z 0 is 2, K is 150.

[0089] According to the step 5, there are 4 types of time series characteristic points of total disorder of mixed traffic flow, which are respectively the starting point T of total disorder increase of mixed traffic flow caused by congestion 1 (t), corresponding to the total disorder of mixed traffic flow caused by congestion, the end point T 2 (t), corresponding to the dissipation of congestion, the total disorder of mixed traffic flow increases starting point T 3 (t) The total disorder increase of mixed traffic flow corresponding to the dissipation of congestion T 4 (t); During the search, the characteristic points of the total disorder time series of these four types of mixed traffic flows meet the following mathematical characteristics:

[0090] (1) The total disorder of mixed traffic flow caused by congestion increases starting point T 1 (t) is the congestion formation stage E S (z 0 , t) corresponds to the turning point from the beginning of low value fluctuation to significant increase, where the range of low value fluctuation is Not more than 300;

[0091] (2) The total disorder of mixed traffic flow caused by congestion increases the terminal point T 2 (t) is the congestion formation stage E S (z 0 , t) the moment corresponding to the turning point when the significant increase ends and turns to low-value fluctuation;

[0092] (3) The starting point T of the total disorder increase of mixed traffic flow corresponding to the dissipation of congestion 3 (t) is the congestion dissipation stage E S (z 0 , t) the moment corresponding to the turning point from the beginning of low-value fluctuation to significant increase;

[0093] (4) The total disorder of mixed traffic flow corresponding to the dissipation of congestion increases the end point T 4 (t) is the congestion dissipation stage E S (z 0 , t) the moment corresponding to the turning point when the significant increase ends and turns to low-value fluctuation;

[0094] The total disorder time series of mixed traffic flow in a target lane covers a number of characteristic points of various types. In the complete process of a single congestion starting to form → the end of congestion formation → congestion continuation → congestion beginning to dissipate → congestion completely dissipates, the total disorder time series shows a trend of total disorder increasing → total disorder decreasing → total disorder fluctuating at a low value → total disorder increasing again → total disorder decreasing again. At the same time, this process covers one characteristic point of each of the four types of total disorder time series, and the order of appearance is T 1 (t), T 2 (t), T 3 (t) and T 4 (t); For the same complete congestion process, the various characteristic points of the total disorder time series of the mixed traffic flow in the target lane always appear before the characteristic points of the corresponding categories of the average speed time series of the mixed traffic flow in the target lane. In this embodiment, Figure 5 As shown;

[0095] According to step 6, the end time T of a single congestion change in the target lane is predicted a ′ and T a Average speed of mixed traffic flow at time ′ in:

[0096] E S (z 0 , t) The change start time Tp includes T 1 (t) and T 3 (t); E S (z 0 , t) Change end time T a Including T 2 (t) and T 4 (t); The time when the target lane congestion changes and starts T p 'Including T 1 ′(t) and T 3 ′(t); the target lane congestion change end time T a 'Including T 2 ′(t) and T 4′(t);

[0097] Average speed of mixed traffic flow corresponding to the start time of congestion change Including the average speed of mixed traffic flow corresponding to the starting point of congestion Average speed of mixed traffic flow corresponding to the starting point of congestion dissipation Average speed of mixed traffic flow corresponding to the end of congestion change Including the average speed of mixed traffic flow corresponding to the congestion formation end point Average speed of mixed traffic flow corresponding to the end point of congestion dissipation

[0098] By observing T p , T p ′、T a , And calculate each time E S (t) Solution T 2 ′(t) and T 4 ′(t):

[0099]

[0100] where c w With c w ′ is the polynomial coefficient, w=1, 2, 3, 4; △E S (z 0 , t) is E at time t S (z 0 , t) compared to the change at the previous moment, △E S (z 0 , t) = E S (z 0 , t)-E S (z 0 , t-1); s is the average speed of mixed traffic flow in the monotonic interval [Tp′, T a ′].

[0101] Furthermore, s in step 6 is approximately equal to the average speed of the mixed traffic flow in the monotonic interval [T p ′,T a ], that is,

[0102] In this embodiment, the least square method is used to calibrate the coefficients of the polynomial that represents the relationship between the time series of the average speed of the single-lane mixed traffic flow and the time series of the total disorder of the single-lane mixed traffic flow, and record T p , T p ′、T a and (Unit: m / s) and summarize the observed values ​​of c w With c w The calibration results of ′ are shown in Table 1; and The observed values ​​are 6.79 m / s and 10.41 m / s respectively. According to step 6, the approximate value of s in this embodiment is 7.67×10 -3 With 6.38×10 -3 ; Substitute

[0103] △E S (z 0 , t), according to the above steps, we can obtain T 2 ′(t) and T 4 The prediction results of ′(t) are shown in Table 2.

[0104] Table 1 Observed values ​​of parameters and calibration results of polynomial coefficients

[0105]

[0106] Table 2 Congestion change endpoint and its corresponding mixed traffic flow average speed prediction results

[0107]

[0108] In summary, the present invention proposes a lane-level congestion prediction method for urban intelligent connected mixed traffic flow. In practical application, the real-time trajectory information of vehicles is collected by road test equipment such as road induction coils, bayonet cameras, and ultrasonic monitors according to the intelligence level of urban roads and the penetration rate of intelligent connected vehicles in mixed traffic flows. By predicting the end time of congestion formation, the end time of congestion dissipation, and the corresponding average speed of mixed traffic flows, a theoretical basis and model support are provided for lane-level congestion prediction and traffic flow organization of urban intelligent connected mixed traffic flows, so that traffic control measures can be taken in time to achieve the purpose of preventing and alleviating congestion, thereby improving the operation efficiency and safety of the transportation system.

Claims

1. A lane-level congestion prediction method for urban intelligent networked mixed traffic flow, characterized in that: The following steps are involved: Step 1: Obtain the instantaneous displacement and instantaneous speed of the intelligent networked vehicle; A coordinate system is established with the intersection of the stop line of the downstream intersection adjacent to the data collection area and the center line of the road as the origin, the driving direction for congestion prediction as the positive direction of the X-axis, and the positive direction of the Y-axis rotated 90 degrees counterclockwise along the positive direction of the X-axis; Collected by vehicle or road test equipment and t is the data acquisition time, a is the number of the intelligent connected vehicle passing through the data collection area at time t, a = 1, 2, ..., m(t), m(t) is the number of intelligent connected vehicles passing through the data collection area at time t, is the instantaneous displacement of the intelligent connected vehicle a at time t, which includes two parts, is the instantaneous displacement abscissa of the intelligent connected vehicle a at time t, is the instantaneous displacement ordinate of the intelligent connected vehicle a at time t, is the instantaneous speed of the intelligent connected vehicle a at time t; Step 2: Calculate the instantaneous displacement and instantaneous speed of the non-intelligent networked vehicle; b is the number of non-intelligent connected vehicles passing through the data collection area at time t, b = 1, 2, ..., n(t), n(t) is the number of non-intelligent connected vehicles passing through the data collection area at time t, is the instantaneous displacement of the non-intelligent connected vehicle b at time t, which includes two parts: is the instantaneous displacement abscissa of the non-intelligent connected vehicle b at time t, is the instantaneous displacement ordinate of the non-intelligent connected vehicle b at time t, is the instantaneous speed of the non-intelligent connected vehicle b at time t; For non-intelligent connected vehicles within the sensing range of intelligent connected vehicles, the data collected by vehicle-mounted or road test equipment and For non-intelligent connected vehicles that are not within the perception range of intelligent connected vehicles, the instantaneous displacement of the intelligent connected vehicles in front and behind and the instantaneous speed of the intelligent connected vehicles in front and behind are estimated. and in, For Functions related to q(t) and Q(t), For Functions related to q(t) and Q(t), for q(t), Q(t), q′(t) and Q′(t); at time t, the number of the adjacent intelligent connected vehicles before and after the non-intelligent connected vehicle numbered b is a f with a r The numbers of the non-intelligent connected vehicle numbered b and the adjacent intelligent connected vehicles on the left and right are a and f ′ and a r ′; At time t, along the X-axis, the intelligent connected vehicle a f With intelligent connected vehicles r The number of non-intelligent connected vehicles between is Q(t), and non-intelligent connected vehicle b is the number of intelligent connected vehicles a. r The q(t)th non-intelligent connected vehicle in the neighborhood; at time t, along the Y-axis, the intelligent connected vehicle a f ′ and intelligent connected vehicles a r ′ is Q′(t), and non-intelligent connected vehicle b is the number of connected vehicles between intelligent connected vehicle a and r q′(t)th non-intelligent connected vehicle in the ′ neighbor; is the intelligent connected vehicle a at time t r The instantaneous speed, is the intelligent connected vehicle a at time t r The instantaneous displacement abscissa, is the intelligent connected vehicle a at time t r The instantaneous displacement ordinate of is the intelligent connected vehicle a at time t f The instantaneous speed, is the intelligent connected vehicle a at time t f The instantaneous displacement abscissa, is the intelligent connected vehicle a at time t f The instantaneous displacement ordinate, is the intelligent connected vehicle a at time t r ''s instantaneous speed, is the intelligent connected vehicle a at time t f ′’s instantaneous speed; Step 3: Calculate the average speed of mixed traffic flow on the target lane is the average of the instantaneous speeds of all intelligent connected vehicles and all non-intelligent connected vehicles on the target lane at time t: A(t) is the number of mixed traffic flow vehicles on the target lane at time t, that is, the sum of all intelligent connected vehicles m′(t) and all non-intelligent connected vehicles n′(t) on the target lane, A(t) = m′(t) + n′(t); Step 4: Search for the time series characteristic points of the average speed of the mixed traffic flow on the target lane; There are 4 types of time series feature points of average speed of mixed traffic flow, which are defined as the starting point of congestion formation T1′(t), the end point of congestion formation T2′(t), the starting point of congestion dissipation T3′(t) and the end point of congestion dissipation T4′(t). When searching, these 4 types of time series feature points of average speed of mixed traffic flow meet the following mathematical characteristics: (1) The starting point of congestion formation T1′(t) is The moment when the high-speed fluctuation turns to the maximum point of monotonous decline, where the range of high-speed fluctuation is Not less than 40km / h; (2) The congestion formation end point T2′(t) is The moment corresponding to the minimum point from monotonically decreasing to slow fluctuation, where the range of slow fluctuation is No more than 30 km / h; (3) The starting point of congestion dissipation T3′(t) is The moment corresponding to the minimum point from slow fluctuation to monotonically rising; (4) The congestion dissipation endpoint T4′(t) is The moment corresponding to the maximum point when the monotonic rise turns to high-speed fluctuation; Step 5, calculating the total disorder of the mixed traffic flow on the target lane and searching for the time series characteristic points of the total disorder of the mixed traffic flow; The length of the data collection area is L, the number of lanes in the driving direction for congestion prediction is Z, and the lane number to be predicted is z0; the data collection area is divided into Z times K road grids of the same size, and the total disorder E of the mixed traffic flow on the target lane at time t is calculated. S (t): Where z is the number of the road grid along the Y axis, z = 1, 2, ..., Z, k is the number of the road grid along the X axis, k = 1, 2, ..., K, K is the total number of road grids on a single lane, E C (z, k, t) is the mixed traffic flow disorder on the road grid (z, k) at time t, E S (z0, t) is the total disorder of mixed traffic flow on the target lane at time t, i is the vehicle number, d z,k,i (t) is the distance from the geometric center of vehicle i to the geometric center of the road grid numbered (z, k) at time t, is the average inter-vehicle distance of mixed traffic flow in the data collection area at time t, v i (t)cosδ i With v j (t)cosδ j They are respectively the vehicle i and vehicle j at time t. Component velocity in direction, v i (t)sinδ i With v j (t)sinδ j At time t, vehicle i and vehicle j are perpendicular to The component velocity of the direction, is the time when vehicle i and vehicle j are moving along The displacement difference in the direction, G represents the vehicle type, G∈{I,N}, At time t, vehicle i is perpendicular to vehicle j. Directional displacement difference; There are 4 types of mixed traffic flow total disorder time series feature points, which are the starting point T1(t) of the mixed traffic flow total disorder increase corresponding to congestion, the end point T2(t) of the mixed traffic flow total disorder increase corresponding to congestion, the starting point T3(t) of the mixed traffic flow total disorder increase corresponding to congestion dissipation, and the end point T4(t) of the mixed traffic flow total disorder increase corresponding to congestion dissipation. When searching, these 4 types of mixed traffic flow total disorder time series feature points meet the following mathematical characteristics: (1) The starting point T1(t) of the total disorder increase of mixed traffic flow corresponding to congestion is the congestion formation stage E. S (z0, t) is the time corresponding to the turning point from the beginning of low value fluctuation to significant increase, where the range of low value fluctuation is Not more than 300; (2) The total disorder increase end point T2(t) of the mixed traffic flow corresponding to the congestion is the congestion formation stage E S (z0, t) The moment corresponding to the turning point when the significant increase ends and the fluctuation turns to a low value; (3) The starting point of the total disorder increase of mixed traffic flow corresponding to the congestion dissipation is T3(t), which is the congestion dissipation stage E. S (z0, t) is the time corresponding to the turning point from the beginning of low-value fluctuation to significant increase; (4) The end point T4(t) of the total disorder increase of mixed traffic flow corresponding to the congestion dissipation is the congestion dissipation stage E S (z0, t) The moment corresponding to the turning point when the significant increase ends and the fluctuation turns to a low value; Step 6: Predict the end time T of a single congestion change in the target lane a ′ and T a Average speed of mixed traffic flow at time ′ E S (z0, t) Change start time T p Including T1(t) and T3(t); E S (z0, t) Change end time T a Including T2(t) and T4(t); the target lane congestion change starts at T p ' includes T1'(t) and T3'(t); the target lane congestion change end time T a ' includes T2'(t) and T4'(t); Average speed of mixed traffic flow corresponding to the start time of congestion change Including the average speed of mixed traffic flow corresponding to the starting point of congestion Average speed of mixed traffic flow corresponding to the starting point of congestion dissipation Average speed of mixed traffic flow corresponding to the end of congestion change Including the average speed of mixed traffic flow corresponding to the congestion formation end point Average speed of mixed traffic flow corresponding to the end point of congestion dissipation By observing T p , T p ′、T a , And calculate each time E S (t) Solution T2′(t) and T4′(t): where c w With c w ′ is the polynomial coefficient, w=0,1,2,3,4; △E S (z0, t) is E at time t S The change of (z0, t) compared with the previous moment, △E S (z0, t) = E S (z0,t)-E S (z0, t-1); s is the average speed of mixed traffic flow in the monotonic interval [Tp′, T a ′].

2. A lane-level congestion prediction method for urban intelligent networked mixed traffic flow according to claim 1, characterized in that: The length of the data collection area shall not be less than 50 meters, and the width shall be all Z lanes in the driving direction for which congestion prediction is to be performed. The time interval between the data acquisition time t and the previous time t-1 shall not exceed 2 seconds.

3. A lane-level congestion prediction method for urban intelligent networked mixed traffic flow according to claim 1, characterized in that: according to Calculation of q(t), Q(t), q′(t) and Q′(t) and The specific method is:

4. A lane-level congestion prediction method for urban intelligent networked mixed traffic flow according to claim 1, characterized in that: In the calculation, s is approximately equal to the average speed of the mixed traffic flow in the monotonic interval [T p ′,T a ], that is,

Images