JP2008165332A - Traffic flow simulation mismatching decision device and traffic flow simulation mismatching decision method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a traffic flow simulation mismatching decision device and a traffic flow simulation mismatching decision method for finding out a spot where a simulated traffic flow is mismatched with an actual traffic flow. <P>SOLUTION: A sensor information input part 310 inputs a sensor traffic volume measured by a sensor 140 about a plurality of spots of a road network. A simulation part 330 simulates the traffic volumes of a plurality of spots based on the sensor traffic volume by using the parameter of a traffic flow model. A mismatching decision part 340 searches the regression line of the sensor traffic volume and the traffic volume of a simulated traffic flow, and decides the spot where the simulated traffic flow is mismatched with the actual traffic flow from an error to the regression line. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a traffic flow simulation inconsistency determination device and a traffic flow simulation failure that determine whether or not the simulation result of traffic flow at each point matches the actual traffic flow at the corresponding point for a plurality of points on the road network. The present invention relates to a method for determining consistency.

  For the purpose of effective operation of the road traffic control system, the traffic flow of the road network is simulated. Usually, in the traffic flow simulation, parameters relating to a vehicle travel model such as vehicle acceleration, speed at which dense waves are transmitted, and vehicle start delay are set to simulate vehicle behavior.

  The traveling model needs to approximate the actual traveling state of the vehicle. Therefore, based on the measurement results of an ultrasonic vehicle detector (hereinafter referred to as “sensor”) that detects the vehicle using ultrasonic waves, and the captured image of the camera that captures the road network, the simulation results are obtained from the actual traffic flow. For example, Patent Document 1 discloses a technique for adjusting the parameters of the travel model so as to approximate to the above.

By the way, in the simulation of traffic flow, not only a vehicle running model but also a traffic flow model indicating characteristics of roads and intersections is set. The traffic flow model includes, for example, parameters such as a saturated traffic flow rate [units / h], which is the maximum number of passing vehicles per unit time, a branching rate [%] at an intersection, and a link capacity [units]. In the road network, since many roads and intersections are connected to each other, if there is an inconsistency in the traffic flow model, the influence spreads. Therefore, in order to increase the reproducibility of traffic flow simulation, it is necessary to approximate the traffic flow model to the characteristics of actual roads and intersections over the entire road network.
JP 2000-222669 A

  However, the characteristics of roads and intersections depend on the surrounding environment and the usage pattern of the road network, and are different at each point. Therefore, when the parameter setting of the traffic flow model is inappropriate, the difference between the actual traffic flow and the simulated traffic flow is absorbed even if the parameter of the travel model is adjusted by applying the technique described in Patent Document 1. There is a fear that I can not finish.

  On the other hand, if the simulated traffic flow finds a portion that is not close to the actual traffic flow, measurement and investigation may be performed only on that portion, and the setting value of the traffic flow simulation can be adjusted efficiently.

  The present invention has been made in view of this point, and provides a traffic flow simulation inconsistency determination device and a traffic flow simulation inconsistency determination method for finding a point where a simulated traffic flow is inconsistent with an actual traffic flow. .

  The traffic flow simulation inconsistency determination device of the present invention includes actual measurement information input means for inputting actual measurement values of traffic flow information for a plurality of points on the road network, and the plurality of points when the traffic flow on the road network is simulated. A regression line between a simulated value of traffic flow information for the point and the measured values of the plurality of points is obtained, and a point where the simulated traffic flow is inconsistent with the actual traffic flow is determined based on an error with respect to the regression line. It adopts a configuration comprising an alignment determination means.

  The traffic flow simulation inconsistency determination method of the present invention simulates the actual value of traffic flow information at a plurality of points on the road network and the traffic flow information at the plurality of points when the traffic flow on the road network is simulated. An information input step for inputting a value, a regression line calculating step for calculating a regression line between the measured value and the simulated value for the plurality of points, and an error with respect to the regression line. An inconsistency determining step for determining a point inconsistent with the traffic flow is provided.

  According to the present invention, a point where the simulated traffic flow is inconsistent with the actual traffic flow is found. Thereby, the setting value of the traffic flow model used for the traffic flow simulation can be easily adjusted, and the reproducibility of the traffic flow simulation can be improved.

  Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a system configuration diagram showing a configuration of a traffic flow simulation system using a traffic flow simulation mismatch determination device according to an embodiment of the present invention.

  In FIG. 1, a traffic flow simulation system 100 is a system that simulates the traffic flow of a vehicle 130 traveling on a road 120 and controls the traffic light 110 so that traffic congestion does not occur on the road 120 based on the simulation result. In addition, the traffic flow simulation system 100 is a system that determines a portion that is inconsistent with the actual traffic flow based on the measurement data obtained by the sensor 140 provided above the road 120 with respect to its own simulation result. .

  The traffic flow simulation system 100 includes a signal control device 200, a traffic flow simulation mismatch determination device 300, and a display device 400. The signal control device 200 is connected to the sensor 140, the traffic light 110, and the traffic flow simulation mismatch determination device 300. The display device 400 is connected to the traffic flow simulation mismatch determination device 300.

  The sensor 140 includes the number of vehicles 130 that have passed directly under each predetermined time interval (hereinafter referred to as “traffic volume”), the ratio of the passage time of the vehicle 130 (hereinafter referred to as “occupation rate”), The average speed is measured and transmitted to the signal control device 200 as sensor information. Although the sensor 140 does not directly measure the average speed, it can be easily calculated from the traffic volume and the occupation rate by the signal control device 200 or the like, for example. Do.

  The signal control device 200 controls the traffic light 110 based on the sensor information and the traffic flow simulation result. The signal control device 200 includes a model parameter database 210 and a signal control unit 220.

  The model parameter database 210 accumulates sensor information sent from the sensor 140 and stores in advance the running model and traffic flow model parameters used when simulating the traffic flow on the road 120. The parameters of the travel model include, for example, the acceleration of the vehicle, the speed at which the behavior between vehicles is transmitted, the deceleration rate used when changing the travel lane, the traffic volume, the queue length, and the like. The parameters of the traffic flow model include, for example, a road network structure, a link saturated traffic flow rate, a branching rate, a link capacity, and regulation information.

  The signal control unit 220 refers to the model parameter database 210, simulates the traffic flow about several minutes ahead, generates signal control information for smoothing the traffic on the road network based on the result, and sends it to the traffic light 110. send.

  The traffic flow simulation inconsistency determination device 300 obtains sensor information, a travel model parameter (hereinafter referred to as “travel model parameter”) and a traffic flow model parameter (hereinafter referred to as “traffic volume model parameter”) from the signal control device 200. The traffic flow is simulated based on the information. Then, the traffic flow simulation inconsistency determination device 300 statistically detects the inconsistency of the traffic volume model parameter with respect to the actual traffic flow from the simulation result. The traffic flow simulation mismatch determination device 300 includes a sensor information input unit 310, a parameter input unit 320, a simulation unit 330, a mismatch determination unit 340, and a mismatch parameter extraction unit 350.

  FIG. 2 is a block diagram illustrating a configuration of the traffic flow simulation inconsistency determination device 300.

  The sensor information input unit 310 receives sensor information from the signal control procedure 200 shown in FIG. 1 and outputs the sensor information to each of the mismatch determination unit 340 and the simulation unit 330.

  The parameter input unit 320 inputs travel model parameters and traffic flow model parameters (hereinafter collectively referred to as “model parameters”) from the signal control device 200 shown in FIG. 1, and sends them to the simulation unit 330 and the mismatch parameter extraction unit 350, respectively. Output.

  The simulation unit 330 simulates the traffic flow on the road 120 in an arbitrary time zone after the time when the sensor information is measured based on the sensor information and the model parameters. Then, the simulation unit 330 outputs information such as traffic volume obtained as a simulation result (hereinafter referred to as “simulation information” as appropriate) to the inconsistency determination unit 340.

  The inconsistency determination unit 340 determines, based on the sensor information, a simulation result that is inconsistent with an actual traffic flow using a statistical method, and the determination result is used as the inconsistency parameter extraction unit 350 and the figure. 1 to the display device 400 shown in FIG.

  The inconsistency parameter extraction unit 350 outputs traffic flow model parameters (hereinafter referred to as “inconsistency parameters”) related to the points determined by the inconsistency determination unit 340 and the surrounding points (hereinafter referred to as “inconsistency points”). To do.

  The display device 400 shown in FIG. 1 displays sensor information, simulation information, inconsistency points, inconsistency parameters, and the like used in the traffic flow simulation inconsistency determination device 300, and displays various information including these information in the traffic flow simulation system. Notify 100 administrators.

  The signal control device 200 shown in FIG. 1 and the traffic flow simulation inconsistency determination device 300 shown in FIG. 2 are not shown, but are storage media such as a CPU (central processing unit) and a ROM (read only memory) storing a control program, respectively. A working memory such as a random access memory (RAM), a communication circuit, and the like. The CPU executes the control program and realizes the functions of the above-described units.

  Hereinafter, the operation of each part of the traffic flow simulation inconsistency determination device 300 will be described in detail.

  First, the parameter input unit 320 acquires model parameters used in the traffic flow simulation from the signal control device 200 and holds them. Next, the sensor information input unit 310 acquires sensor information actually measured by the sensor 140 (hereinafter referred to as “actual measurement information”) from the signal control device 200 and holds it.

  The simulation unit 330 acquires model parameters from the parameter input unit 320 and acquires measured information from the sensor information input unit 310. Then, the simulation unit 330 simulates the traffic flow based on the model parameters and the actual measurement information, generates simulation information including at least the traffic volume at an arbitrary time, and outputs the simulation information to the inconsistency determination unit 340. The arbitrary time is from the time when the actual measurement information is acquired to several tens of minutes later.

Here, the simulation unit 330 uses the traffic flow simulation to calculate the traffic volume (hereinafter “simulated traffic volume”) from the simulated information of the points 1 to n in the traffic network at the time point u minutes after the time when the actual measurement information was measured. Q _s1 (t + u), Q _s2 (t + u),..., Q _sn (t + u) are calculated. The traffic flow simulation uses the traffic volume (hereinafter referred to as “sensor traffic volume”) Q _d1 (t), Q _d2 (t),. , Q _dn (t). Since the specific method of the traffic flow simulation is a known technique, a description thereof is omitted here.

  The inconsistency determination unit 340 calculates a single regression line from the sensor traffic volume and the simulated traffic volume, and determines that an inconsistency occurs at a point that takes a value that deviates from the single regression line.

  The inconsistency determination unit 340 executes inconsistency determination processing for determining inconsistencies in the traffic flow model parameters based on the sensor traffic volume and the simulated traffic volume.

  FIG. 3 is a flowchart showing the flow of inconsistency determination processing.

First, in step S1000, the inconsistency determination unit 340 reads the sensor traffic volume Q _d1 (t + u) to Q _dn (t + u) from the sensor information input unit 310, and simulates the traffic volume Q _s1 (t + u) from the simulation unit 330. Read ~ Q _sn (t + u).

Next, in step S2000, the inconsistency determination unit 340 uses the detector traffic volumes Q _d1 (t + u) to Q _dn (t + u) as explanatory variables and the simulated traffic volumes Q _s1 (t + u) to Q _sn (t + u) as objective variables. As a result, a single regression line defined by the intercept b ₀ and the slope b ₁ is obtained.

Next, in step S3000, the inconsistency determination unit 340 has a range in which the simulated traffic volumes Q _s1 (t + u) to Q _sn (t + u) can be regarded as normal around the single regression line obtained in step S2000 (hereinafter referred to as “prediction section”). ").

The prediction interval is calculated by a method described below, for example. Here, when the degree of freedom is f and the significance level is α, the t value is t (f; α). When the number of samples is i and the number of constraints is j, the degree of freedom f is f = i−j, α = 0.05, and f = n−2. Also, placing the standard error se, the mean value of the sensor traffic _{Q d Q dave,} the standard deviation of the detectors traffic _{Q d} and sx. At this time, the inconsistency determination unit 340 calculates the predicted section upper limit value L _max and the predicted section lower limit value L _min of the simulated traffic volume Q _s for the arbitrary sensor traffic volume Q _d by the following equations (1) to (1) to Calculate using (3).

In step S4000, the inconsistency determination unit 340 determines whether or not the simulated traffic volume Q _s1 (t + u) to Q _sn (t + u) take values in the prediction section one by one in order. The inconsistency determination unit 340 detects a simulated traffic volume Q _{s *} (t + u) that takes a value outside the prediction interval, and detects this as a statistical outlier, and adds the point to the outlier point list. Add.

  In step S5000, the inconsistency determination unit 340 outputs the outlier spot list listed in step S4000 to the inconsistency parameter extraction unit 350 and the display device 400, and ends a series of processes.

Figure 4 is an explanatory diagram showing the relationship between the outlier and the prediction interval simulated traffic volume Q _s. The vertical axis represents the simulated traffic volume Q _s, a horizontal axis and detector traffic volume Q _d.

A single regression line 502, a predicted section upper limit value L _max and a predicted section lower limit value L _min are calculated for the plot 501 relating to the points 1 to n of the traffic network. Here, as shown in FIG. 4, if there is a plot 504 located outside the prediction interval sandwiched between the prediction interval upper limit value L _max and the prediction interval lower limit value L _min , the point corresponding to this plot is off. Detected as a value.

In this way, when there is a point that takes an outlier, that is, a point that takes a value outside the range where the simulated traffic volume Q _s can be regarded as normal, the outlier point list is output to the inconsistent parameter extraction unit 350. Is done.

  When the outlier list of points is input, the inconsistency parameter extraction unit 350 extracts parameters related to neighboring roads and intersections for each point described in the point list and outputs them to the display device 400. This parameter includes, for example, the saturated traffic flow rate and link capacity of the link (road) determined to be inconsistent by the inconsistency determination unit 340, and the signal control parameters of each node (intersection) located upstream and downstream of the link. Etc. are included.

  The display device 400 displays a map of a road network prepared in advance, and displays sensor information, simulation information, and inconsistency points input from the traffic flow simulation inconsistency determination device 300 in association with the map. . As a result, the administrator of the traffic flow simulation system 100 can easily grasp a portion where there is a mismatch between the actual traffic situation and the simulated traffic situation, and can identify a mismatched portion of the traffic flow model parameters.

  As described above, according to this embodiment, a single regression line is calculated from the sensor traffic volume and the simulated traffic volume, and the simulated traffic volume is inconsistent with the actual sensor traffic volume from an error with respect to the single regression line. Determine. As a result, a point where the simulated traffic is inconsistent with the actual detector traffic, that is, a point where the simulated traffic is inconsistent with the actual traffic is found. In addition, this makes it possible to carry out measurement and investigation only in the inconsistent points, and to efficiently adjust the setting values of the traffic flow simulation.

  In addition to ultrasonic vehicle detectors, various types of vehicle detectors such as optical vehicle detectors, microwave vehicle detectors, and image type vehicle detectors are used as devices for measuring traffic information. Of course it is good. In the present embodiment, the single regression line is calculated from the detector traffic volume and the simulated traffic volume, but the present invention is not limited to this. In other words, if the measured value of traffic flow information for multiple points is combined with the simulated value of traffic flow information for each corresponding point when simulating the traffic flow of the road network, the average speed and queue length Various other information such as the above may be used. Further, as long as there is a correlation between the actually measured value and the simulated value, these do not necessarily have to be the same type of information.

  The traffic flow simulation inconsistency determination device and the traffic flow simulation inconsistency determination method according to the present invention are a traffic flow simulation inconsistency determination device and a traffic flow simulation inconsistency that find a point where a simulated traffic flow is inconsistent with an actual traffic flow. This is useful as a matching determination method.

The system block diagram which shows the structure of the traffic flow simulation system using the traffic flow simulation inconsistency determination apparatus which concerns on one embodiment of this invention The block diagram which shows the structure of the traffic flow simulation inconsistency determination apparatus in this Embodiment Flow chart showing the flow of inconsistency determination processing in the present embodiment Explanatory drawing which shows the relationship between the prediction area in this Embodiment, and the outlier of simulated traffic

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Traffic flow simulation system 110 Traffic light 120 Road 130 Vehicle 140 Sensor 200 Signal control device 210 Model parameter database 220 Signal control unit 300 Traffic flow simulation mismatch judgment device 310 Sensor information input unit 320 Parameter input unit 330 Simulation unit 340 Inconsistency Determination unit 350 Inconsistent parameter extraction unit 400 Display device

Claims (9)

Actual measurement information input means for inputting actual measurement values of traffic flow information for a plurality of points on the road network;
Obtaining a regression line between the simulated value of the traffic flow information for the plurality of points and the measured value for the plurality of points when simulating the traffic flow of the road network, and from the error with respect to the regression line, Inconsistency judging means for judging a point where the inconsistency with the actual traffic flow,
A traffic flow simulation inconsistency judging device having Simulation means for simulating the traffic flow of the road network based on the measured values for the plurality of points using predetermined parameters;
The traffic flow simulation mismatch determination device according to claim 1, further comprising: Inconsistency parameter extraction means for extracting parameters used for simulating the inconsistent point;
The traffic flow simulation inconsistency determination device according to claim 2, further comprising: The traffic flow information is measurement information of a vehicle detector installed at each of the plurality of points.
The traffic flow simulation mismatch determination device according to claim 1. The traffic flow information includes at least traffic volume,
The inconsistency determination means includes
Obtaining a regression line between the measured value and the simulated value of the traffic volume for the plurality of points,
The traffic flow simulation mismatch determination device according to claim 1. Display means for displaying points determined to be inconsistent by the inconsistency determining means;
The traffic flow simulation mismatch determination device according to claim 1, further comprising: Display means for displaying the parameters extracted by the inconsistent parameter extraction means;
The traffic flow simulation inconsistency determination device according to claim 2, further comprising: An information input step for inputting measured values of traffic flow information for a plurality of points on the road network, and simulated values of traffic flow information for the plurality of points when simulating the traffic flow of the road network;
A regression line calculating step for calculating a regression line between the measured value and the simulated value for the plurality of points;
A mismatch determination step for determining a point where the simulated traffic flow is inconsistent with the actual traffic flow from an error with respect to the regression line;
A traffic flow simulation inconsistency determination method comprising: An information input step for inputting measured values of traffic flow information for a plurality of points on the road network, and simulated values of traffic flow information for the plurality of points when simulating the traffic flow of the road network;
A regression line calculating step for calculating a regression line between the measured value and the simulated value for the plurality of points;
A mismatch determination step for determining a point where the simulated traffic flow is inconsistent with the actual traffic flow from an error with respect to the regression line;
A traffic flow simulation inconsistency judgment program for causing a computer to execute.

Images