JP2001076283A - Traffic volume estimating device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an OD(origin-destination) traffic volume capable of accurately reproducing a traffic situation such as the occurrence of a congestion by calculating some demand traffic volume with a passing request at some spot. SOLUTION: The traffic volume estimating device consists of a demand traffic calculating means for calculating a demand traffic volume desired to pass through some spot with a traffic volume expressing the number of passing vehicles per a unit time at the spot and an occupation ratio expressing the rate of a vehicle existing time per the unit time as inputted values, a traffic volume distribution means for distributing a traffic volume with a road network and an initial OD traffic volume as inputs to calculate the traffic volume of each road on the road network, and an OD traffic volume correcting means for correcting an OD traffic volume so that the distributed traffic volume may be coincident with the demand traffic volume.

Description

DETAILED DESCRIPTION OF THE INVENTION

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a traffic flow simulator and a device for obtaining OD traffic used as basic data in estimating and predicting traffic conditions.

[0001]

2. Description of the Related Art Conventionally, when performing a traffic flow simulation or estimating or predicting a traffic situation,
OD traffic may be required. OD traffic in a narrow area can be relatively easily investigated and obtained by a method such as license plate matching. However, when targeting a wide range, the OD traffic between relatively large zones can be obtained from existing survey data, but the detailed OD traffic cannot be surveyed. Therefore, in order to obtain detailed OD traffic in a wide range, it is necessary to perform OD traffic estimation. In the conventional OD traffic volume estimation method, it matches the observed values such as the branch traffic volume at the intersection, the branch probability, and the road traffic volume by the entropy maximization method, residual sum of squares minimization method, maximum likelihood estimation method, etc. Such OD traffic is required.
There is also a method of improving accuracy by making a correction based on the surveyed rough OD traffic. It is also known to perform more detailed estimation of OD traffic volume in consideration of time change. However, all of these approaches take only traffic or probabilities related to traffic as inputs. In addition, verification is performed based on whether only the traffic volume is appropriate.

[0002]

However, when only the traffic volume is input, the O which correctly represents the traffic situation at the time of occurrence of traffic congestion.
D Traffic volume cannot be estimated. It is due to the following: In general, the relationship between the traffic volume and the vehicle density indicating the degree of congestion on the road is as shown in FIG. What is understood from this figure is as follows. There are areas where the traffic volume increases as the vehicle density increases, and areas where the traffic volume decreases as the vehicle density increases. Vehicle density K is Kc
Beyond, congestion becomes severe, and traffic congestion causes a gradual decrease in traffic volume, eventually leading to a stagnant flow. FIG. 13 shows the characteristics of such a traffic flow. Vehicle density K is K
The state below c is called free flow, and the state above Kc is called congestion flow. From this relationship, it can be seen that the traffic volume is Qt in two cases: a free flow with a vehicle density of Ktl and a congestion flow with a vehicle density of Kt2. Therefore, when determining the OD traffic volume that reproduces a certain traffic situation, if only the traffic volume is used, whether the state is free flow,
It was not possible to distinguish whether the traffic flow was congested, and it was not possible to estimate the OD traffic volume representing the actual traffic situation.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and has an object to solve the problem of a traffic flow in a free flow state where a flow is relatively smooth and a traffic flow state in which a flow is stagnant. By introducing the concept of the demand traffic volume representing the traffic volume required to pass through for the traffic volume of, the OD traffic volume more suitable for the actual traffic situation is obtained.

[0004]

According to an aspect of the present invention, there is provided an information processing system comprising: a traffic volume representing a number of passing vehicles per unit time at a certain point; and an occupancy rate representing a vehicle existence time ratio per unit time. A demand traffic calculation means for calculating a demand traffic going to pass through the point as an input value, and a traffic distribution using the road network and an initial OD (starting point / destination) traffic amount as input, and performing traffic distribution on the road network. Traffic distribution means for calculating the traffic volume of each road, and OD for correcting the OD traffic volume so that the allocated traffic volume and the demand traffic volume match.
It is a traffic volume estimation device including a traffic volume correcting means.

Another invention is characterized in that the demand traffic amount calculating means calculates the demand traffic amount separately for the free flow and the congested flow in relation to the occupancy and the traffic volume. further,
According to another invention, the demand traffic volume calculating means approximates, in a traffic congestion flow, a state in which traffic is flowing while congested with a straight line in a relationship between the traffic volume and the occupancy rate. Considering the traffic congestion elimination process, the traffic volume less than the straight line is assumed as the demand traffic volume, and the traffic volume less than the straight line is the demand traffic volume that is larger than the traffic volume predicted with the straight line, The operation is performed.

[0006]

The traffic volume represents the number of passing vehicles per unit time at a certain point. The occupancy represents the ratio of the vehicle existence time per unit time at a certain point. From these two quantities, it can be determined whether the vehicle is in the free flow state or the traffic congestion state. Then, in accordance with each state, a demand traffic volume representing a traffic volume required to pass at that point is calculated. The OD traffic is distributed to each road on the road network, and the traffic on each road is calculated.
In order for this traffic volume to be equal to the above demand traffic volume,
D traffic volume is modified. Thus, the corrected OD traffic accurately represents the demand at each point. Therefore, it is possible to obtain an OD traffic volume that reflects the actual traffic situation.

[0007] In another invention, demand traffic volume is predicted separately for free flow and congested flow in relation to traffic volume and occupancy. Therefore, it is possible to obtain an OD traffic volume that more accurately represents the free flow state and the congested flow state. Further, according to another invention, in the traffic congestion flow area, the traffic congestion traffic volume is calculated by dividing the traffic flow into a straight line and a region divided into two by the straight line. The demand traffic volume is calculated according to the time when traffic is congested but in each case. Then, the OD traffic volume is calculated according to the demand traffic volume. When the travel time is calculated using the OD traffic volume, a value more suitable for the actual situation can be obtained.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described based on specific embodiments. The present invention is not limited to the following examples. FIG. 1 shows the configuration of an embodiment of the present invention. As a procedure, first, the demand traffic volume is calculated by the demand traffic volume calculation means using the traffic volume (sensor volume) and the occupancy (sensor occupancy) as input values. This traffic volume is the number of vehicles passing per unit time at a certain point on the road network and is a measured value. The occupancy is a measured value, which is a ratio of the time during which the vehicle is detected to the entire measurement time in a state measured from directly above at a certain point. Traffic volume and occupancy depend on vehicle density. The higher the vehicle density, the greater the traffic volume. However, when the vehicle density exceeds a certain value and congestion occurs and the vehicle speed decreases, the measured traffic volume decreases. On the other hand, the occupancy increases in proportion to the vehicle density.

In the present embodiment, data obtained from a vehicle sensor installed on a road is used for the traffic volume and the occupancy as input values. The unit time is determined by the time unit of the required OD traffic and the initial OD traffic. In this embodiment, 1
The traffic demand is calculated in units of 15 minutes by using the measurement data obtained by the sensors collected every 15 minutes. Next, the traffic volume distribution means distributes the 15-minute initial OD traffic volume data on the road network and calculates the 15-minute traffic volume of each road. Finally, the OD traffic correction means corrects the OD traffic that passes so that the allocated traffic on the road where the sensor is located and the demand traffic match.

Next, the contents of each means shown in FIG. 1 will be described. First, a method of calculating the demand traffic volume will be described. Traffic flow is roughly divided into free flow and congested flow. If the occupancy of the vehicle measured by the sensor at each unit time at a certain point is plotted on the horizontal axis and the traffic volume is plotted on the vertical axis, the free flow state and the congested flow state can be separated as shown in FIG. In the free flow state, there is no congestion at all,
All vehicles arriving there can pass. That is, the traffic volume measured by the sensor directly represents the demand traffic volume. However, in a traffic congestion state, since many vehicles are going to pass therethrough, the vehicle speed is reduced and traffic congestion occurs. The more traffic congestion, the less vehicles can pass. Therefore, the traffic volume measured by the sensor decreases, but conversely, the number of vehicles desired to pass,
That is, the demand traffic is large. In other words, it can be said that the demand traffic volume increases as the traffic volume measured by the sensor decreases. Specifically, in this example, the demand traffic volume was determined as follows.

In FIG. 2, a straight line a corresponding to a free flow and a straight line b corresponding to a congested flow are determined as shown in FIG.
The way to decide is as follows. First, the time zone of each data,
First, data is divided into a free flow and a congested flow from the shape of a graph representing the relationship between the traffic volume and the vehicle density while referring to other data such as travel time. The equation of the straight line a is obtained by regression analysis of free flow data. As shown in FIG.
From the point on the straight line a, for example, the measured detector traffic volume of 500 vehicles / hour is directly used as the demand traffic volume.

Next, the traffic congestion flow data will be analyzed in more detail. As shown in FIG. It can be seen that there are data shown below, in which traffic is heavy and traffic volume is difficult to flow, and data in a state in which traffic and supply are relatively balanced in traffic between them.

Therefore, the data in the first two states is removed from the congested flow, and regression analysis is performed using the data in the balanced state to obtain a straight line b. Once again, the free flow is on the left side of the intersection of the two straight lines, and in the case of the free flow, the traffic volume measured by the sensor is used as the demand traffic volume. The traffic flow on the right side of the intersection is determined as follows on a straight line, above the straight line, and below the straight line as a congestion flow. First, for the data on the straight line, the detector traffic volume is defined as the demand traffic volume.

On the upper side of the straight line b, it is considered that the traffic amount that has been staying so far when the traffic congestion is resolved flows, and the demanded traffic amount itself is made smaller than the straight line. here,
The value obtained by subtracting the value obtained by multiplying the square of the traffic volume difference between the straight line and the sensor traffic volume by a coefficient from the traffic volume obtained from the straight line corresponding to the occupancy is given as the demand traffic volume.

In order to cause heavy traffic congestion on the lower side of the straight line b, it is necessary to give a traffic amount more than the straight line traffic amount. Therefore, the difference of the traffic amount between the straight line and the traffic amount measured by the sensor is 2
The demand traffic volume is obtained by adding the value obtained by multiplying the power to the coefficient by the traffic volume obtained from the straight line corresponding to the occupancy. However,
Since an excessively large amount of addition cannot be generated because it is meaningless, the upper limit of the traffic difference between the straight line and the traffic measured by the sensor is set to 50 vehicles / lane. The coefficient was 0.04 for both the upper and lower sides of the straight line. In this study, we considered the actual travel time, and as a result, determined the traffic volume of traffic congestion in this way. You only have to consider it. For example, as shown in FIG.
The traffic demand per vehicle / hour is determined.

Next, the traffic distribution means will be described. The traffic distribution method includes the All or Nothing method in which all OD traffic is distributed to one or more routes, such as the shortest time route and the shortest distance route, at one time. There are a split allocation method that considers an increase in travel time of a road due to passing, and a dial allocation method that calculates and allocates an effective route for allocating OD traffic and its selection probability and performs allocation. In addition, a method has been devised in which the above is applied and more detailed distribution is performed for each time zone in consideration of the remaining outflow amount in the previous time zone. In the present invention, the distribution may be performed using any of these methods, and the traffic passing through each road may be obtained using a traffic flow simulation.

In this embodiment, the traffic volume is distributed as follows. First, the input road network is shown in FIG.
As shown in Fig. 5, a road with five intersections on a road with a total length of about 3.5 km was used. In this estimation of OD traffic volume, only the direction of the arrow in the figure is targeted.
Determined as shown in An initial OD traffic volume of 15 minutes is used. In addition, it is assumed that the vehicle detectors are installed at five locations shown in the figure. The allocation time zone is 15 minutes, the same as the initial OD traffic volume, and the road network is small, so the remaining outflow volume in the previous time zone is not taken into account.
The method was used. The route selection characteristics used in the allocation method are also simple, in which only the shortest distance route is selected for one OD pair because the road network is small and no alternative route exists.

As described above, the demand traffic volume and the allocated traffic volume are calculated, and the OD traffic volume correction means corrects the OD traffic volume so that they match. As the method, a residual sum of squares minimizing method, an entropy maximizing method, or the like may be used. This time, the OD traffic volume was corrected according to the procedure shown in FIG. First, the difference between the allocated traffic volume and the demand traffic volume in the entire time zone on the road where the sensor exists is calculated. If the convergence condition is satisfied, the calculation ends. If not, first, increase the passing OD traffic on the road where the allocated traffic is smaller than the demand traffic. Next, the passing OD traffic on the road where the allocated traffic is larger than the demand traffic is reduced. Finally, if the traffic volume on all roads after the increase / decrease exceeds the road capacity, the passing OD
Reduce traffic. After that, the difference between the allocated traffic volume and the demand traffic volume is calculated again. The above OD traffic volume correction is repeated until the convergence condition is satisfied.

As described above, OD traffic estimation was performed. A traffic flow simulation was performed using the estimated OD traffic to confirm the reproduction accuracy of the traffic situation based on the estimated OD traffic. The block density method is used for traffic flow simulation. The results of comparing the measured values and the simulated values for the traffic volume, occupancy rate, and travel time of the entire section at each point are shown below. 6 and 7 show the time change of the traffic volume and the time change of the occupancy at a certain point. FIG. 8 shows the time change of the 30-minute average travel time of the entire section. For comparison, the same point when OD traffic estimation was performed to match the traffic volume measured by the sensor without using the demand traffic volume as in the past, and the traffic flow simulation was performed using it. 9 and 10 show the time change of the traffic volume and the time change of the occupancy rate in
FIG. 11 shows the time change of the 30-minute average travel time of the entire section. 6 and 9, the traffic volume relatively well agrees with the actually measured value in both the conventional method and the present method. However, FIG.
Looking at 0, when the actual method and the present method are used at the time of congestion in the evening, the occupancy rate increases and congestion occurs, whereas no congestion occurs when the conventional method is used. You can see that. This is because although the decrease in the traffic volume in the evening in the measured values is due to traffic congestion, the conventional method cannot distinguish between free flow and congested flow, so it is mistakenly regarded as free flow. That's why. Further, from FIGS. 8 and 11, it can be seen that the increase in travel time due to traffic congestion is not reproduced in the conventional method, but the traffic congestion occurs in the present method and the section travel time increases in the same manner as the measured value.

The present invention does not merely estimate the OD traffic from the measured passing traffic per unit time,
The occupancy rate, which is the percentage of time that the vehicle occupies the road surface, is measured, and the traffic volume measured by the occupancy rate is divided into a free flow and a congested flow. It is a feature of the present invention that the free flow and the traffic congestion flow, which could not be known only by the conventional passing traffic volume, are divided according to the occupancy. In a traffic congestion state, the measured passing traffic volume does not necessarily indicate the degree of congestion.
Therefore, the travel time cannot be properly calculated, and the feature is that the demand traffic volume is obtained in consideration of this correction. The method of calculating the demand traffic volume by this correction is arbitrary.
Another feature of the present invention is that the demand traffic volume is calculated separately for a congested state, a congested but flowing state, and a completely congested state.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an apparatus according to a specific embodiment of the present invention.

FIG. 2 is a measurement diagram showing the relationship between the occupancy and the traffic volume.

FIG. 3 is an explanatory diagram showing regions divided for calculating a demand traffic volume in the measurement diagram.

FIG. 4 is an explanatory diagram showing a relationship between a road network, a measurement position, and an occurrence / arrival point of OD traffic.

FIG. 5 is a flowchart showing a procedure for calculating an OD traffic volume so that the allocated traffic volume matches the demand traffic volume.

FIG. 6 is a characteristic diagram showing a traffic volume simulated for each time zone using the OD traffic volume obtained by the apparatus of the present embodiment, and actual measured values thereof.

FIG. 7 is a characteristic diagram showing the occupancy simulated for each time zone using the OD traffic volume obtained by the apparatus of the present embodiment and the actually measured values.

FIG. 8 is a characteristic diagram showing a travel time simulated for each time zone using the OD traffic volume obtained by the apparatus according to the present embodiment and its actually measured value.

FIG. 9 is a characteristic diagram showing a traffic volume simulated for each time zone using the conventional uncorrected OD traffic volume and its actually measured value.

FIG. 10 is a characteristic diagram showing occupancy rates simulated for each time zone using the conventional uncorrected OD traffic volume and their actual measurement values.

FIG. 11 is a characteristic diagram showing a travel time simulated for each time zone using the conventional uncorrected OD traffic volume and its actually measured value.

FIG. 12 is an explanatory diagram showing a relationship between a calculated demand traffic volume and a measured traffic volume in a free flow and a congested flow in the apparatus according to the embodiment.

FIG. 13 is a characteristic diagram showing a relationship between vehicle density and traffic volume.

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Eiji Teramoto, inventor 41 F-term in Toyota Central R & D Labs.

Claims (3)

[Claims] 1. A demand traffic volume that is going to pass through a point is calculated by using as input values a traffic volume representing the number of vehicles passing per unit time at a certain point and an occupancy rate representing a vehicle existence time ratio per unit time. Means for calculating traffic demand, traffic distribution means for performing traffic distribution by inputting a road network and initial OD (starting point-to-destination) traffic, and calculating traffic volume of each road on the road network; OD so that the calculated traffic volume and the demand traffic volume match.
A traffic volume estimation device comprising OD traffic volume correction means for correcting traffic volume. 2. The demand traffic calculating means calculates the demand traffic by dividing the demand into the free flow and the traffic congestion based on the relationship between the occupancy and the traffic. A traffic volume estimating device according to the above. 3. The traffic demand calculation means, in the relationship between the traffic volume and the occupancy, approximates a straight line in the traffic congestion flow while the traffic is congested. The part with a large amount is regarded as a congestion elimination process,
Calculating the traffic volume smaller than the traffic volume predicted by the straight line as the demand traffic volume, and calculating the traffic volume smaller than the traffic volume of the straight line as the demand traffic volume. The traffic volume estimating device according to claim 2, characterized in that: