JP3402263B2 - Traffic information management device and traffic information management method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a traffic information management device and method, and more particularly to a traffic information management device and method for detecting a vehicle detected at an upstream point at a downstream point.

[0002]

2. Description of the Related Art Conventionally, there are cases where a field control vehicle is monitored in a traffic control system. this is,
It tracks the travel of a specific vehicle.

[0003]

The most reliable method of tracking the vehicle in the state of the art is to use a license plate reader. In this system, a license plate of a vehicle to be tracked is read by a reader at one location on a road, and a reader provided at another position detects a vehicle having the license plate.

In such a system, the space to be detected is generally specified. Also, enter the license plate of the vehicle to be tracked into the system in advance,
It is also possible to follow it.

For example, the purpose is to set or detect in real time a vehicle that ignores a signal, a vehicle that violates parking / speeding, another runaway vehicle, or a vehicle that is desired to be tracked for some reason, and track this from that location. In order to reach, it is not possible with a single device and requires multiple license plate readers. However, there is a problem that the license plate reading device is expensive and there is a limit to its widespread use throughout the target road network.

Therefore, the present invention has an object of manufacturing a traffic information management device capable of searching for a vehicle at a low cost.

[0007]

[0008]

[0009]

[0010]

[0011]

[0012]

[0013]

[0014]

[0015]

[0016]

According to one aspect of the present invention, a traffic information management device is installed at an upstream point of a road,
First collection means for collecting the traffic volume, the characteristic amount of the passing vehicle and the passing time of the vehicle, and the traffic amount, the characteristic amount of the passing vehicle and the passing time of the vehicle installed at a downstream point of the road. Second collecting means, an acquiring means for acquiring a reference time estimated to be a travel time between the upstream point and the downstream point, the acquired reference time and the vehicle collected by the first collecting means. A predicting unit that predicts a time when the vehicle that has passed through the upstream point reaches the downstream point based on the passing time of the second point, and the characteristic amount collected by the first collecting unit of the characteristic amounts collected by the second collecting unit. Deviation from the second
Information of the group of vehicles collected by the first collecting means and the group of vehicles collected by the second collecting means based on the deviation of the passing time of the vehicle from the predicted time collected by the collecting means Matching means for matching the vehicle information, calculating means for calculating the travel time between the upstream point and the downstream point based on the matching result, and a reliability calculating means for calculating the reliability of the travel time. An estimating means for estimating the time when the vehicle detected at the upstream point arrives at the downstream point based on the travel time and the reliability, and the vehicle detected at the upstream point based on the estimated time. And a detecting means for detecting at a downstream point, and the matching means uses the second collecting means.
Deviation of collected vehicle transit times from predicted times
An evaluation value is calculated based on the amount .

According to the present invention, the travel time between the upstream point and the downstream point is calculated based on the matching result, so that the travel time of the vehicle between the upstream point and the downstream point can be calculated accurately and efficiently. It is possible to provide a traffic information management device that can perform calculation and can detect a vehicle with high reliability based on the result.

Preferably, each of the first and second collecting means includes a loop type sensor and an identifying means for identifying the vehicle length of the vehicle as a characteristic amount based on the output of the sensor.

When the loop type sensor is adopted as the collecting means in this way, it becomes easy to collect the characteristic amount of the vehicle.

Preferably, each of the first and second collecting means comprises an ultrasonic sensor.

When the ultrasonic type sensor is adopted as the collecting means in this way, the characteristic amount can be collected easily.

Preferably, each of the first and second collecting means includes a sensor for collecting a running sound of the vehicle.

In this way, when the sensor for collecting the traveling sound is adopted as the collecting means, the collection of the characteristic amount becomes easy.

Preferably, each of the first and second collecting means is a camera for obtaining an image of the vehicle, and a singular number from a group of vehicle width, vehicle height, vehicle length, vehicle color, brightness and pattern from the image. It also includes an image processing device that extracts a plurality of selected feature amounts.

When the camera is adopted as the collecting means in this way, the collection of the characteristic amount becomes easy.

Preferably, each of the first and second collection means comprises an optical vehicle sensor.

When the optical vehicle sensor is adopted as the collecting means in this way, the characteristic amount can be collected easily.

Preferably, each of the first and second collection means comprises a microwave vehicle detector.

When the microwave type vehicle detector is adopted as the collecting means as described above, the characteristic amount can be collected easily.

Preferably, the first and second collecting means collect a plurality of types of characteristic quantities, and the matching means performs matching by weighting a plurality of types of characteristic quantities.

In this way, a plurality of types of feature quantities are collected,
If matching is performed by weighting a plurality of types of feature amounts, more accurate travel time calculation can be performed.

Preferably, the matching means determines the deviation of the number of vehicles from the predicted passing time of the passing times of the vehicles collected by the second collecting means.

In this way, if the deviation of the number of vehicles is determined, the deviation can be easily determined.

Preferably, the matching means uses only the characteristic amount of the vehicle satisfying a specific criterion and the passing time of the vehicle for matching.

As described above, when only the characteristic amount of the vehicle satisfying the specific criterion and the passing time of the vehicle are used for the matching, the processing in the traffic information management device becomes easy.

Preferably, the first and second collecting means collect only the characteristic amount of the vehicle satisfying a specific criterion and the passing time of the vehicle.

As described above, by collecting only the characteristic amount of the vehicle satisfying the specific criterion and the passing time of the vehicle, the processing in the traffic information management device becomes easy.

According to another aspect of the present invention, a traffic information management device is installed at an upstream point of a road, collects a characteristic amount of a passing vehicle and a passing time of the vehicle, and downstream of the road. It is installed at a point, and a second collection means for collecting the passage time of the feature amount and the vehicle of a vehicle passing, by the feature amount and the second collecting means one vehicle collected by the first collection means Deviation from the collected feature amount of one vehicle and the collected by the first collecting means
The passing time of the one vehicle, the passing time of one vehicle collected by the second collection means, based on the reference time as the estimated travel time between these two points, a plurality of target vehicles each
The evaluation value is calculated for each of the
Matching means for matching the group of vehicle information collected by the first collecting means with the group of vehicle information collected by the second collecting means based on the evaluation value ;
Based on the result of matching, a calculation means for calculating the travel time between the upstream point and the downstream point, a reliability calculation means for calculating the reliability of the travel time, and the vehicle detected at the upstream point reaches the downstream point. Then, the estimation unit estimates the estimated time based on the travel time and the reliability, and the detection unit detects the vehicle detected at the upstream point at the downstream point based on the estimated time.

According to the present invention, since the travel time between the upstream point and the downstream point is calculated based on the matching result, the travel time of the vehicle is calculated accurately and efficiently, and based on the result. Therefore, it is possible to provide a traffic information management device that can detect a vehicle with high reliability.

According to another aspect of the present invention, a traffic information management method includes a first collecting step of collecting a characteristic amount of a passing vehicle and a passing time of the vehicle at an upstream point of a road, and a downstream point of the road. In the second step, a second collecting step of collecting a feature amount of a passing vehicle and a passing time of the vehicle, an obtaining step of obtaining a reference time estimated to be a travel time between an upstream point and a downstream point, Based on the acquired reference time and the passing time of the vehicle collected in the first collecting step, the predicting step of predicting the time when the vehicle passing through the upstream point reaches the downstream point, and the collecting step by the second collecting step The deviation of the collected characteristic amount from the characteristic amount collected in the first collecting step and the deviation of the passing time of the vehicle collected in the second collecting step from the predicted time. Based on the matching step of matching the information of the group of vehicles collected by the first collecting step with the information of the group of vehicles collected by the second collecting step, based on the matching result, the upstream point Calculation step to calculate the travel time between the travel point and the downstream point, a reliability calculation step to calculate the reliability of the travel time, and the travel time at which the vehicle detected at the upstream point is estimated to reach the downstream point. And a reliability, and a detection step of detecting a vehicle detected at an upstream point at a downstream point based on the estimated time, and a matching step.
Of the vehicles collected by the second collection step
Based on the amount of deviation of the transit time from the predicted time
Calculate the evaluation value .

According to the present invention, a traffic information management method capable of accurately and efficiently calculating the travel time of a vehicle between an upstream point and a downstream point and detecting the vehicle with high reliability based on the result. Can be provided.

According to another aspect of the present invention, in a traffic information management method, a first collecting step of collecting a feature amount of a passing vehicle and a passing time of the vehicle at an upstream point of a road, and a downstream point of the road. in the second and collection step, the feature amount and the second collecting step of one <br/> vehicles collected by the first collecting step of collecting the features and passing time of the vehicle passing vehicles 1 collected by
Deviation of a feature quantity of the base of the vehicle, and the passage time of the first one vehicle collected by the collecting step, the passing time of one vehicle collected by the second collecting step,
An evaluation value is calculated for each of a plurality of vehicles of interest based on the reference time that is the estimated travel time between these two points , and
A matching step of matching the group of vehicle information collected by the first collecting step with the group of vehicle information collected by the second collecting step based on the plurality of evaluation values obtained by the above ; A calculation step to calculate the travel time between the upstream point and the downstream point based on the matching result, a reliability calculation step to calculate the reliability of the travel time, and the vehicle detected at the upstream point reaches the downstream point. Then, the estimation step of estimating the estimated time based on the travel time and the reliability, and the detection step of detecting the vehicle detected at the upstream point at the downstream point based on the estimated time.

According to the present invention, the traffic information management method capable of accurately and efficiently calculating the travel time of the vehicle between the upstream point and the downstream point and detecting the vehicle with high reliability based on the result. Can be provided.

[0044]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [First Embodiment] FIG. 1 is a block diagram showing the configuration of a road traffic control system according to a first embodiment of the present invention. Referring to the figure, the road traffic control system is roughly divided into a traffic control center 100.
A traffic information collecting device 200 installed on the road, a traffic information providing device 300 installed on the road, a CRT 400 displaying video data and a characteristic amount of a vehicle, and information obtained from the traffic information providing device 300 and displayed. Information display board 50 for performing
0 and a vehicle 6 that obtains information from the traffic information providing device 300
00 and.

The traffic control center 100 is a device for traffic information processing and system monitoring, and is composed of a traffic information processing device and an information monitoring device.

The traffic information processing device measures traffic volume, speed and occupancy rate, collects OD information, measures congestion degree and congestion length, compression wave analysis, accident event detection, and runaway vehicle detection. Specific vehicle tracking, earthquake detection, travel time measurement, existing vehicle number estimation, travel time prediction, inflow control processing, and traffic flow simulation processing are performed.

The information monitoring device comprises a system display board, a system monitoring device, and a system intervention device.
The information monitoring device allows the user to obtain traffic information.

The traffic information collecting device 200 is composed of ultrasonic type vehicle detectors installed at least at two locations (upstream and downstream) on the road, and collects vehicle feature quantities.

The upstream point and the downstream point mentioned here are as shown in FIGS. Here, the white triangle indicates the upstream point, the black triangle indicates the downstream point, and the arrow indicates the flow of the vehicle. 2 shows a case of a single road, FIG. 3 shows a case of a branch, and FIG. 4 shows a case of an intersection.

Alternatively, even if there is a branch or an intersection,
The upstream and downstream points may be set along the route, such as Tomei Expressway, National Route 2, Koshu Kaido, and Meiji Dori.

The traffic information providing device 300 includes the information display board 5
00 and automobile 600 are provided with information.

FIG. 5 is a view showing the external appearance of the road traffic control system. Referring to the figure, the traffic information collecting device includes ultrasonic vehicle detectors 201a and 201b and primary processing devices 203a and 203b that process signals from the ultrasonic vehicle detectors 201a and 201b.

Ultrasonic vehicle detectors 201a, 201b
Is provided on the road R and detects that the vehicle V passing through the road has passed the position and the height of the vehicle (vehicle height). The detection signal is the primary processing devices 203a and 203b.
Entered in. From the primary processing devices 203a and 203b, the time when the vehicle passes through the upstream point or the downstream point and the vehicle height of the vehicle are output and sent to the traffic control center 100.

Here, the ultrasonic type vehicle detector 201 is used.
It is assumed that a is provided downstream of the road from the ultrasonic vehicle detector 201b. Ultrasonic vehicle detector 201
The point where "a" is installed is called "downstream point", and the point where the ultrasonic vehicle detector 201b is installed is called "upstream point".

In the present embodiment, a vehicle group including a fixed number (N vehicles) of vehicles (which is defined as a vehicle of interest) satisfying a certain standard at an upstream point is measured as one vehicle group. Subject to. At the downstream point, it is considered that the number of vehicles that make up this vehicle group and the order of the vehicles have changed, so the characteristic amount of the vehicle group obtained at the upstream point and the characteristic amount of the vehicle obtained at the downstream point By evaluating and as a whole according to the predetermined matching evaluation criteria,
Correlate vehicle groups.

FIG. 6 is a graph in which the vehicle height (feature amount) detected by the ultrasonic vehicle detector 201b provided at the upstream point is shown on the vertical axis, and the time when the detection is performed is shown on the horizontal axis. is there. As shown in the figure, the characteristic amount of the vehicle detected with the passage of time is recorded.

In order to improve the performance of the matching process, the vehicles to be matched (vehicles of interest) are limited to those having a characteristic amount with a relatively low appearance frequency (here, the vehicle height is relatively high). . Further, in the present embodiment, N = 10 vehicles (1) to (10) are considered as vehicles of interest.
Is to be used. 10 vehicles are detected at time t
_It is assumed that the period is from _u to time T.

FIG. 7 is a graph in which the vertical axis indicates the vehicle height (feature amount) detected by the ultrasonic type vehicle detector 201a provided at the downstream point, and the horizontal axis indicates the time when the detection is performed. is there. The detection here is performed from the reference time t _d . The reference time t _d is determined by the time t _u and the travel time measured last time.

As shown in FIG. 7, at the downstream point, the order of the vehicles of interest and the number of vehicles of interest constituting the vehicle group are different from those of the upstream point (FIG. 6). For more details, see FIG.
In comparison with FIG. 7, vehicles (2) and (5) are not detected in FIG. This is because the detection error of the sensor occurs, the lanes of the vehicles (2) and (5) are changed, the vehicle moves to another branch road, or the vehicle stops.

At the downstream point, a vehicle not detected at the upstream point (vehicle indicated by ●) is newly detected.

Matching is performed based on the data shown in FIGS. 6 and 7, and the vehicle group (1) shown in FIG.
The time (travel time) for (10) to move to the downstream point is determined.

For example, considering each vehicle, it can be determined as follows whether the vehicle (7) passing through the upstream point at time t ₇ in FIG. 6 corresponds to the vehicle detected at the downstream point. Will be judged.

First, the reference time t _d is obtained by adding the travel time between the upstream point and the downstream point obtained previously to the time t _u when the vehicle detection is started at the upstream point. Then, by adding the reference time t _d a (t ₇ -t _u), determining an expected time for the vehicle (7) (time considered will pass through a point downstream).

T before and after the expected time for this vehicle (7)
_A vehicle that is considered to be closest to the vehicle (7) is determined using a predetermined function within a time range of ₀ (t ₀ is a predetermined time). Here, t ₀ is determined by how much the characteristics of travel time between the upstream point and the downstream point of the vehicle of interest vary on average.

The following process is performed to determine the vehicle that is considered to be closest to the vehicle (7).

First, time (t _d + (t ₇ −t _u )) ± t
_In the range of ₀ , a candidate vehicle corresponding to the target vehicle (here, vehicle (7) in FIG. 6) is obtained. In this, a vehicle having a characteristic amount relatively close to the characteristic amount of the vehicle of interest is set as a candidate vehicle. However, the number m of candidate vehicles is a constant value m
If there are ₀ or more vehicles, it is considered that there are too many candidate vehicles and reliable handling is not possible, and it is handled in the same manner as when there is no candidate vehicle. Note that m ₀ is determined here by the detection error of the feature amount and the number N of vehicles of interest.

Then, for each candidate vehicle, the difference (feature amount difference) x between the feature amount and the target vehicle is calculated. Further, the deviation y of the time when each candidate vehicle is detected at the downstream point from the expected time is calculated.

Then, the vehicle having the smallest value of a | x | + b | y | is regarded as the vehicle closest to the vehicle of interest, and the processing is performed. Here, the variables a and b are predetermined constants.

One vehicle group (eg, (1) to FIG. 6)
The minimum value of a | x | + b | y | is obtained for each of the target vehicles included in (10)) that have candidate vehicles, and the average value thereof is the reference time (here, t _d ) Is evaluated value.

However, if the number of vehicles of interest without candidate vehicles exceeds a fixed value n ₀ , it is determined that vehicles cannot be associated at the reference time, and the evaluation value is set to infinity. Here, n ₀ is substantially determined by the probability of vehicle inflow and outflow between N and the upstream and downstream points.

For example, referring to FIG. 8, vehicles A and C at downstream points are considered as candidate vehicles with respect to vehicle A at the upstream point, and vehicle C at downstream points with respect to vehicle C of interest.
Is a candidate vehicle, and the target vehicle D at the upstream point
Assuming that the vehicles E, F, and G at the downstream point are candidate vehicles and the vehicle E at the downstream point is a candidate vehicle with respect to the target vehicle E at the upstream point, the target vehicles A to The number m of each candidate vehicle of E is 2, 0,
It becomes 1, 3, 1. In addition, the number of target vehicles without candidate vehicles is 1 (only the target vehicle B). Further, the number n of vehicles of interest having candidate vehicles is four.

The reference time t _d is gradually shifted as shown in I to V of FIG. 9, and the evaluation value is calculated at each reference time. The one with the best evaluation value is determined as the time when the vehicle arrives.

10 to 12 show the traffic control center 100.
3 is a flowchart showing a process performed in.
Referring to the figure, the system is initialized in step S101. In this initial setting, conditions such as the vehicle of interest, the characteristic amount, and the processing cycle are set. Also, various constants are set.

In step S103, the traffic control center 100 refers to the current time. Step S10
At 5, it is determined whether or not the current timing is to start the periodic processing based on the referred time. Until YES, the processing from step S103 is repeated.

If YES at step S105, at step S107, the latest N vehicles of interest (here, N = 10) are selected based on the data obtained at the upstream point. Specifically, the vehicle (1) to (10) shown in FIG. 6 is selected.

In step S109, the passage times and the feature quantities of the vehicles of interest (1) to (10) are detected and set. In step S111, the reference times are sequentially shifted, and matching with the data obtained at the downstream point is started. For example, the reference time is assumed in the initial state is a reference time t _d as shown in FIG.

The vehicle of interest selected in step S113 is sequentially referred to. In step S115, the expected time of passage of each vehicle of interest at the downstream point is calculated. In step S117, the area around the expected time (± t ₀ ) is searched, and candidate vehicles for the vehicle of interest are searched. In step S119, it is determined whether or not there is a candidate vehicle, and if YES, step S1
If NO in step 21, the process returns to step S113 to process the next vehicle of interest.

In step S121, it is determined whether the number of candidate vehicles is a certain number or more. If YES, the process returns to step S113, and if NO, the process proceeds to step S123.

In step S123, the deviation amount x of the characteristic amount from the vehicle of interest and the deviation y from the expected time are recorded for each candidate vehicle. Next, in step S125, the evaluation value is calculated and recorded by the formula a | x | + b | y | for each candidate vehicle.

In step S127, the minimum evaluation value obtained for each candidate vehicle is accumulated in the accumulated evaluation value. In step S129, it is determined whether the processing has been completed for all the target vehicles. If NO, the process returns to step S113 to start the processing of the next target vehicle. If YES, the process proceeds to step S131.

At step S131, the target vehicle in which the upstream point and the downstream point are associated (the target vehicle with the candidate vehicle)
The number n of is referred to. In step S133, it is determined whether or not the number is equal to or less than the constant value n _0. If YES, the evaluation value is set to infinity in step S157. On the other hand, if NO, then in step S135 the cumulative evaluation value is divided by the corresponding number n of the target vehicles. Then, a value obtained by dividing the infinity evaluation value or the cumulative evaluation value in step S137 by the number n of the corresponding attention vehicles is recorded as the evaluation value at the reference time.

In step S139, it is determined whether the processing is completed for all reference times to be investigated. If NO, the process returns to step S111 to start matching at the next reference time. If YES, step S
Proceed to 141.

In step S141, the recorded evaluation value for each reference time is referred to. In step S143, it is checked whether there is a time zone in which the evaluation value suddenly decreases. If it is determined in step S145 that there is a time zone in which the evaluation value suddenly decreases, the process proceeds to step S147. If NO in step S145, the flow advances to step S151 to refer to the time series data of travel times measured up to the previous processing cycle. The tendency of the change in travel time measured in step S153 is grasped, and the travel time in this cycle is estimated. The process is terminated with the travel time estimated in step S155 as the travel time of the cycle.

In step S147, it is determined whether there is only one time zone when the evaluation value suddenly decreases. If NO, the process proceeds to step S151. If YES, in step S149, the travel time of the cycle is determined from the reference time when the evaluation value is the minimum, and the process ends.

The determinations in steps S143 and S145 are to select a time zone (a plurality of reference times) in which the vehicle may have passed from the obtained evaluation values for each reference time. In this case, the evaluation value of this possible time zone needs to be sufficiently small compared to all the time zones. If it is not small enough, matching is not possible for this train of interest and travel time cannot be measured, and the travel time for that cycle is calculated based on the travel time data up to the previous cycle. (S151 to S155).

In the present embodiment, the reference time having the best evaluation value within the time zone thus obtained is selected, and the travel time is determined based on the reference time ( S149).

For example, when the reference time is shifted with reference to FIG. 13, the evaluation value in a certain specific time zone is steeply smaller than that in other time zones (a steep valley is observed). It is a condition for adopting the reference time obtained by matching.

On the other hand, as shown in FIG. 14, when there is no time zone in which the evaluation value is sufficiently small (A) or when there are a plurality of valleys of the same level (B), the measurement is performed only by the periodic processing. Is determined to be insufficient, and the travel time in this cycle is determined by comprehensively evaluating travel time measured in time series up to that time.

As described above, in the present embodiment, the following data are used as the data relating to the matching evaluation criteria.

(1) Data relating to comparison between each vehicle of interest and its corresponding candidate vehicle Deviation of feature amount (x) Deviation from expected time (y) Number of candidate vehicles (m) (2) Attention Data in consideration of vehicle sequence Number of vehicles of interest with corresponding candidate vehicles (n) Note that, based on the above data, the evaluation value for each reference time is obtained by the function shown in the following equation (1). You may do it.

[0091]

[Equation 1]

In the equation (1), Σ is a process for obtaining the total amount in the vehicle of interest, and min is a process for comparing the evaluation values between the candidate vehicles.

Alternatively, the following equation (2) may be used instead of the above equation (1).

[0094]

[Equation 2]

In the equation (2), a and b are constants for adjusting numerical values between variables.

[Second Embodiment] Since the configuration of the road traffic control system in the second embodiment is the same as that in the first embodiment, the description thereof will not be repeated here. The road traffic control system according to the second embodiment is characterized in that more detailed matching is performed in consideration of the change in the rank of the vehicle of interest. That is, in the road traffic control system according to the second embodiment, the data described below is used as the data used for matching.

(1) Data relating to comparison between each vehicle of interest and candidate vehicle corresponding thereto Deviation of feature amount (x) Deviation from expected time (y) (2) Data considering the train of targeted vehicle Corresponding candidates Number of target vehicles in which vehicles exist (n) Time difference between target vehicles or difference in the number of vehicles (y
₂ ) Rank change of the vehicle of interest (z) FIG. 15 is a diagram for explaining the rank change of the vehicle of interest. In the figure, A to F indicate the vehicle of interest, and ∘ indicates a vehicle that is not the vehicle of interest. At the upstream point, attention vehicle A
It is assumed that the ranks of to F are the first to the sixth.

It is assumed that the target vehicle B is not detected at the downstream point, but the target vehicles A, D, C, E, and F are detected in this order. That is, the target vehicles A, C, D, E, F
The respective rankings are 1, 3, 2, 4, and 5. As a result, the rank change Z of the vehicle of interest is A, C, D, E, F.
0, 0, 2, 1, 1, respectively.

16 to 18 are flowcharts showing the processing performed by the road traffic control system according to the second embodiment. Differences between this flowchart and those shown in FIGS. 10 to 12 will be described below.

First, after the process of setting the passing time and the feature amount of the vehicle of interest in FIG. 16 (S209), the process of determining and setting the number of vehicles between the vehicle of interest (S210) is performed. This is specifically performed to obtain the time difference (y ₂ ) between the vehicles of interest.

When it is determined in step S247 in FIG. 18 that there is a time zone in which the evaluation value suddenly decreases at only one location, the processing from step S257 is performed. This is to perform more detailed matching while sequentially shifting the reference time in the time zone in which it is determined that the evaluation value suddenly decreases.

In step S257, the process of starting the matching is performed while shifting the reference time. Step S
At 259, the data of each target vehicle and each candidate vehicle is read.

In step S261, the process of calculating the evaluation value at the reference time is performed by the following equation (3) or (4).

[0104]

[Equation 3]

In the expressions (3) and (4), the variables a, b, c and d are constants for adjusting the numerical values among the variables.

In step S263, it is determined whether or not the processing has been completed for all candidate vehicles. If NO, the process proceeds to step S259, and if YES, step S259.
Proceed to 265.

In step S265, it is determined whether the processing has been completed for all reference times to be investigated,
If NO, the process returns to step S257, and if YES, the process proceeds to step S267.

In step S267, the travel time of the cycle is determined from the reference time with the smallest evaluation value.

As described above, in the present embodiment, more detailed matching is performed in the time zone in which the evaluation value suddenly decreases, so that the travel time can be determined more accurately.

In the present embodiment, detailed matching is performed only in the time zone when the evaluation value suddenly decreases, but detailed matching may be performed at all reference times.

[Third Embodiment] FIG. 19 is a diagram showing the structure of a road traffic control system according to a third embodiment of the present invention. Referring to the figure, in the present embodiment, traffic information collecting device 200 is a loop type vehicle detector (coil) 231a, 231b, 233 provided on a road.
a, 233b.

In this embodiment, the passing time of the vehicle is obtained at the upstream point and the downstream point by the loop type vehicle detector. Further, the vehicle speed and the vehicle length are required as the characteristic quantities of the vehicle.

For example, the loop type vehicle detector 233a or 233b is used as the first loop, the loop type vehicle detector 231.
When a or 231b is the second loop, the pulse shown in FIG. 20 is generated in the first or second loop when the vehicle passes the upstream point or the downstream point.

Time t ₁ shows the rising point of the pulse of the first loop, and time t ₂ shows the falling point. Also, at time t
₃ indicates the rising point of the pulse of the second loop, and time t ₄ indicates the falling point.

These times t _{1 to} t ₄ indicate that the vehicle is in the position as shown in FIG. That is, time t ₁ indicates the time when the vehicle V is approaching the first loop, and time t ₂ indicates the time when the vehicle V has passed the first loop. Further, time t ₃ indicates the time when the vehicle V approaches the second loop, and time t ₄ indicates the time when the vehicle V passes the second loop.

Here, let the diameter of the loop type vehicle sensor be r _l
And the length of the vehicle and l _v, and the distance between the first and second loop vehicle detectors and l _l, the velocity of the vehicle V and _{v, v = l l / (} t 3 -t 1 ) (7) l _v = {v (t ₂ −t ₁ ) −r _l } / 2 (8) holds. As a result, the vehicle speed and the vehicle length can be obtained as the characteristic amounts of the vehicle. Then, similarly to the first and second embodiments, the travel time can be calculated.

Since a plurality of feature amounts are used for matching in this embodiment, the feature amount shift (x)
It is necessary to use a vector as the value of. Further, the following formula (9) may be used instead of the formula (2), and the formula (10) may be used instead of the formula (4).

[0118]

[Equation 4]

In these equations, x ^T represents the transpose of the vector x. [Example of Actual Field Data] Below is shown an example of field data actually obtained at the upstream point and the downstream point when only the vehicle length is selected as a feature amount for reference.

FIG. 22 is a graph showing the relationship between the detection result of the feature amount and the time at the upstream point (in front of the Shibuya ramp on the Metropolitan Expressway Line 3), and FIG. 23 is at the downstream point (Roppongi on the Metropolitan Expressway Line 3). It is a graph which shows the relationship between a feature-value and time. Shibuya on-ramp, Shibuya off-ramp, and Takagi-machi on-ramp exist between these upstream points and downstream points. In addition, this data was obtained from the driving lane during heavy traffic on weekdays, and there are many data for large trucks. The distance between the upstream and downstream points is 4
It is a little less than km.

As can be seen from the figure, it is apparent that pattern matching of a vehicle having a long vehicle length (large vehicle) is easy at first glance. Further, in this case, it is considered that the travel time can be measured with high accuracy even if the detection accuracy of the feature amount is somewhat poor.

Originally, on an expressway in a city, a large vehicle does not change lanes except an off-ramp that goes down to a branch or an ordinary road, and the probability of overtaking other large vehicles is small. Further, it is considered that the number of lane changes is particularly small during heavy traffic congestion. Therefore, if the location is the same, it is considered that the detection accuracy of a large vehicle improves the accuracy of travel time measurement.

In addition, before the branching point of the expressway, before the off-ramp with a lot of outflow traffic, and before the on-ramp with a lot of inflow traffic, the lane changes are often disturbed and the traffic is unfavorable as a measurement point. Also, on holidays, the running probability of large vehicles is quite low, which is not desirable for measurement. In this case, it is conceivable to lengthen the unit of measurement, but if the detection accuracy of the feature quantity of the vehicle is good, or if there are multiple types of feature quantities to be adopted, even if such measurement conditions are bad It may be possible to detect travel time.

In the measurement on the intercity expressway, it is considered that the distance between the upstream point and the downstream point is long (about several kilometers to 10 km), and the configuration of the vehicle of interest is likely to change. However, on an inter-city expressway, there is at most one inflow and outflow, and it is considered that the probability of inflow and outflow is not so high, and the running probability of large vehicles is very high. It is considered that the travel time can be measured without any problem by adopting a logic that considers the driving characteristics of a large vehicle.

However, in the time zone when the traffic volume becomes large and the traffic lane or the overtaking lane starts to become congested, the lane change of the vehicle generally becomes large, and how much the large vehicle contributes to the travel It has some influence on the accuracy of time detection.
However, large vehicles are more likely to maintain their own lanes than ordinary vehicles and tend to return to their original lanes immediately after changing lanes. This is convenient for improving the detection accuracy of travel time.

Further, with respect to the ordinary road, in general, there is a large amount of vehicle inflow and outflow between the upstream point and the downstream point (assuming about 2 km). Therefore, the distribution of the traveling characteristics of how much a large-sized vehicle travels straight without turning right and left and the limit condition of the matching probability are related to the measurement accuracy.

[Fourth Embodiment] FIG. 24 is a diagram showing the structure of a road traffic control system according to a fourth embodiment of the present invention. Referring to the figure, in the present embodiment, traffic information collection device 200 is configured with cameras 251a and 251b provided on the road.

In this embodiment, the camera 251
The passing times of the vehicle are obtained at the upstream point and the downstream point by a and 251b. Further, vehicle length, vehicle width, vehicle height, vehicle speed, vehicle color, and other data are obtained as the characteristic amount of the vehicle.

On the basis of these plurality of characteristic quantities or a single characteristic quantity among them, vehicle group matching is performed as in the first and second embodiments, and the travel time is determined. To be done.

[Fifth Embodiment] FIG. 25 is a block diagram showing the structure of a road traffic control system according to the fifth embodiment of the present invention. Referring to the drawings, in the present embodiment, a voice recognition device is adopted as traffic information collecting device 200, as compared with the road traffic control system shown in FIG.

FIG. 26 is a diagram showing the structure of the road traffic control system in the present embodiment. Referring to the figure, the traffic information collecting device 200 is provided with a microphone 27 provided on a road R.
1a and 271b and primary processing devices (speech recognition devices) 272a and 272b that process voice from a microphone. The passing time of the vehicle is obtained at the upstream point and the downstream point from the voices from the microphones 271a and 271b. In addition, a voice when the vehicle passes is required as a feature amount of the vehicle.

In the traffic control center 100, the frequency characteristic (frequency spectrum) is recognized from the running sound of the vehicle (engine sound, wind noise, tire sound, etc.), and matching is performed using the frequency characteristic as a feature amount.

In the present embodiment, since the matching can be performed by the voice when the vehicle is running,
The travel time can be calculated accurately and efficiently.

[Sixth Embodiment] The external appearance of a road traffic control system according to the sixth embodiment is the same as that shown in FIG. In the present embodiment, an optical vehicle detector is used instead of the ultrasonic vehicle detectors 201a and 201b of FIG. The optical vehicle detector determines the passing time of the vehicle at the upstream point and the downstream point.

Further, the vehicle height is obtained as the characteristic amount of the vehicle. Matching is performed based on this feature amount.

[Seventh Embodiment] The external appearance of a road traffic control system according to the seventh embodiment is the same as that shown in FIG. In the present embodiment, a microwave type vehicle detector is used instead of the ultrasonic type vehicle detectors 201a and 201b in FIG.

The microwave type vehicle detector determines the passing time of the vehicle at the upstream point and the downstream point.

Further, the vehicle speed and the vehicle height obtained by the Doppler effect can be obtained as the characteristic amount of the vehicle. Also, as with the case of using the loop type vehicle detector,
The vehicle length can also be obtained as a feature amount. In other words, the output of the vehicle detector turned off when the output of the vehicle detector turned on because the front of the vehicle approached the vehicle detector, and the output of the vehicle detector turned off when the rear part of the vehicle passed the vehicle detector. The time until time is obtained, and the vehicle length is obtained from the time and the vehicle speed.

[Eighth Embodiment] In the above-described embodiment, the travel time is calculated by matching the information of a group of vehicles. However, the vehicle license plate is used at the upstream point and the downstream point. By identifying, the time (travel time) that the vehicle travels from the upstream point to the downstream point may be calculated (license plate matching).

[Ninth Embodiment] As shown in FIG. 27, the vehicle V to the ground systems 201a and 201b are connected.
The travel time may be calculated based on the uplink information sent to. That is, using a device capable of two-way communication such as an optical beacon between the vehicle and the ground system, measure the time when the uplink information from the vehicle was obtained at the upstream point and the downstream point, and calculate the travel time from the time difference. It may be calculated.

[Modification of the Embodiment] In the above-described embodiment, the passing vehicle at the upstream point is used as a reference for matching with the passing vehicle at the downstream point. Instead, conversely, the passing vehicle at the downstream point may be used as a reference to match the past passing vehicle at the upstream point. By doing so, it is not necessary to wait for the vehicle to be detected at the downstream point, and real-time processing becomes possible.

In addition, in the third and fourth embodiments, the vehicle length is detected as the feature amount. However, in order to improve the performance of the matching process, the matching target is set to a vehicle length with a relatively low frequency of appearance. It is efficient to limit to the vehicles that have. Specifically, a vehicle having a vehicle length equal to or greater than a certain value (for example, a large truck) or a vehicle having a vehicle length less than or equal to a certain value (such as a light vehicle) may be a vehicle of interest.

When the device is activated,
The range of the reference time to be evaluated is set to be sufficiently large, and the travel time measured in the previous cycle and the maximum value of the travel time that fluctuates in one cycle are used as a reference to determine the reference time to be evaluated. The range may be narrowed.

The reliability of the calculated travel time (variation error from the average) may be calculated based on the evaluation value.

Further, the cycle of measuring the travel time may be changed in response to changes in traffic conditions such as congestion and non-congestion.

Further, the cycle of measuring the travel time may be changed by the ratio of the number of passing vehicles of interest to the total number of vehicles.

In addition, the travel time is measured on the target road by the number of vehicles flowing into the road to be measured between the upstream point and the downstream point or the number of vehicles flowing out of the target road. You may make it change with the ratio with respect to the number of vehicles.

The cycle of measuring the travel time may be changed according to the distance between the upstream point and the downstream point.

Furthermore, the processing described in the first to ninth embodiments may be performed for each lane of the road to be measured.

In the above-mentioned road traffic control system, the information obtained from the traffic information collecting device 200 is transmitted to the traffic control center 100, and the traffic control center 10
Although the travel time and the like are measured at 0, information may be exchanged between the traffic information collecting devices 200 and the travel time may be measured by the traffic information collecting device 200. Further, the processing may be performed by a distributed processing device separately installed on the road.

The information may be transmitted by using a dedicated wired line, a public line or a wireless line.

When using a plurality of types of feature amounts for matching, each feature amount may be weighted.

Although the deviation from the expected time is used for matching in the embodiment, the deviation of the number of vehicles from the expected time may be used instead.

In addition, for example, both the ultrasonic type sensor and the optical type vehicle detector can determine the vehicle height. Also,
Both the loop type sensor and the microwave type vehicle detector can determine the vehicle length. Further, the camera (image input device) can determine the vehicle height and the vehicle length. In this way, different sensors can collect the same feature amount, so that matching can be performed even if different sensors are used at the upstream point and the downstream point.

The reliability of the calculated travel time may be calculated. This is the end of matching,
When the travel time is finally obtained, the reliability of the travel time is also obtained at the same time. For example, if the travel time finally obtained is Td and the time difference (or the number of vehicles) of each vehicle with respect to it is y _i , the reliability V (y) corresponds to the variation, and the following equation (11) ).

[0156]

[Equation 5]

In the equation (11), n indicates the number of vehicles. The smaller the reliability obtained by the equation (11), the higher the reliability is (reliable).

[Effects of the Embodiment] The above-described embodiments have the following advantageous effects as compared with the related art.

(1) In the case where there is a large amount of inflow / outflow traffic or overtaking because the matching evaluation criteria for the entire vehicle group are set based on the ambiguous characteristic amount of each vehicle in the vehicle group of the upstream point and the downstream point. However, it becomes possible to associate vehicle groups. Further, even if the matching of the individual vehicles cannot be determined, highly accurate measurement values can be obtained in terms of average travel time.

(2) Since the characteristic amount is not a rough numerical value such as a large vehicle type or a small vehicle type, but an analog numerical value, the measurement accuracy can be improved.

(3) Since the evaluation criteria mainly consider the deviation of the feature amount and the deviation from the reference time, highly accurate measurement is possible.

(4) To enhance the matching effect,
Vehicles that have a high frequency of appearance, such as feature quantities of ordinary vehicles, are excluded from the matching target, and therefore accurate measurement is possible.

(5) If the reliability of the travel time is calculated along with the measurement of the travel time, the value obtained by the system can be used easily.

(6) By using the image processing apparatus as in the fifth embodiment, a plurality of ambiguous feature quantities can be extracted. Then, by performing matching as the entire vehicle group, it is possible to further improve the measurement accuracy.

[Tenth Embodiment] The road traffic control system in this embodiment uses the travel time data in a predetermined section obtained as in the first to ninth embodiments, Track the vehicle.

FIG. 28 is a diagram showing the structure of the road traffic control system in the present embodiment. Referring to the figure, a road traffic control system includes a traffic control center 100, a traffic information collecting device 200, a traffic information providing device 300,
The configuration including the CRT 400 is the same as in the first to fifth embodiments.

However, in the present embodiment, the road traffic control system is provided with an input device 401 for inputting data on a vehicle (specific vehicle) to be tracked.

For example, when the input device 401 is composed of a mouse or a digitizer and the traffic information collecting device 200 includes a camera, the CRT 400 displays a road and an image of a vehicle passing on the road, and traces from the road. The operator can instruct which vehicle he / she wants.

Further, the operator inputs the characteristic amount of the vehicle (vehicle height, vehicle length, vehicle width, color, shade, balance of vehicle color or shade, one or a plurality of patterns, shapes) by the input device 401. Thus, the vehicle can be tracked.

In addition, the data of a plurality of feature quantities is converted into CRT40.
Alternatively, the operator may indicate the characteristic amount of the vehicle to be tracked, and the characteristic amount of the vehicle may be displayed. In addition, the data of the characteristic amount of the vehicle obtained by the traffic information collecting device 200 is C
It may be displayed on the RT400 and the operator may be instructed from the feature quantity of the vehicle to be traced.

Further, when the vehicle is displayed on the CRT 400, the characteristic amount of the vehicle may be displayed on the image. At this time, when the time for searching for the vehicle is limited, it is effective to display the feature amount only within that time.

Further, the image of the vehicle and the feature amount may be displayed in an overlapping manner. At this time, it is effective to allow the image to stand still based on the input from the input device 401.

Further, when the tracking of the specific vehicle is wrong, the operator can change the specific vehicle by using the input device 401.

Further, when there are a plurality of vehicles having the specified characteristic amount, the operator can reduce the number by the input device.

FIG. 29 is a plan view showing the structure of the traffic information collecting device. Referring to the figure, a plurality of locations P1 to R1 of the road R
The traffic information collecting devices 200a to 200f are provided at P6, and the characteristic amount of the vehicle is detected at that point.

Further, the travel times T1 to T5 between the respective traffic information collecting devices are obtained on the basis of this characteristic amount as in the first to ninth embodiments.

In each of the calculated travel times T1 to T5, the time t _{0 for} detecting the vehicle is calculated according to the reliability of the travel time. For example, if a particular vehicle is detected by the traffic information collecting device 200a, the detection of the specific vehicle by the traffic information collecting device 200b is started when the detected time elapsed time T1-t ₀ from the time T1 + t _The detection by the traffic information collection device 200b is continued until it becomes ₀ .

If the traffic information collecting device 200b detects a specific vehicle, the traffic information collecting device 200c similarly detects the specific vehicle.

By doing so, the specific vehicle can be traced in order from the upstream point to the downstream point of the road.

30 and 31 are flow charts showing the processing of the road traffic control system when a camera is used as the traffic information collecting apparatus 200 as in the fourth embodiment.

Referring to the figure, initial setting is performed in step S301. In this initial setting, the operator inputs the feature amount of the vehicle (specific vehicle) to be tracked via the input device 401. Alternatively, the vehicle to be the specific vehicle may be selected from the image of the vehicle displayed on the CRT.

In step S303, it is determined whether the input of the characteristic amount is based on the image data (for example, based on the selection of the displayed vehicle). When it is determined in step S305 that the input is image data, the feature amount of the vehicle is recognized from the image data in step S307.

In step S309, the input characteristic amount is transmitted to each of the traffic information collecting devices 200a to 200f. Each traffic information collection device starts detection of passage of a vehicle having the input characteristic amount.

By referring to the time in step S311, it is determined whether or not the current timing is the cycle of processing. If the timing is correct, in step S313, each of the traffic information collecting devices 200a to 20a
Time-series data of the characteristic amount of the vehicle is received from 0f.

In step S315, the travel times T1 to T5 between the traffic information collecting devices and the reliability of the travel times are calculated.

In step S317, it is determined whether or not the specific vehicle is already being tracked. If YES, step S333.
If NO, the process proceeds to step S319.

[0187] In step S319, it is determined whether there is a notification from the traffic information collecting device that has detected the specific vehicle. If there is a contact in step S321,
Proceed to step S323, and if there is no contact, step S323.
Return to 311.

At step S323, the travel time and reliability from the contacted traffic information collecting device to the traffic information collecting device at the downstream point are referred to. Then, in step S325, the arrival time of the specific vehicle in the traffic information collecting device at the downstream point is predicted.

[0189] In step S327, the particular vehicle is the time range t ₀ expected is determined based on the reliability of the travel time to pass through the traffic information collection device.

In step S329, the traffic information collecting device is installed.
In addition, the feature amount of the tracking vehicle and the time range t ₀And send. So
Then, in step S331, it is determined that the current state is tracking the vehicle.
To do. Then, it returns to step S311. The traffic information
In the collecting device, the time range t₀Within the
Departure is done.

If YES in step S317, the flow advances to step S333 to check the video data of the traffic information collecting device on the downstream side. Here, the video data is sent from the traffic information collecting device to the traffic control center 100, and the
It is assumed that the feature amount 400 is displayed by being superimposed on the image. In addition, CRT40
It is displayed with a mark on 0.

In step S335, the operator determines whether or not the specific vehicle is correctly detected. If NO, in step S337 the input device 401 is used to correct the detection error.
Or add a specific vehicle.

In step S339, it is determined whether or not the tracking of the specific vehicle is defective, and if NO, it is determined in step S341 whether the tracking is completed. If NO, the travel time and reliability to the traffic information collecting device on the further downstream side are referred to in step S343, and the arrival time of the specific vehicle in the traffic information collecting device at the downstream point is predicted in step S345.

In step S347, the time range t ₀ passing through the traffic information collecting device at the downstream point is determined, and the time range t ₀ is determined.
At 349, the characteristic amount of the specific vehicle and the time range t ₀ are transmitted to the traffic information collecting device at the downstream point. Then, step S31
Return to 1.

If YES in step S339 or S341, the tracking state is canceled in step S351, and the process returns to step S311.

[Eleventh Embodiment] The device configuration of the road traffic control system according to the eleventh embodiment is the tenth embodiment.
Since it is similar to that of the embodiment, description thereof will not be repeated here.

In the tenth embodiment, the operator has designated the specific vehicle, but in the eleventh embodiment, the specific vehicle is automatically determined by the device. Specifically, step S301 in FIG.
At, the vehicle to be tracked is automatically determined based on the data obtained from each traffic information collecting device.

Here, for example, the following vehicles are determined as vehicles to be tracked. (1) Vehicles with a speed equal to or higher than a certain value (2) Vehicles judged to have recklessly changed lanes by image processing (3) Vehicles judged to ignore signals by image processing (4) Image processing A vehicle that is considered to have violated parking If the vehicle to be tracked is automatically determined as in the present embodiment, the burden on the operator is reduced. Further, the tenth embodiment and the eleventh embodiment may be combined to automatically determine the specific vehicle and input the specific vehicle from the operator at the same time.

When a vehicle whose speed is equal to or higher than a certain value is tracked (vehicle which is in violation of speed), the downstream of the vehicle to be tracked is based on the average speed of the vehicles in the vicinity and the speed ratio of the vehicle to be tracked. The estimated time of arrival at the point may be changed.

[Effects of Tenth and Eleventh Embodiments] As described above, in the tenth and eleventh embodiments, a loop type or ultrasonic vehicle detector which is widely used in road networks nationwide. Alternatively, it is possible to construct a system, which is slightly more expensive than that, but cheaper than using a license plate reading device, by using an image processing device capable of confirming various traffic conditions.

Further, by using the time-series data of the ambiguous feature amount of each vehicle and narrowing the time range in which the specific vehicle arrives at the downstream device based on the travel time between the devices and its reliability, By reducing the number of vehicles similar to the specific vehicle as much as possible, it is possible to configure a system with a high probability of detecting the specific vehicle.

Further, by displaying the feature amount and image data of all the vehicles in the time range in a superimposed manner on the CRT (display device) of the control system, it is possible to confirm whether or not the tracking is properly performed, It is possible to reduce the number of vehicles when there are vehicle tracking candidates, or to correct this with man-machine when the tracking is wrong.

Further, the specific vehicle can be set in real time by using the input device 401 or the like, and the system side can automatically detect the vehicle to be tracked such as a runaway vehicle.

A plurality of feature quantities may be obtained by using a plurality of traffic information collection devices according to the first to seventh embodiments. In addition, the feature quantity is individually communicated (two-way beacon, ET
C)). Furthermore, one or more of license plate, vehicle shape, pattern, color, color balance, vehicle shade, and shade balance are extracted from the image data,
These may be used as the characteristic amount.

[Twelfth Embodiment] FIG. 32 is a block diagram showing the structure of a traffic information management apparatus according to the twelfth embodiment of the present invention.

This traffic information management device, like the tenth and eleventh embodiments, detects a vehicle detected at the upstream point at the downstream point.

Referring to the figure, the traffic information management device includes a travel time estimation unit 701 for estimating travel time between an upstream point and a downstream point, and a travel time estimated by the travel time estimation unit 701. A travel time reliability estimation unit 703 for estimating reliability, and an arrival time estimation unit for estimating time at which a vehicle detected at an upstream point is estimated to reach a downstream point based on the estimated travel time and reliability. 705,
The specific vehicle extraction unit 70 that detects the vehicle detected at the upstream point at the downstream point based on the estimated time.
7, a specific vehicle output unit 709 that displays the detected vehicle on a display or the like, and a specific vehicle designation unit 711 that designates the vehicle to be detected.

In estimating the travel time in the travel time estimating unit 701, as described in the first to seventh embodiments, matching between a plurality of vehicles (group matching) is performed.
May be performed, license plate matching may be performed as in the eighth embodiment, or travel time may be estimated based on the uplink information as in the ninth embodiment. Good.

When the travel time is calculated using the license plate or the uplink as in the eighth and ninth embodiments, the travel time reliability estimation unit 703 calculates the reliability based on the variation of the travel time. It will be estimated.

The specific vehicle extraction unit 707 detects the vehicle length, vehicle width, vehicle height, brightness (vertical and horizontal) of the vehicle detected at the upstream point,
A vehicle is detected at a downstream point based on one or more data of hue (vertical and horizontal).

The specific vehicle designation unit 711 may specify the vehicle manually or automatically.

For example, as a method of manually specifying a vehicle to be detected, a method of letting the user specify the numerical values of the vehicle length, vehicle width, and vehicle height, a method of clicking the vehicle in the image with the mouse, How to specify (for example, by inputting the name and color of a specific vehicle type, or by inputting the classification of large vehicles, trucks, buses, etc.) ・ Identify the vehicle by having the user scan a photo The method can be considered.

As a method of automatically identifying a vehicle, a method of identifying a vehicle in which a speed violation is detected by an orbis or an image sensor, or a vehicle parked by an image sensor (especially a vehicle parked for a long time) Is a vehicle to be detected, and a method in which a vehicle in which reckless lane change is detected is detected (particularly by detecting the speed of the vehicle in the lateral direction) is included. .

The arrival time estimation unit 705 estimates the time at which the vehicle detected at the upstream point passes the sensor at the downstream point from the travel time and the reliability. For example, when the time at which the sensor passes through the upstream point is 9:00, the travel time from the upstream point to the downstream point is 20 minutes, and the reliability (error) of the travel time is ± 2 minutes. Is to estimate the arrival time of the vehicle to the downstream point from 9:18 to 9:22. At this time, the specific vehicle extraction unit 707
Detects the vehicle only within the time range from 9:18 to 9:22.

FIG. 33 is a flow chart showing the processing of the specific vehicle extraction unit 707. In the following description, the upstream point is point A, the downstream point is point B, and at time t
Vehicle a _1, a _2, which has passed the point, ..., and a _n, a time t + delta
vehicle b ₁ , which has passed point B at t (Δt is travel time),
It is assumed that b ₂ , ..., B _m are matched (corresponding).

Referring to FIG. 33, at the points A and B respectively, the vehicle detector detects the size of the vehicle,
A luminance distribution and a hue distribution are obtained.

Here, the vehicle length, the vehicle width, and the vehicle height are obtained as the size of the vehicle. Car obtained by the A point length, vehicle width and vehicle height, respectively, and _{l a, w a, h a} ,
The vehicle length, vehicle width, and vehicle height obtained at point B, respectively,
Let l _b , w _b , h _b .

The luminance distribution and the hue distribution are obtained in the vertical direction (traveling direction) and in the lateral direction (direction orthogonal to the traveling direction) of the vehicle.

The evaluation value p ₁ is obtained from the size of the vehicle obtained at the points A and B.

The luminance distribution and the hue distribution of the vehicle obtained at the points A and B are normalized. Based on the normalized luminance distribution, the evaluation value (p ₂ , p ₃ or p
₄ ) is calculated. An evaluation value (p ₅ or p ₆ ) is calculated based on the normalized hue distribution.

From the obtained evaluation value, it is judged which vehicle the vehicle passing through the point A hits at the point B (match judgment).
Is performed.

Hereinafter, how to obtain each of the evaluation values p _{1 to} p ₆ will be described. (1) Method of obtaining evaluation value based on size The difference in size between the vehicles a _i and b _j is defined by the weighted sum as in Expression (12).

[0223] _{p 1 = α | l a -l} b | + β | w a -w b | + γ | h a -h b | ... (1 2) In this case, α, β, γ, respectively vehicle length, vehicle width And vehicle weight, and when an image type vehicle detector is adopted, α ≧ β ≧ γ ≧ 0. This is because, in general, the vehicle length is easy to obtain and therefore highly reliable, and conversely, the vehicle height is difficult to obtain, and therefore highly reliable. If γ = 0, the vehicle height will not be used for size determination.

That is, when only a part of the vehicle length, the vehicle width, and the vehicle height can be obtained, the weight is set to 0 for the items that cannot be obtained. For example, when the vehicle detector is a loop type in which only the vehicle length is required, β = γ = 0. Similarly, in the case of the ultrasonic type where only the vehicle height is required,
Let α = β = 0.

(2) How to obtain evaluation value based on luminance distribution When the vehicle detector is an image type, matching can be performed using the luminance distribution of the vehicle. When the vehicle detector is a color camera, RGB is H (hue), S (saturation), I
(Brightness) At this time, the same vehicle may appear different in size on the image, depending on the position and angle of the camera used for shooting. Therefore, normalization is performed so that the sizes on the image are the same. In the following explanation,
The brightness of the vehicle at the upstream point (point A) after normalization is U ₁ ,
U _2, U _3, ..., and U _N, D ₁ brightness of the vehicle at a point downstream (B _{point), D 2, D 3,} ..., is set to D _N.

The following expression (13) shows a method for obtaining the evaluation value p ₂ by using the variance of the brightness difference, and the expression (14)
Indicates a method of obtaining the evaluation value p ₃ using correlation,
Expression (15) shows a method of obtaining the evaluation value p ₄ using the sum of the absolute values of the brightness differences. Any of these methods may be adopted.

[0227]

[Equation 6]

In the expression (13), the smaller the evaluation value is, the higher the degree of coincidence is, and p ₂ = 0 when the luminance is completely coincident. That is, there is a relation of p ₂ > 0. Further, in the equation (13), μ _S is the average of δ.

In the equation (14), there is a relation of 1> p ₃ > -1. Incidentally, the expression of p ₃ - although code is attached, which is so that it is p ₃ smaller high degree of coincidence, it is assumed that by reversing the sign.

In the equation (15), there is a relation of p ₄ > 0, and the smaller p ₄ is, the higher the degree of agreement is.

(3) How to obtain evaluation value based on hue When a color camera is used, hue (0 ° ≦ H ≦ 36
Regarding 0 °, the same things as in formulas (13) to (15) can be considered. However, it is desirable to target only flashy vehicles because the reliability is low in the case of a nearly achromatic (simple) vehicle. Specifically, the determination based on the hue may be performed only on the vehicle having the saturation (0 ≦ S ≦ 1) of 0.1 to 0.2 or more.

When the evaluation value is obtained by using the variance of the hue difference, the following equation (16) may be used. When the evaluation value is obtained by using the sum of hue angles, the equation (17) is used.
Can be used. Also in these expressions, the smaller the evaluation value, the higher the degree of agreement.

[0233]

[Equation 7]

The following methods are conceivable as methods for finally determining whether or not the vehicle is the same vehicle by using these evaluation values after obtaining each evaluation value. In practice, it is effective to consider the vehicles with the highest degree of coincidence as the same vehicle in order based on the weighted sum (or points) of the evaluation values and exclude them from the candidates. If such exclusion is performed, only combinations with a low degree of coincidence remain, but if only combinations with an evaluation value equal to or higher than a certain value (combinations with a low degree of coincidence) remain, it may be considered that correspondence cannot be performed.

(1) Method using weighted sum of evaluation values The following expression (18) is obtained m × n times, and each vehicle a
Vehicle b _j (1 ≦ 1) for which P is minimum for _i (1 ≦ i ≦ n)
j ≦ m) is regarded as the same vehicle. However, in Expression (18), q is a weight term, and q ≧ 0.

[0236]

[Equation 8]

(2) Method Using Weighted Sum of Points Referring to FIG. 34 (a), n × m (here, n = m =
4), and as shown in (b), 3 points, 2 points, and 1 point are given in order from the highest matching degree (lowest evaluation value). This point is calculated for all evaluation values, the total of the points is calculated, and the vehicle with the highest point is regarded as the same vehicle.

When the evaluation values are summed up with reference to the bottom table of FIG. 34 (a), vehicles n = 1, 2, 3, 4 at the upstream point and vehicles m = 1 at the downstream point. ，
2, 3, and 4 respectively correspond (the evaluation value is the lowest).

In the table in the center of FIG. 34 (a), there are two vehicles with an evaluation value of "3" among the vehicles with n = 1 at the upstream point.
Since there are vehicles, referring to the table in the center of FIG. 34B, 1.5 points, which is the average of 2 points and 1 point, is given to each vehicle.

In the total points of FIG. 34B, the point in the combination of the vehicle of n = 1 at the upstream point and the vehicle of m = 2 at the downstream point is "5", and the vehicle at n = 1, m = 1. The total number of points is larger than “4.5”, but in the vehicle of n = 2 at the upstream point, the point of combination with m = 2 is larger than that, so m = 1 and n = 1 should be associated. , M = 1 corresponds to n = 1.

Further, when the evaluation values (or points or the like) are equal or very close to each other and the combination cannot be limited to one combination, a plurality of candidates may be manually selected.

Next, referring to FIGS. 35 to 40, a concrete result of matching the vehicles at the upstream point (FIGS. 35 to 37) and the downstream point (FIGS. 38 to 40) will be described.

In FIG. 35, the truck (specific vehicle) at the lower right of the image is detected at the downstream point. Also in FIG. 36, the lower right track of the image is detected at the downstream point. In FIG. 37, the tank truck located closest to the center lane is detected at the downstream point.

In FIG. 38, the track of FIG. 35 is shown in the lower right part of the image. In FIG. 39, the track of FIG. 36 is shown in the lower right part of the image. In FIG. 40, the tank truck of FIG. 37 is shown in front of the center lane.

Images of roof portions of these trucks and tank trucks were photographed and matching was performed.

Referring to FIG. 41, the average value of the luminance of pixels in the horizontal direction of the image of the roof portion of the truck or tank truck is taken, and the vertical luminance distribution of the average value is used as vertical luminance for matching data. ing.

42 to 44 are graphs showing the results of matching. The value obtained by multiplying each horizontal axis value by 5 cm is the vertical position in the vehicle. The vertical axis represents luminance. 42 to 44, the data before normalization is used. 42 to 44, the thick line in the graph indicates the vehicle brightness at the downstream point, and the thin line indicates the vehicle brightness at the upstream point.

FIG. 42 is a comparison result of the trucks of FIGS. 35 and 38, FIG. 43 is a comparison result of the trucks of FIGS. 36 and 39, and FIG. 44 is a comparison result of the tank truck of FIGS. 37 and 40. is there.

FIGS. 45 to 47 are graphs corresponding to FIGS. 42 to 44 and show the states in which matching is performed on the basis of the data after the normalization. By normalization, the size of the image data (horizontal axis of the graph) is unified to “600” at the upstream point and the downstream point.

As can be seen from FIGS. 45 to 47, if the same vehicle is used at the upstream point and the downstream point, the brightness values are substantially the same. This facilitates detection of the vehicle detected at the upstream point at the downstream point.

As for the roof data of the vehicle shown in FIGS. 35 to 40, as shown in FIG. 48, the luminance of the pixels in the vertical direction is first averaged, and the average value is taken as the luminance in the horizontal direction (the luminance in the horizontal direction). The results obtained by making the determination are shown below. 49 to 51 are diagrams showing the data after the normalization, and are diagrams corresponding to FIGS. 45 to 47, respectively.

As can be seen from these figures, it is easy to perform vehicle matching using the lateral luminance distribution. In particular, in FIG. 50, a valley is formed in the central portion of the graph, but this is because there is a vertical line in the roof portion of the truck in FIGS. 36 and 39, and this portion is located at the upstream point and the downstream point. It shows that we are doing.

Although the vehicle detected at the downstream point is output by the specific vehicle output unit 709 of FIG. 32, the output may be performed by the following method.

(1) A method of outputting a sensor number, a route name, a position of the sensor, a detected vehicle (specific vehicle) number and a feature amount by text display or voice.

For example, "a vehicle similar to the specific vehicle number 3 passes through the detector 12 at the launch intersection of National Route 2; vehicle length: 10.5 m, vehicle width: 2.2 m, vehicle height: 2.6 m", etc. Is displayed and voice output is performed.

(2) When the sensor is of the image type and the image taken by the specific vehicle output unit 709 can be displayed, the specific vehicle is highlighted by computer graphics as shown in FIG. Alternatively, the feature amount of the specific vehicle may be displayed like the characters in the balloon of FIG.

Of course, if there are a plurality of candidates for the specific vehicle or if the output specific vehicle is incorrect, the user (controller) may manually select or correct.

The embodiments disclosed this time are to be considered as illustrative in all points and not restrictive. The scope of the present invention is shown not by the above description but by the claims, and is intended to include meanings equivalent to the claims and all modifications within the scope.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a road traffic control system according to a first embodiment of the present invention.

FIG. 2 is a first diagram showing specific examples of upstream points and downstream points.

FIG. 3 is a second diagram showing specific examples of upstream points and downstream points.

FIG. 4 is a third diagram showing specific examples of upstream points and downstream points.

FIG. 5 is a diagram showing a configuration of a road traffic control system according to the first embodiment.

FIG. 6 is a diagram showing a feature amount of a vehicle group obtained at an upstream point.

FIG. 7 is a diagram showing a feature amount of a vehicle group obtained at a downstream point.

FIG. 8 is a diagram for explaining a relationship between a vehicle of interest and a candidate vehicle at an upstream point and a downstream point.

FIG. 9 is a diagram for explaining how to shift the reference time.

FIG. 10 is a flowchart showing a process performed by the road traffic control system according to the first embodiment.

11 is a flowchart following FIG.

FIG. 12 is a flowchart following FIG. 11.

FIG. 13 is a diagram showing a result of matching when a valley of evaluation values exists.

FIG. 14 is a diagram showing a result of matching in which an evaluation value cannot be adopted.

FIG. 15 is a diagram for explaining the processing of the road traffic control system in the second embodiment.

FIG. 16 is a flowchart showing a process performed by the road traffic control system according to the second embodiment.

FIG. 17 is a flowchart following FIG.

FIG. 18 is a flowchart following FIG.

FIG. 19 is a diagram showing a configuration of a road traffic control system according to a third embodiment of the present invention.

FIG. 20 is a diagram showing pulses obtained with the system of FIG.

21 is a diagram showing the relationship between the pulse of FIG. 20 and the position of the vehicle.

FIG. 22 is a diagram showing a measurement result of a vehicle length (upstream point).

FIG. 23 is a diagram showing a measurement result of a vehicle length (downstream point).

FIG. 24 is a diagram showing a configuration of a road traffic control system according to a fourth embodiment of the present invention.

FIG. 25 is a block diagram showing a configuration of a road traffic control system according to a fifth embodiment of the present invention.

FIG. 26 is a diagram showing a configuration of a road traffic control system according to a fifth embodiment.

FIG. 27 is a diagram showing a configuration of a road traffic control system according to a ninth embodiment.

FIG. 28 is a diagram showing a configuration of a road traffic control system in the tenth embodiment.

FIG. 29 is a diagram showing a configuration of a road traffic control system in the tenth embodiment.

FIG. 30 is a flowchart showing a process performed by the road traffic control system in the tenth embodiment.

FIG. 31 is a flowchart following FIG. 30.

FIG. 32 is a block diagram showing the configuration of a road traffic control system in the twelfth embodiment.

33 is a flowchart showing a process performed by the specific vehicle extraction unit 707 of FIG. 32.

FIG. 34 is a diagram showing a method of matching determination based on an evaluation value.

FIG. 35 is a first diagram showing an image captured at an upstream point
FIG.

FIG. 36 is a second showing an image taken at an upstream point
FIG.

FIG. 37 is a third view showing an image taken at an upstream point.
FIG.

FIG. 38 is a first showing an image taken at a downstream point.
FIG.

FIG. 39 is a second showing an image taken at a downstream point
FIG.

FIG. 40 shows a third image taken at a downstream point.
FIG.

FIG. 41 is a diagram showing how to obtain vertical luminance.

FIG. 42 is a diagram showing a result of vehicle matching shown in FIGS. 35 and 38;

FIG. 43 is a diagram showing a result of vehicle matching shown in FIGS. 36 and 39;

44 is a diagram showing a result of vehicle matching shown in FIGS. 37 and 40. FIG.

45 is a diagram showing a result of matching using the normalized data, and a diagram corresponding to FIG. 42. FIG.

FIG. 46 is a diagram showing a result of matching using the normalized data, and a diagram corresponding to FIG. 43.

FIG. 47 is a diagram showing a result of matching using the normalized data, and a diagram corresponding to FIG. 44.

FIG. 48 is a diagram for explaining how to obtain horizontal luminance.

FIG. 49 is a diagram showing a result of vehicle matching shown in FIGS. 35 and 38 according to lateral luminance.

FIG. 50 is a diagram showing a result of vehicle matching shown in FIGS. 36 and 39 according to lateral luminance.

51 is a diagram showing a result of vehicle matching shown in FIGS. 37 and 40 based on lateral luminance. FIG.

FIG. 52 is a diagram showing a specific example of an output method of a specific vehicle.

[Explanation of symbols]

100 Traffic Control Center 200 Traffic Information Collection Devices 201a, 201b Ultrasonic Vehicle Detectors 231a, 231b, 233a, 233b Loop Vehicle Detectors 251a, 251b Cameras (Image Processing Devices) 271a, 271b Microphones

Continuation of front page (56) Reference JP-A-7-262488 (JP, A) JP-A-11-39587 (JP, A) JP-A-9-91588 (JP, A) JP-A-9-167296 (JP , A) (58) Fields surveyed (Int.Cl. ⁷ , DB name) G08G 1/01 G07C 9/00 G08G 1/015 G08G 1/017 G08G 1/04

Claims (14)

(57) [Claims] 1. A first collecting means installed at an upstream point of a road for collecting a characteristic amount of a passing vehicle and a passing time of the vehicle, and a characteristic amount of a vehicle passing at a downstream point of the road and Second collecting means for collecting the passage time of the vehicle; acquiring means for acquiring a reference time estimated to be travel time between the upstream point and the downstream point; and the acquired reference time and Based on the passing time of the vehicle collected by the first collecting means, a predicting means for predicting a time at which the vehicle passing through the upstream point reaches the downstream point, and a collecting means collected by the second collecting means. The first of the feature quantity
The deviation from the feature quantity collected by the collecting means of
Information of the group of vehicles collected by the first collecting means and the second collecting means based on the deviation of the passing time of the vehicle collected by the collecting means from the predicted time. Matching means for performing matching with the information of the group of vehicles that have been set, calculation means for calculating travel time between the upstream point and the downstream point based on the result of the matching, and reliability of the travel time. Reliability calculating means for calculating, estimating means for estimating the time at which the vehicle detected at the upstream point arrives at the downstream point, based on the travel time and the reliability, the estimated based on the time, and detecting means for detecting a vehicle detected at said point upstream in the downstream point in the matching unit, by said second collecting means Income
Deviation of the passing time of the collected vehicles from the predicted time
A traffic information management device that calculates an evaluation value based on the amount of traffic. The method according to claim 2 wherein each of the first and second collection means includes a loop sensor, the identifying means for identifying a feature value of the car length of the vehicle based on an output of the sensor, according to claim 1 Traffic information management device described in. 3. The traffic information management device according to claim 1 , wherein each of the first and second collecting means includes an ultrasonic sensor. Wherein each of said first and second collection means includes a sensor for collecting running noise of the vehicle, according to claim 1
Traffic information management device described in. 5. Each of the first and second collecting means includes a camera for obtaining an image of a vehicle, and a group of a vehicle width, a vehicle height, a vehicle length, a vehicle color, a luminance and a pattern from the image. The traffic information management device according to claim 1 , further comprising an image processing device that extracts a single or a plurality of selected feature amounts. 6. The traffic information management device according to claim 1 , wherein each of the first and second collection means includes an optical vehicle detector. 7. The traffic information management device according to claim 1 , wherein each of the first and second collecting means includes a microwave type vehicle detector. Wherein said first and second collection means collects feature amounts of a plurality of types, the matching means performs the matching by performing the weighting of the plurality of types of characteristic amounts, of claims 1 to 7 The traffic information management device according to any one. Wherein said matching means, by determining the number displacement of the vehicle, determining a deviation from the predicted time of passing time of the vehicle collected by said second collecting means, according to claim 1 The traffic information management device according to any one of 1 to 8 . Wherein said matching means, utilizes only passage time of the feature and the vehicle in the vehicle that meet certain criteria for matching, the traffic information management apparatus according to any one of claims 1 to 9. Wherein said first and second collection means collects only passage time of the feature and the vehicle in the vehicle that meet certain criteria, the traffic information management according to any one of claims 1 9 apparatus. 12. A first system installed at an upstream point of a road, for collecting a characteristic amount of a passing vehicle and a passing time of the vehicle.
And collecting means, disposed downstream point of the road, and a second collection means for collecting features and passing time of the vehicle passing vehicles, of the one vehicle collected by the first collection means The difference between the characteristic amount and the characteristic amount of one vehicle collected by the second collecting means, the passing time of the one vehicle collected by the first collecting means, and the second collecting and passing time of one vehicle collected by means, on the basis of the reference time becomes the estimated travel time between these two points, a plurality
The evaluation value is calculated for each of the
Matching means for matching the group of vehicle information collected by the first collecting means with the group of vehicle information collected by the second collecting means based on the plurality of evaluated values obtained ; Based on the result of matching, a calculating means for calculating a travel time between the upstream point and the downstream point, a reliability calculating means for calculating a reliability of the travel time, a vehicle detected at the upstream point Estimating means for estimating a time estimated to reach the downstream point based on the travel time and the reliability; and a vehicle detected at the upstream point based on the estimated time, the vehicle being detected at the downstream point. Traffic information management device, comprising: 13. A first collecting step of collecting a characteristic amount of a vehicle passing through at an upstream point of the road and a passing time of the vehicle, and a characteristic amount of a vehicle passing at a downstream point of the road and passing of the vehicle. A second collection step of collecting time; an acquisition step of acquiring a reference time estimated to be a travel time between the upstream point and the downstream point; the acquired reference time and the first A prediction step of predicting a time when a vehicle passing through the upstream point arrives at the downstream point based on the passing time of the vehicle collected in the collecting step; and the feature quantity collected in the second collecting step, Deviation from the feature amount collected in the first collection step,
The information of the group of vehicles collected in the first collecting step and the second collection based on the deviation of the passing time of the vehicle collected in the second collecting step from the predicted time. A matching step of matching with the information of the group of vehicles collected by the step, a calculation step of calculating a travel time between the upstream point and the downstream point based on the result of the matching, A reliability calculation step of calculating reliability, an estimation step of estimating a time at which the vehicle detected at the upstream point reaches the downstream point, based on the travel time and the reliability, based on the estimated time, the vehicle detected at said point upstream and a detecting step of detecting at said downstream point, said map In the ching step, the second collecting step
Said predicted time of passage of the vehicle collected by
A traffic information management method that calculates an evaluation value based on the amount of deviation from . 14. A first collecting step of collecting a characteristic amount of a vehicle passing through at an upstream point of a road and a passing time of the vehicle, and a characteristic amount of a vehicle passing through at a downstream point of the road and a passage of the vehicle. a second collecting step of collecting time, one collected by the feature and the second collection step of the first one vehicle collected by the collecting step of
The deviation of the feature amount of the vehicle, and the passage time of the first one vehicle collected by the collecting step, the passing time of the one vehicle collected by the second collecting step, these Calculate the evaluation value for each of the multiple vehicles of interest based on the reference time that is the estimated travel time between the two points
Then, the group of vehicle information collected in the first collecting step and the group of vehicle information collected in the second collecting step are matched based on the plurality of evaluation values obtained thereby. A matching step, a calculation step of calculating a travel time between the upstream point and the downstream point based on the result of the matching, a reliability calculation step of calculating a reliability of the travel time, and the upstream point An estimation step of estimating the time at which the detected vehicle arrives at the downstream point based on the travel time and the reliability, and a time detected at the upstream point based on the estimated time. And a detecting step of detecting a vehicle at the downstream point.