JP4320625B2 - Traffic signal control method and control apparatus for implementing the method - Google Patents

Description

The present invention relates to a traffic signal control method for optimally controlling the light color period of a traffic light installed at an intersection, and a control device for implementing the method .

  FIG. 10 is a graph depicting the cumulative number of vehicles passing through a specific inflow path at an intersection with respect to time. The cumulative number is drawn as a diagonal line. The slope q during the red light period represents the number of vehicles flowing into the intersection per unit time, that is, the inflow traffic. The slope r in the blue hour represents the number of vehicles flowing out from the intersection per unit time, that is, the outflow traffic. The difference q−r corresponds to the number of queues in the inflow path, and a value obtained by integrating the number of queues over time is called a total delay in the inflow path, and is used as an indicator of traffic congestion at the intersection.

FIG. 11 is a graph for explaining a dilemma sensitive control method. The vertical axis represents the vehicle travel speed (km / h), and the horizontal axis represents the distance (m) from the vehicle position to the stop line at the intersection.
In the figure, a curve L1 represents a relationship between a traveling speed and a distance that can be stopped if the vehicle decelerates, and a straight line L2 indicates that the vehicle is within the yellow signal time when a yellow signal is generated while the vehicle is traveling. It represents the relationship between travel speed and distance that can reach the stop line at the intersection. The vehicle in the region A on the left side of the straight line L2 determines that the vehicle will cross the intersection when it becomes a yellow signal during traveling. The vehicle in the region B on the right side of the curve L2 determines to decelerate and stop at the intersection when it becomes a yellow signal during traveling.

A region C between the straight line L2 and the curve L2 is called a dilemma zone, and vehicles in this region C are mixed with vehicles with different judgments for stopping or passing, so that there is a high possibility that a rear-end collision will occur. There is a high possibility of head-on accidents due to vehicles entering the intersection.
To extend the green light time when it detects the vehicle in such a dilemma zone, abort the green light if not detected, the control to prevent the accident that the "dilemma sensitive control".

The contents of the dilemma sensitive control will be specifically described.
Based on the measured speeds of all the vehicles sensed between time t-T (T is the monitoring target time) and t, and the current time in the sensitive range is t. Or B. When any of the above conditions is satisfied, the green light is cut off and switched to the yellow light.
A. Even if the vehicle travels at the lower speed limit, it passes through the downstream end of the dilemma zone.

B. Even if you drive at the maximum speed, you have not arrived at the upstream end of the dilemma zone.
JP 2001-134893 A JP-A-4-163700

Minimizing the total delay at the intersection described with reference to FIG. 10 is effective for eliminating traffic congestion. Therefore, a traffic signal control method that optimally controls the light color period of a traffic light is desired.
On the other hand, the dilemma sensitive control described with reference to FIG. 11 is a process for preventing the occurrence of an accident.

In the traffic signal control method for optimally controlling the light color period of the traffic light, there is a need for a traffic signal control method that simultaneously satisfies the two requirements of reducing traffic congestion and reducing traffic accidents.
Accordingly, the present invention simultaneously satisfies the two requirements of reducing traffic congestion and reducing traffic accidents by performing optimal dilemma sensitive control in a traffic signal control method for optimally controlling the light color period of a traffic light. It is an object of the present invention to provide a traffic signal control method capable of performing the above and a control device for implementing the method .

The traffic signal control method of the present invention obtains the traffic volume flowing in from each inflow path of the intersection, calculates the traffic volume predicted value of each inflow path, and based on the traffic volume and the traffic volume predicted value, over a predetermined period. The cost C serving as an indicator of traffic jam at the intersection is calculated, the optimal time Gopt for switching the color from blue to another color that minimizes the cost C is obtained, and the increase from the minimum value of the cost C is within the threshold ΔC. In this period, when it is determined that the vehicle traveling toward the controlled intersection will not be in a situation of traveling at the intersection at the end of the yellow light, the green light is discontinued, and it is determined that the situation will occur. A dilemma sensitive control process for extending the green light is performed.

With this method, a period in which the increase from the minimum value of the cost C is within the threshold ΔC is set, and by performing the dilemma sensitive control process within the set period, traffic congestion is reduced and traffic is reduced. Two requests to reduce accidents can be met simultaneously.
Further, by performing the dilemma sensitive control process in a period in which the increase from the minimum value of the cost C is within the threshold ΔC, the dilemma sensitive control process is effectively performed in the vicinity of the time when the signal is scheduled to switch from blue to another color. Can meet the demand to reduce traffic accidents. The smaller the threshold value ΔC, the more difficult the dilemma sensitive control is executed. If importance is placed on reducing traffic congestion, the threshold value ΔC may be decreased. Since the dilemma sensitive control becomes easier to execute as the threshold value ΔC is larger, the threshold value ΔC may be increased if importance is placed on accident prevention.
The lamp color switching optimum timing Gopt is obtained, and a period in which the increase from the minimum value of the cost C is within the threshold ΔC is set. In this period, a series of processes for performing the dilemma sensitive control process is performed within the green signal period. It is desirable to carry out it a plurality of times for each fixed period n. The fixed period n is, of course, a period shorter than the green light period. By performing multiple times in this way, the traffic volume measurement value of each inflow route changes from moment to moment, and the traffic color prediction value that flows in from each inflow route at the intersection also changes. The time Gopt can be determined more accurately.

If the time sopt until the cost C reaches the minimum value is included in the “predetermined period n during the green light period”, a dilemma sensitive control process is performed , and the “predetermined period n during the green light period” is set. If it exceeds, a process of not performing the dilemma sensitive control process is also useful. This is because (1) more than once before the time when the signal is scheduled to switch from blue to another color, “require the optimum time Gopt for light color switching and set the period related to the optimum time Gopt for light color switching. In this period, if there is an opportunity to perform a series of processes of performing dilemma sensitive control processing, (2) the time when the signal is scheduled to switch from blue to another color without performing the dilemma sensitive control processing comes. Until now, there will no longer be an opportunity to perform “a series of processes for obtaining a lamp color switching optimal timing Gopt, setting a period related to the lamp color switching optimal timing Gopt, and performing a dilemma sensitive control process in this period”. In this case, the dilemma sensitive control process is performed .

By this process, the opportunity to perform the dilemma sensitive control process can be given only once before the optimal color Gopt of the lamp color switching. This is effective when the dilemma sensitive control process is to be performed only once during the green light period.
Since the minimum value Gmin and the maximum value Gmax of the green light period are usually set regardless of whether the traffic volume is large or small, the period for calculating the minimum value of the cost C is from the minimum value Gmin to the maximum value Gmax of the green light period. It is desirable to be within range.
Moreover, the control apparatus which implements the traffic signal control method of this invention is a control apparatus concerning the invention of the traffic signal control method mentioned above substantially the same invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
In the present invention, the traffic volume is measured at the intersection. Here, the “traffic volume” means the number of vehicles passing through the measurement point per unit time. For traffic measurement, an ultrasonic vehicle detector that detects vehicles on the road is installed on each inflow path at the intersection, or image processing that detects vehicles by shooting the road with a camera and processing the images. A vehicle detector is installed.

FIG. 1 is a diagram showing controlled object intersections and signal cycles. The inflow path to the intersection is indicated by 1 to 4.
One cycle of the signal is composed of a series of lamp color periods, and the lamp color of each inflow path is determined for each lamp color period. When the lamp color period for one cycle is completed, the next cycle starts.
In the example of FIG. 1, one cycle is composed of six lamp color periods, and the lamp color periods are expressed as 1 to 6, respectively. If attention is paid to the inflow channels 1 and 3, the lamp color period 1 is the blue period. The light color period 2 is a yellow period, and the light color periods 3 to 6 are red periods. If attention is paid to the inflow channels 2 and 4, the lamp color periods 1 to 3 are red periods. The lamp color period 4 is a blue period, the lamp color period 5 is a yellow period, and the lamp color period 6 is a red period.

The present invention optimizes the length of the light-colored period 1 according to the traffic volume of the inflow paths 1 and 3, that is, the blue time, and optimizes the length of the light-colored period 4 according to the traffic volume of the inflow paths 2 and 4 Turn into. The lengths of the lamp periods 2, 3, 5, and 6 are not changed from the default values.
Hereinafter, an embodiment in which the length of the lamp period 1 is changed according to the traffic volume of the inflow channels 1 and 3 will be described. In the embodiment in which the length of the lighting period 4 is changed according to the traffic volume of the inflow paths 2 and 4, the control contents are the same except that the subscript indicating the inflow path is different.

Hereinafter, the definition and calculation formula of each variable will be described.
Current time t: Current time (t> 0) counted from the start of the green signal. Let t = 0 when the green signal starts. In particular, t during the blue period is sometimes referred to as Gnow.
T: Cycle length (sum of default lengths for all lighting periods).
R: Total time other than green light in one cycle.

qi (t): Number of inflowing vehicles per unit time of inflow path i at time t.
Qi: Number of outflows per unit time from the queue of inflow path i. A fixed value.
Mi: Number of queues in the inflow path i at the start of the green light (t = 0).
mi (t, s): The number of queues for the inflow path i at time t. mi (t, s) is expressed as the following equation.

mi (t, s) = Mi (t = 0)
mi (t, s) = max {mi (t-1) -Qi + qi (t), 0}
(0 <t ≦ s, s + R <t ≦ T)
mi (t, s) = mi (t-1) + qi (t) (s <t≤s + R)
C (s): Cost when the lamp color is switched from blue to yellow after elapse of time s from the current time t (unit: number × time). C (s) is expressed by the following equation.

C (s) = Σ Σmi (t, s)
(I is from 1 to 4, t is from 1 to T)
Gmin: Minimum value of the blue light color period. No matter how little the traffic is, it cannot be shortened any further.
Gmax: Maximum value of the blue light color period. No matter how much traffic you have, you cannot make it longer.

Gnow: Elapsed time from the start of the blue light period to the current time.
Gopt: Time to switch the blue light to the next color based on the calculated optimal blue light color period.
F (G): A flag indicating whether or not the dilemma sensitive control is permitted with the elapsed time G from the start of the light color period for the current light color period.
By the way, the matrix number mi (t, s) and the number of inflowing vehicles qi (t) are not obtained only by past values, and the predicted values of one cycle ahead must be used. There are various methods for obtaining the predicted value. For example, the traffic volume measured by a sensor installed upstream of the intersection is calculated assuming that the traffic flow into the intersection is delayed by the required time from the sensor to the intersection, or a moving average or index based on the measured traffic volume. Using a smoothing method, a change in the number of inflowing vehicles qi (t) from the past to the present is obtained, and it can be obtained on the assumption that the tendency of the change will continue in the future.

  FIG. 2 is a graph plotting the value of the number of inflowing vehicles qi (t) measured at the past t <0 and the predicted value of the number of inflowing vehicles qi (t) toward the future one cycle ahead. Since one cycle is about several tens of seconds at the longest, the accuracy of the predicted value obtained in this way is sufficiently good from experience. In addition, when determining the predicted value of the inflowing vehicle number qi (t), not only the value obtained by extending the inflowing vehicle number qi (t), but also the inflowing vehicle number qi (t It is also effective to collect the data of) and determine the predicted value of the inflowing vehicle number qi (t) in consideration of the accumulated value.

FIG. 3 and FIG. 4 are graphs showing the time change of the number of queues mi (t, s) in the inflow path i. The time t at the start of the green signal is set to zero. The number of queues in the inflow path i at that time is Mi. Since usually Qi> qi (t), the number of matrixes mi (t, s) decreases from the start of the green light. However, it does not become 0 or less.
When the time s elapses from the current time Gnow, the lamp color changes from yellow to red, and the number of queues mi (t, s) increases. The sum of the number of queues mi (t, s) over one cycle T is the area of the portion indicated by hatching. If this area is summed over all the inflow channels, the cost C (s) can be obtained.

The difference between FIG. 3 and FIG. 4 is that in FIG. 3, the number of queues mi (t, s) becomes 0 during the period of the green light, whereas in FIG. It is.
FIG. 5 is a flowchart showing the overall processing of the traffic signal control method of the present invention.
This processing is performed in real time every 0.1 seconds.
First, it is determined whether or not the current lamp color period is a blue period (step S1).

If it is the blue period, the counter U determines whether or not the period of “blue time optimization process” n seconds (n is 6 seconds, for example) has passed (step S2).
If n seconds have elapsed, the blue time optimization process (step S3) is performed. This blue time optimization process will be described later.
When the blue time optimization process is completed, U = 0 is set (step S4).

Next, it is confirmed whether or not the dilemma sensitive control is permitted at the elapsed time Gnow (that is, the current time) from the start of the lamp color period. This confirmation is confirmed based on whether or not the flag F (Gnow) set in the procedure of FIG. 7 is 1 (step S5).
If the dilemma sensitive control is permitted, the dilemma sensitive control is performed (step S6). Since the content of the dilemma sensitive control is well known, it will be briefly described. This is a process of extending or shortening the timing Gopt for ending the green signal set in step T8 of FIG. 6 when a vehicle in a so-called dilemma zone is detected.

Thereafter, the counter U is advanced by 0.1 in accordance with the actual passage of time (step S7).
If the dilemma sensitive control is not permitted in step S5, the counter U is advanced by 0.1 without performing the dilemma sensitive control (step S7).
Next, it is determined whether or not the elapsed time Gnow from the start of the lamp color period is a timing Gopt for shifting to the next lamp color period (step S8). If Gnow = Gopt, the lamp color is the next color. Switch to yellow (step S9).

Next, the blue time optimization process (step S3) will be described with reference to FIG. This process is performed every n seconds.
First, the lamp color switching time s is set to the time (Gmin-Gnow) from the elapsed time Gnow from the start of the lamp color period to the present to the minimum value Gmin of the lamp color period (step T1). Here, the minimum value Gmin of the lamp color period is a minimum value (set value) that the lamp color period 1 should not be shorter than this. If the elapsed time Gnow exceeds Gmin, the lamp color switching time s is set to 1. This means that the lamp color is switched after one second. In this sense, the lamp color switching time s is the time that must be waited for the minimum even if the signal is switched earliest from now.

Next, the minimum cost Copt is set to a number that can be regarded as infinite (step T2).
Next, the cost C (s) for switching the lamp color at time s is calculated. As described above, the cost C (s) is the sum of the number of queues mi (t, s) of the inflow path i at time t over one cycle for all inflow paths. The number of queues mi (t, s) of the inflow path i can be obtained based on the number of inflowing vehicles qi (t) per unit time of the inflow path i at time t.

It is determined whether or not C (s) <Copt is satisfied (step T4). If it is satisfied, Copt is replaced with the value of C (s), and sopt is set to the value of s (step T5). sopt is a lamp color switching time when the cost C (s) becomes the minimum cost Copt.
Next, it is determined whether the lamp color switching time s has reached the time (Gmax−Gnow) from the elapsed time Gnow from the start of the lamp color period to the maximum value Gmax of the lamp color period. Here, the maximum value Gmax of the lamp color period is the maximum value (set value) of the lamp color period 1 that the lamp color period 1 should not be longer than this.

If (Gmax−Gnow) has not been reached, 1 is added to s (step T7), and the next cost C (s) is calculated (step T3). Here, “1” added to s does not match the actual passage of time, but is a shorter time that can be processed by the computer.
It is determined whether or not C (s) <Copt is satisfied (step T4). If it is satisfied, Copt is replaced with the value of C (s), and sopt is set to the value of s (step T5). If C (s) ≧ Copt, step T5 is skipped. In this way, the minimum value of C (s) is obtained.

If the lamp color switching time s reaches the maximum lamp color period value Gmax (YES in step T6), the Copt obtained so far is stored, and the elapsed time Gnow from the start of the lamp color period to the present is stopped. Is added to obtain the optimal color change timing Gopt (step T8). Then, permission of dilemma sensitive control is determined (step T9), and the process returns to step S4 in FIG.
In this way, it is possible to obtain the optimal color change timing Gopt that minimizes the cost C (s).

Next, the procedure for determining permission of the dilemma sensitive control in step T9 will be described with an example.
FIG. 7 is a flowchart for explaining an exemplary procedure for determining permission of dilemma sensitive control. FIG. 8 is a graph in which the obtained cost C (s) is plotted with s as the horizontal axis.

First, the lamp color switching time s is set to the time (Gmin−Gnow) from the elapsed time Gnow from the start of the lamp color period to the present to the minimum value Gmin of the lamp color period (step U1). This is the same procedure as step T1 in FIG.
It is determined whether or not the difference between the cost C (s) and the minimum cost Copt at the time s is within the threshold ΔC (step U2). If the difference is within the threshold ΔC, the variable A is set to 1 (step U3), If larger than ΔC, the variable A is set to 0 (step U4). A = 1 means that dilemma sensitive control is permitted, and A = 0 means that dilemma sensitive control is not permitted.

And times s, s + 0.1, s + 0.2,. . . , S + 0.9, the value of the flag F permitting dilemma sensitive control is set to A. That is,
F (Gnow + s + i) = A (i is from 0 to 0.9)
(Steps U5 to U8). This indicates that the value A of the flag F is fixed for 1 second and the determination of whether or not to permit dilemma sensitive control is not changed.

The threshold ΔC is the value of the smaller flag F period is shortened as the 1, Kukunaru a dilemma sensitive control is executed. If it is important to reduce traffic congestion, the threshold value ΔC may be reduced. The larger the threshold ΔC, the longer the period during which the value of the flag F is 1 and the dilemma sensitive control becomes easier to execute. If importance is placed on accident prevention, the threshold value ΔC may be increased.
Next, 1 is added to s (step U10), and the processing after step U2 is repeated. Thereby, the value of the flag F for the next one second is determined.

If the lamp color switching time s reaches the time (Gmax−Gnow) from the elapsed time Gnow from the start of the lamp color period to the current value Gmax (Gmax−Gnow) (step U9), the process is terminated. Return to immediately before the return in FIG.
As described above, the value of the flag F indicating whether or not the dilemma sensitive control is permitted can be set over the time from the minimum value Gmin to the maximum value Gmax.

  In the dilemma permission determination control of FIG. 7 described so far, the flag F indicating whether or not to permit the dilemma sensitive control based on whether or not the difference between the cost C (s) and the minimum cost Copt is within the threshold ΔC. The value was determined. However, if the time sop that minimizes C (s) arrives by the next blue time optimization (after n seconds), the dilemma sensitive control is performed during the next blue time optimization. There is also a method of permitting.

FIG. 9 is a flowchart for explaining another procedure for determining permission of dilemma sensitive control.
First, the lamp color switching time s is set to the time (Gmin−Gnow) from the elapsed time Gnow from the start of the lamp color period to the present to the minimum value Gmin of the lamp color period (step V1). This is the same procedure as step U1 in FIG. It is determined whether the time sopt is within n seconds (step V2). If it is within n seconds, it is determined whether the time s is within n seconds (step V3). The variable A is set to 1 (step V4). If any exceeds n seconds, variable A is set to 0 (step V5). And times s, s + 0.1, s + 0.2,. . . , S + 0.9, the value of the flag F permitting dilemma sensitive control is set to A. Ie
F (Gnow + s + i) = A (i is from 0 to 0.9)
(Steps V5 to V9). This means that the decision to permit or disallow dilemma sensitive control is not changed for 1 second.

Next, 1 is added to s (step V11), and the processes after step V2 are repeated. Thereby, the value of the flag F for the next one second is determined.
If the lamp color switching time s reaches the time (Gmax−Gnow) from the elapsed time Gnow from the start of the lamp color period to the current value Gmax (Gmax−Gnow) (step V10), the process is terminated. Return to immediately before the return in FIG.

As described above, if the time until the cost C (s) reaches the minimum value is within n seconds, which is the period of the blue time optimization processing, the permission determination of the dilemma sensitive control is made and the time is less than n seconds. If it is earlier, it is possible to refer to the next blue time optimization process without deciding whether to permit dilemma sensitive control.
Although the embodiments of the present invention have been described above, the embodiments of the present invention are not limited to the above-described embodiments, and various modifications can be made within the scope of the present invention.

It is a figure which shows the example of distribution of the lamp color for every inflow path in an intersection, and the inflow path in 1 cycle lamp color period. It is the graph which plotted the predicted value of inflow vehicle number qi (t) toward the future of 1 cycle ahead while plotting the value of inflow vehicle number qi (t) measured in the past. It is a graph which shows the time change of the queue number mi (t, s) in the inflow path i. It is a graph which shows the time change of the queue number mi (t, s) in the inflow path i. It is a flowchart showing the whole process of the traffic signal control method of this invention. It is a flowchart showing a blue time optimization process. It is a flowchart for demonstrating the procedure which determines permission of dilemma sensitive control. It is the graph which plotted cost C (s) by making s into a horizontal axis. It is a flowchart for demonstrating the other procedure which determines permission of dilemma sensitive control. It is the graph which drawn the cumulative number with respect to time which accumulated the passing number of vehicles which flow in from the specific inflow way of an intersection in time. It is a graph for demonstrating the general dilemma sensitive control method.

Explanation of symbols

1,2,3,4 Inflow channel

Claims (5)

(A) Obtain the traffic volume flowing in from each inflow path at the intersection,
(B) Calculate the traffic forecast value for each inflow channel,
(C) Based on the traffic volume and the traffic volume prediction value, calculate a cost C that serves as an indicator of traffic congestion at an intersection over a predetermined period;
(D) Finding the optimal time Gopt for switching the color from blue to another color that minimizes the cost C,
(E) setting a period during which the increase from the minimum value of the cost C is within the threshold ΔC;
In (f) this period, it aborts the green light when the vehicle traveling towards the intersection of the controlled object is determined not to become a situation in which the traveling intersection at yellow light ends, extend the green light when it is determined that becomes the situation A traffic signal control method comprising performing a dilemma sensitive control process.   The traffic signal control method according to claim 1, wherein the steps (c) to (e) are performed a plurality of times every predetermined period n within one green light period. In the procedure (e) , if the time sopt until the cost C reaches the minimum value is within the “predetermined period n during the green signal period”, a dilemma sensitive control process is performed, and the “during the green signal period” 3. The traffic signal control method according to claim 2, wherein the dilemma sensitive control process is not performed if the predetermined period n is exceeded. In the procedure of the (c), the period for calculating the minimum value of the cost C is the traffic signal according the minimum value Gmin of the green light period claims 1 over the range of the maximum value Gmax to any one of claims 3 Control method. A control device for implementing a traffic signal control method, comprising a computer,
Means for obtaining the traffic volume measured by the vehicle detector installed in each inflow path of the intersection;
Means for calculating a predicted traffic volume for each inflow channel;
Means for calculating a cost C that serves as an indicator of traffic congestion at an intersection over a predetermined period based on the traffic volume and the traffic volume prediction value;
Means for obtaining an optimal timing Gopt for switching the color of the lamp from blue to another color that minimizes the cost C;
Means for setting a period during which an increase from the minimum value of the cost C is within a threshold ΔC;
In this period, abort the green light when the vehicle traveling towards the intersection of the controlled object is determined not to become a situation in which the traveling intersection at yellow light ends, dilemma sensitive to extend the green light when it is determined that becomes the situation And a control device for performing control processing.

Images