JPH0714093A - Traffic signal controlling method - Google Patents

Abstract

PURPOSE:To smooth stream of pedestrians and vehicles extending over plural crossings. CONSTITUTION:The deviation of speed between actual speed and target traveling speed is obtained on the basis of the detection output of a speed sensor P1. Then, in order to calculate signal time for making the speed deviation zero, the sensitivity to show influence to the average traffic amount to blue time of each signal S0, S1 is obtained in respect of a road P1. Further, in order to minimize the deviation between the target speed set for the road P1 and the actual speed, the inclination to the blur time of a set evaluation function FV and each signal S0, S1 is obtained. This value is obtained by considering the sensitivity, and when this value exceeds a set value, the blue time of the signal S1 is increased, and the blue time of the signal S0 is decreased. Namely, the traffic state of the road P1 is monitored, and the blue time of each signal S0, S1 is set while considering mutually the traffic state related to the signals S0, S1. Thus, the effect of traffic congestion can be suppressed to the minimum.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a traffic signal control method, and more particularly to a traffic signal control method suitable for traffic control by adjusting signal times of traffic signals at intersections of road networks.

[0002]

2. Description of the Related Art Conventionally, when controlling traffic on a road network, for example, JP-A-3-46100 and JP-A-2-46100 are available.
As described in JP-A-287700, a method of optimizing the flow of traffic at an intersection alone based on road information is adopted. When controlling the signal time of the traffic signal at each intersection based on the road information, a method called SCOOT is adopted which predicts the arrival pattern of the vehicle group on the downstream intersection side between the adjacent intersections and determines the signal parameter. There is.

In consideration of the length of traffic congestion, "Simulation on adaptive control of traffic congestion" Shimizu
Hikari: A traffic engineering research presentation (November 1993), pages 117 to 119, a control method has been proposed in which a traffic flow is linearly modeled to minimize the congestion length.

[0004]

Since the above-mentioned prior art is not a method of comprehensively evaluating a traffic network, even if optimal control can be performed at each intersection, it is not always optimal from the viewpoint of the road network. I can't. That is, even if it is possible to set the optimum signal time at each intersection alone,
If the signal times of traffic signals at multiple intersections are not set and the signal times are not set, the road network may be divided into a congested area and a quiet area. There is a problem that it does not become.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a traffic signal control method capable of adjusting the signal times of a plurality of traffic lights in association with each other to smooth the flow of traffic across a plurality of intersections.

[0006]

In order to achieve the above-mentioned object, the present invention controls traffic signals based on vehicle speed information by controlling one road between adjacent intersections of a road network. Speed detection processing for detecting the speed of a vehicle traveling on each of a plurality of control target roads as a target road, each detection speed detected by the speed detection processing, and an average target travel speed set for each control target road The speed deviation calculation process for calculating each speed deviation and the signal adjustment time for suppressing each speed deviation to zero based on each speed deviation calculated in the speed deviation calculation process are calculated for the traffic signals at each intersection. Traffic including signal adjustment time calculation processing and signal time adjustment processing for adjusting the signal time of each traffic signal on each control target road according to the signal adjustment time calculated by the signal adjustment time calculation processing It is obtained by employing the signal control method.

In the travel time measuring process for measuring the travel time of a vehicle traveling on each of a plurality of control target roads with a road between adjacent intersections of the road network as one control target road, and the travel time measuring process. Speed calculation processing for calculating the actual traveling speed of the vehicle on each control target road from the measured values, each calculated speed calculated by the speed calculation processing, and the average target set for each control target road A speed deviation calculation process for calculating a speed deviation from the traveling speed,
Based on each speed deviation calculated by the speed deviation calculation processing, the signal adjustment time calculation processing for calculating the signal adjustment time for suppressing each speed deviation to zero for each traffic signal at each intersection, and the signal adjustment time calculation processing According to the calculated signal adjustment time, a traffic signal control method including a signal time adjustment process of adjusting the signal time of each traffic signal on each control target road is adopted.

Furthermore, a speed detection process for detecting the speed of a vehicle traveling on a plurality of control target roads with a road between adjacent intersections of the road network as one control target road, and a vehicle on each control target road. Congestion length detection processing for detecting the congestion length, traffic volume detection processing for detecting the traffic volume indicating the number of vehicles passing through each control target road within a fixed time, the speed detection processing and the congestion length detection processing, Effect of change in signal time of each traffic signal on average speed of vehicles on each control target road based on each detection value detected by the traffic volume detection processing and current signal time of each traffic signal of each control target road The speed sensitivity calculation processing for calculating the speed sensitivity indicating the degree of the speed deviation, and the speed deviation calculation for calculating the speed deviation between the detected speed detected by the speed detection processing and the average target traveling speed set for each control target road Processing and The speed deviation calculated by the speed deviation calculation processing is evaluated by the speed evaluation processing in relation to the change in the signal time of each traffic signal, and the evaluation value obtained by the speed evaluation processing and the speed sensitivity calculation processing are calculated. Signal adjustment time calculation processing for calculating the signal adjustment time for suppressing the speed deviation to zero based on the speed sensitivity for the traffic signal at each intersection, and each control according to the signal adjustment time calculated by the signal adjustment time calculation processing The traffic signal control method including a signal time adjustment process for adjusting the signal time of each traffic signal on the target road is adopted.

Each of the above control methods can be applied to each traffic light on a single controlled road.

Further, in applying the above control methods,
The speed evaluation processing, when evaluating the relationship between each speed deviation and all other speed deviations, weights each speed deviation determined from the positioning of each controlled road on the traffic network,
It is desirable to include a process of calculating an evaluation value for minimizing the sum of the weighted speed deviations.

Further, when the above control methods are applied, each process is executed periodically, and the signal adjustment time calculation process is a process for calculating a signal adjustment time for stepwise suppressing the speed deviation to zero. It is desirable to include.

In order to control the traffic signal based on the information about the traffic congestion length of the vehicle, which is the road information, the road between adjacent intersections in the road network is defined as one control target road, and Congestion length detection processing for detecting the congestion length, and congestion length deviation for calculating the congestion length deviation between each congestion length detected by the congestion length detection processing and the average target congestion length set for each controlled road Calculation processing, signal adjustment time calculation processing for calculating the signal adjustment time for suppressing each congestion length deviation to zero on the basis of the congestion length deviation calculated in the congestion length deviation calculation processing, for the traffic signal at each intersection, and A traffic signal control method including a signal time adjustment process for adjusting the signal time of each traffic signal on each controlled road according to the signal adjustment time calculated by the signal adjustment time calculation process is adopted. Than it is.

The above control method can be applied to each traffic light on a single controlled road. Further, when the control method is applied, each process is executed cyclically, and the signal adjustment time calculation process may include a process for calculating a signal adjustment time for gradually reducing the congestion length deviation to zero. desirable.

Next, the traffic signal is controlled on the basis of the traffic volume information of the road information. The road between adjacent intersections of the road network is set as one control target road, and each of the plurality of control target roads is traveled. Traffic volume detection processing for detecting the traffic volume indicating the number of vehicles passing through within a certain time, and the average target traffic volume set for each traffic volume and each control target road detected by the traffic volume detection processing Traffic volume deviation calculation processing for calculating a traffic volume deviation with the traffic volume deviation, and a signal adjustment time for suppressing each traffic volume deviation to zero based on the traffic volume deviation calculated in the traffic volume deviation calculation processing. Traffic including a signal adjustment time calculation process for calculating the signal adjustment time and a signal time adjustment process for adjusting the signal time of each traffic signal on the controlled road according to the signal adjustment time calculated by the signal adjustment time calculation process. It is obtained by employing the signal control method.

Further, a traffic volume detection for detecting a traffic volume indicating the number of vehicles passing through each of a plurality of control target roads within a certain time period with a road between adjacent intersections of the road network as one control target road. Sensitivity indicating the degree of influence of changes in the signal time of each traffic signal on the average traffic volume of each controlled road based on each traffic volume detected by the traffic volume detection processing and the current signal time of each traffic signal And a traffic volume deviation calculation process for calculating a traffic volume deviation between each traffic volume detected by the traffic volume detection processing and the average target traffic volume set for each control target road. And a traffic volume evaluation process that evaluates the traffic volume deviation calculated by the traffic volume deviation calculation process in relation to the change in the signal time of each traffic light, and the evaluation value and the traffic volume obtained by the traffic volume evaluation process. Calculated by the sensitivity calculation process Signal adjustment time calculation processing for calculating the signal adjustment time for suppressing each traffic deviation to zero on the basis of the obtained traffic sensitivity, and the signal adjustment calculated by the signal adjustment time calculation processing The traffic signal control method includes a signal time adjustment process for adjusting the signal time of each traffic signal on the controlled road according to time.

In applying the above control method, the traffic volume evaluation process uses the traffic of each control target road for each traffic volume deviation when evaluating the relationship between each traffic volume deviation and all other traffic volume deviations. It is desirable to include a process of assigning a weight determined from the position on the network and calculating an evaluation value for maximizing the sum of the weighted traffic volume deviations. Furthermore, when the control method is applied, each process is executed periodically, and the signal adjustment time calculation process includes a process for calculating a signal adjustment time for stepwise suppressing the traffic volume deviation to zero. desirable.

In order to control the traffic signal based on the information related to the direction guidance among the traffic information, a group of vehicles traveling on a plurality of control target roads with a road between adjacent intersections of the road network as one control target road. Traffic volume detection processing for detecting the traffic volume indicating the number of vehicles passing within a fixed time, and a saturation degree indicating the ratio of each traffic volume detected by the traffic volume detection processing to the maximum traffic capacity of each controlled road. Saturation degree calculation processing to calculate, saturation degree deviation calculation processing to calculate the saturation degree deviation between each saturation degree calculated in the saturation degree calculation processing and the average target saturation degree set for each control target road, A signal adjustment time calculation process for calculating a signal adjustment time for suppressing each saturation deviation to zero based on the saturation deviation calculated in the saturation deviation calculation process, and the signal adjustment time calculation process place In which calculated according to the signal adjustment time signal time of each traffic signal in each control target road intended to constitute a traffic signal control method comprising a signal time adjustment process for adjusting, respectively.

Further, a traffic volume detection for detecting a traffic volume indicating the number of vehicles passing through a plurality of control target roads within a predetermined time, with a road between adjacent intersections in the road network as one control target road. Processing, a saturation degree calculation processing for calculating a saturation degree indicating a ratio of each traffic volume detected by the traffic volume detection processing and the maximum traffic capacity of each control target road, and a saturation calculated by the saturation degree calculation processing Saturation degree sensitivity calculation processing to calculate the saturation degree sensitivity indicating the degree of the influence of the change in the signal time of each traffic signal on the average saturation degree of each control target road based on the degree and the current signal time of each traffic signal, and Saturation degree deviation calculation processing for calculating the saturation degree deviation between each saturation degree calculated in the saturation degree calculation processing and the average target saturation degree set for each control target road, and the saturation degree deviation calculation processing Each saturation deviation Saturation evaluation processing to evaluate in relation to the change in the signal time of the unit, each saturation degree based on the evaluation value obtained in the saturation evaluation processing and the saturation sensitivity calculated in the saturation deviation calculation processing Signal adjustment time calculation processing for calculating the signal adjustment time for suppressing the deviation to zero for the traffic light at each intersection, and the signal time of each traffic light on each controlled road according to the signal adjustment time calculated by the signal adjustment time calculation processing A traffic signal control method including a signal time adjustment process for adjusting each of the above is adopted.

In applying the above control method, the saturation evaluation processing is carried out by evaluating the relationship between each saturation deviation and all the other saturation deviations when the saturation deviation is evaluated by the traffic network of each control target road. It is preferable to include a process of assigning a weight determined from the above-mentioned positioning and calculating an evaluation value for maximizing the sum of the weighted saturation deviations.

Furthermore, when applying each of the control methods, each process is executed periodically, and the signal adjustment time calculation process calculates a signal adjustment time for stepwise suppressing the saturation deviation to zero. It is desirable to include processing.

When adapting a traffic signal control based on speed, congestion length, traffic volume and saturation, the signal time adjustment process changes the signal time with the green time, the red time and the yellow time as one signal cycle. When doing so, it is desirable to include the process of adjusting the ratio of green time to other signal times in each signal cycle.

Similarly, as the information that affects the signal time of each traffic light on the controlled road, the timings of the red time and the green time that are paired with each traffic light and that are included in the signal time are in an inverse relationship. It is preferable that the method includes a road information detection process of detecting road information related to a traffic signal, and the signal adjustment time calculation process includes a process of calculating a signal adjustment time in consideration of the road information detected by the road information detection process.

Further, when each control method is applied, the traffic jam time detection processing for detecting the traffic jam time information of each control target road and the traffic jam information for sequentially storing the traffic jam time information detected by the traffic jam time detection processing. A signal adjustment time calculation process including a storage process and a priority order setting process for setting a priority order for each traffic signal related to a signal time for suppressing a traffic jam based on the information about the traffic jam time stored in the traffic jam information storage process. Preferably includes a process of calculating the signal adjustment time in consideration of the priority set in the priority setting process.

[0024]

According to the above-mentioned means, when the speed of the vehicle is detected as the road information related to the road to be controlled, the speed deviation between the detected speed and the average target traveling speed is calculated. Based on the calculated speed deviation, the signal adjustment time for suppressing the speed deviation to zero is calculated in association with the signal time of each traffic signal, and the signal time of each traffic signal is adjusted according to the calculated signal adjustment time. . That is, the signal times of the traffic signals are set in association with each other in order to minimize the influence of traffic congestion. Therefore, the flow of traffic can be smoothed.

When the congestion length and the traffic volume are considered in addition to the speed in the road information, the vehicle speed, the congestion length and the traffic volume of the vehicle are detected. Then, based on each detection information and the current signal time of each traffic light, the speed sensitivity indicating the degree of influence of the change in the signal time of each traffic light on the average speed of the vehicle on the controlled road is calculated, and the detected speed And the average deviation of the average traveling speed is calculated. The relationship between the speed deviation and the change in the signal time of each traffic signal is evaluated, and the signal time for suppressing the speed deviation to zero is calculated for each traffic signal based on this evaluation value and the speed sensitivity. That is, the signal time of each traffic signal is set in consideration of the speed of the vehicle, the congestion length, and the traffic volume, and the signal time of each traffic signal is adjusted according to the set signal adjustment time. In this case, the signal time of each traffic signal is adjusted in consideration of the speed of the vehicle, the length of traffic jam, and the traffic volume, so that the flow of traffic can be smoothed according to the traffic situation of the road network and the traffic can be decentralized. And traffic congestion can be reduced.

Further, in the case of traffic signal control mainly on information about traffic congestion length of road information, traffic signal control mainly on information on traffic volume, and further traffic information control mainly on information on saturation. In the same way as when controlling traffic signals based on speed information, the signal time of each traffic signal on the controlled road is set in consideration of the signal time of each other, so the flow of traffic across multiple intersections is smooth. Can be
That is, when the traffic congestion length (traffic volume or saturation) of the vehicle is detected as the road information related to the controlled road, the detected congestion length (traffic volume or saturation) and the average target congestion length (average traffic volume or average saturation). Traffic congestion length deviation (traffic volume or saturation deviation). Then, based on the calculated speed deviation (traffic volume or saturation deviation), the signal adjustment time for suppressing this speed deviation (traffic quantity or saturation deviation) to zero is calculated in association with the signal time of each traffic signal, The signal time of each traffic light is adjusted according to the calculated signal adjustment time. At this time, the signal time of each traffic light is set in association with each other in consideration of traffic conditions in order to minimize the influence of traffic congestion.

[0027]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is an overall configuration diagram showing an embodiment when the present invention is applied to a road network. In FIG. 1, the roads between adjacent intersections in the road network are the control target roads P1 and P.
2, ... Pn is set, and each control target road P
1 to Pn are sensor groups for detecting road information, which are speed sensors V1, V2, ... Vn and congestion length sensors L1 and L.
2, ... Ln, traffic volume sensors K1, K2, ... Kn are installed.

The speed sensor Vi is composed of, for example, an ITV camera, an ultrasonic sensor, an optical sensor, or an image sensor, and the vehicle 1 traveling on each controlled road Pi.
The speed of 0 is detected and the detected speed is output to the controller 101. Each congestion length sensor Li detects the congestion length of the vehicle 10 traveling on each controlled road Pi,
The detection output is output to the controller 101. The traffic volume sensor Ki detects the traffic volume (the number of vehicles passing per unit time) on each control target road Pi, and outputs a detection output to the controller 101.

The controller 101 includes a speed deviation minimization unit 102, a traffic flow model 103, and a speed deviation calculation unit 104, and each unit is composed of, for example, a microcomputer. The speed deviation calculation unit 104 includes a calculator Vn1, and each speed sensor V1, V2, ... Vn and a target speed (average target traveling speed set for each control target road Pi) V10, V20 ,. The speed deviation from Vn0 is calculated, and the calculated speed deviation signal is used as the speed deviation minimum lower part 1
It is designed to output to 02. Each target speed V10,
.. Vn0 is set corresponding to each control target road Pi. For example, when giving priority to an emergency vehicle, this target speed is set to 50 km / h on only one road, and the other speeds are set on other roads. Is set to 15 km / h.

The lower portion of the speed deviation minimum 102 takes in the signal from the speed deviation calculation unit 104 and the signals from the respective congestion length sensors Li and the traffic volume sensor Ki, and suppresses the speed deviation according to the input signal and the traffic flow model 103. Is calculated, and the signal times of the traffic signals S0, S1, ... Sn at each intersection are adjusted according to the calculated signal parameters. When adjusting this signal time,
When the signal cycle of each traffic light S0 to Sn is set to 1 cycle and 2 minutes, the ratio (split) between the blue time and the other time of the red time, the yellow time and the blue time included in the signal time is adjusted. It is supposed to be done.

Here, when the controller 101 adjusts the signal time of each of the traffic signals S0 to Sn based on the road information, in order to smooth the flow of traffic over a plurality of intersections, a control index as shown below is used. It is supposed to control the signal time.

First, it is possible to level the traffic by reducing the deviation between the target speed and the actual travel speed of the vehicle 10 traveling on each road Pi.

Therefore, in order to evaluate the deviation between the target speed and the actual speed in association with each road Pi, the following evaluation function FV is defined, and the value of this evaluation function FV is minimized. It is supposed to control.

[0034]

[Equation 1]

Here, vsi, vri, and vli are the actual speeds within one cycle of the straight lane, the right turn lane, and the left turn lane, respectively, and vsoi, vroi, and vloi are the targets of the straight lane, right turn lane, and left turn lane, respectively. It is the traveling speed. Further, wsi, wry, and wli are weights of the respective lanes (weights considering whether or not the lane is a main line).
In the case of using the following equation (1), only straight lanes will be considered to simplify the description.

When the deviation between the target speed and the actual speed is evaluated using the equation (1), the traveling speed is affected by the congestion length and traffic volume of each road. As shown in FIG. 2, the road traffic information about the two intersections q and q ′ is used as a traffic flow model, and the information about the traffic flow model 103 is stored in the memory.

When the intersections q and q'are used in the model, the average travel time t within one cycle between two adjacent intersections q and q'is expressed by the following equation (2).

[0038]

[Equation 2]

Here, lsp: traffic congestion length of straight lane p,
d: headway distance kp ', kp: road p', traffic volume (vehicles / second) on p, grq, grq ': green time VC of intersection q, q', VF: moving speed in non-congestion area, non-congestion This is the speed of movement within the area.

Here, VF is the free traveling speed of the vehicle traveling in the non-congested section, and VC is the traveling speed of the vehicle traveling in the congested section, and these values are detected by the speed sensor Vi. VF = 40km
It is also possible to use preset values such as / h and VC = 10 km / h.

Using the above equation (2), the average traveling speed vsp of the road p is represented by the following equation (3).

[0042]

[Equation 3]

Here, L represents the distance between the intersections q and q '.

Next, based on the expressions (2) and (3), the green times grq and grq 'of the traffic lights installed at the intersections q and q'.
When the speed sensitivity α indicating the degree of influence of the change on the average speed vsp of the road p is obtained, the speed sensitivity α indicating the relationship between the average speed vsp and the green time grq of the intersection q is expressed by the following equation (4). To be done.

[0045]

[Equation 4]

The equation (4) shows that the larger the value of α, the higher the average speed vsp as the green time of the intersection q increases. That is, the larger the value of α, the greater the change in speed when the green time of the signal is changed.

The relationship between the green time grq 'at the intersection q'and the average speed vsp is the opposite of that of the equation (4). That is, at the intersection q ', if the green time of the intersection q'is increased, the average speed vsp of the road p is decreased.

On the other hand, the speed sensitivity α obtained by differentiating the evaluation value obtained by differentiating the evaluation function FV with the average speed vsp of the arbitrary road p and the average speed vsp of the road p with the green time grq of the intersection q. Since the product of and indicates the relationship between the evaluation function FV and the green time at the intersection q, the relationship between the evaluation function FV and the green time at the intersection q is expressed by the following equation (5).

[0049]

[Equation 5]

From the equation (5), the increasing / decreasing direction of the signal time at the intersection q is determined. That is, the evaluation function FV is calculated according to the equation (5).
The increasing / decreasing direction of the green time at the intersection q is determined so that Then, based on the value obtained according to the equation (5), the blue time at the intersection q is realized by the algorithm shown in the following equation (6).

[0051]

[Equation 6]

Here, TH is a threshold value. This threshold value is set to prevent a sudden change in the blue time, because a sudden change in the blue time poses a safety problem.

The formula (6) is obtained only when the value obtained by the formula (5) exceeds the threshold value TH, that is, when the deviation between the target speed and the actual speed exceeds a certain amount. , A fixed value STEP is added to the present green time, or a fixed value STEP is subtracted from the green time. When the value of the equation (5) is within the range of the threshold value TH, the next signal cycle is performed in the same green time as the previous one.

The fixed value STEP (second) set as the correction step size is determined as follows, for example.

The number of vehicles that can be increased in one signal cycle is one in which the inflow is increased by 1 STEP and the outflow is decreased by 1 STEP, and the increased number of vehicles is two or less per one signal cycle (the upstream green signal increases, and the downstream green signal increases. Condition (when the green signal of is decreased and the number of green signals is increased by about two in one signal cycle), the condition according to the following equation (7) is set.

[0056]

[Equation 7]

Here, k is the inflow / outflow traffic volume (vehicle /
Seconds) are the same for simplicity. And k = 200
Assuming that there are 0 units / 3600 seconds, a value as shown in equation (8) is obtained.

[0058]

[Equation 8]

From equation (8), STEP is appropriate for a few seconds on a normal road.

The data relating to the equations (1) to (8) are set in the lower part 102 of the speed deviation, and the data showing the relationship between the signal time at each intersection and the evaluation function FV is also set in the lower part 102 of the speed deviation. Has been done.

Next, the operation of this embodiment will be described with reference to the flow chart shown in FIG.

First, as road information about each road Pi, when the traveling speed of the vehicle 10 is detected by the speed sensor Vi, the congestion length is detected by the congestion length sensor Li, and the traffic volume is detected by the traffic volume sensor Ki, These pieces of detection information are input to the controller 101 (step 31). When the road information about each road Pi is detected, the controller 101 obtains the deviation between the running speed of each road Pi and the target speed, and also determines the speed sensitivity α indicating the slope of the average running speed of each road with respect to the green time. Calculate (step 32). For example, the speed sensitivity α indicating the slope of the average traveling speed vsp with respect to the green time grp is calculated according to the equation (2). Next, the slope of the evaluation function FV with respect to the green time grp is calculated according to the equation (5) (step 3).
3).

Next, the split change processing of the signal time at each intersection is executed based on the value calculated in step 33 (step 34). In this case, the calculated value is the threshold value ± TH
It is determined whether or not the signal exceeds, and the signal time is adjusted according to the determination result. Further, at this time, it is judged whether or not the corrected signal time exceeds the upper limit value or the lower limit value (step 35), and only when the corrected blue time does not exceed the upper limit value or the lower limit value. Adjust the green time at each intersection according to the modified green time.

For example, the value of STEP is 2 seconds, the target traveling speed of all target roads is 50 km / h, the actual speed of a certain road p is 30 km / h, and ws is all 1,
When the threshold TH is 350, the speed sensitivity α of the road p to the traffic signal q1 at the upstream intersection is 10.0, and the speed sensitivity α of the road p to the traffic signal q2 at the downstream intersection is −1.
If it is 0.0, the condition of the expression (6) is expressed by the following expressions (9) and (10) for each traffic light.

That is, for the traffic signal q1, the condition of the expression (6) is expressed by the following expression (9).

[0066]

[Equation 9]

Regarding the traffic signal q2, the condition of the expression (6) is expressed by the following expression (10).

[0068]

[Equation 10]

From equations (9) and (10), each traffic signal q1,
The green time of q2 is adjusted according to the following equation (11).

[0070]

[Equation 11]

That is, a process of reducing the upstream blue time and increasing the downstream blue time is performed. For example, the upstream blue time is 6
If the downstream green time is 90 seconds at 0 seconds, the upstream blue time is changed to 58 seconds and the downstream blue time is changed to 92 seconds.

When the processing of steps 31 to 35 is completed, the control time t is set to t + T and the time width T is set in order to periodically and continuously perform the processing of steps 31 to 35.
Accordingly, the green time adjustment control is executed for each traffic light. That is, signal adjustment control for stepwise suppressing the speed deviation to zero is executed.

As described above, in the present embodiment, when adjusting the signal time of each traffic light Si of the controlled road Pi, the signal time of each traffic light is taken into consideration based on the road information of each road Pi and the mutual signals are controlled. Since the time is set, the flow of traffic over a plurality of intersections can be made smooth.

In the above embodiment, the signal time of each traffic light is adjusted based on the speed information of the road information. However, the controller 101 is provided with a traffic jam length deviation calculating section and a traffic jam length deviation minimum lower portion. By setting a model related to the traffic congestion length as a traffic flow model, the green time of each traffic light can be adjusted based on the traffic congestion length. For example, the congestion length deviation between the congestion length detected by the congestion length sensor Li and the average target congestion length set for the control target road Pi is calculated, and the congestion length deviation is set to zero based on the calculated congestion length deviation. The signal adjustment time for suppressing the traffic light is calculated for the traffic light at each intersection, and the signal time of each traffic light is adjusted according to the calculated signal adjustment time. Also in this case, it is desirable to periodically perform each process and calculate the signal adjustment time for stepwise suppressing the congestion length deviation to zero.

In the above embodiment, the green time of the traffic light related to the straight road in the road network is adjusted. However, when the green time of the traffic light at each intersection is adjusted, the traffic light paired with each traffic light is used. It is necessary to consider the road information related to the traffic lights whose red time and green time in the signal time are in the opposite relationship to each other. That is, since the traffic signals for each road are installed at the intersection where the vertical road and the horizontal road intersect, if the green time of one traffic light is increased, the green time of the other traffic light must be reduced. As a result, the green time of both traffic lights cannot be increased at the same time.

Therefore, an embodiment in which a constraint condition between the traffic lights (a condition that the green time of both traffic lights cannot be increased at the same time) is added to the algorithm described in the above embodiment will be described below. Then, FIG. 4 shows such a constraint condition between traffic lights and a neural network model for optimizing the evaluation function FV of the expression (6).

In FIG. 4, the output of each neuron of the neural network is made to correspond to the increase / decrease of the green time of each traffic light of the equation (6). That is, when the output of the neuron is 1, the green time of the corresponding traffic light is extended by STEP seconds,
When the output is -1, it is shortened by STEP seconds, and when the output is zero, it is not changed. Furthermore, the condition for avoiding the competition in the same intersection is expressed by the weight of the neural network.

Specifically, the intersection g is provided with a neuron gv in the vertical direction and a neuron gh in the horizontal direction. Neuron gv is intersection g
The speed information of the roads 31, 32, 43, and 44 running in the north-south direction is monitored, and a request for increasing or decreasing the green time is received according to the monitoring result. For example, for the inflow roads 32 and 43, when the speed is lower than the target speed, a request to increase the green time in the north-south direction is received, and conversely, when the speed is higher than the target speed, a request to decrease the green time is received. I am supposed to receive it. In addition, roads 31 and 44 that are outflow routes
For, the sign is reversed, and when the speed is lower than the target speed, it is requested to reduce the green time in the north-south direction, and conversely, when the speed is higher than the target speed, it is requested to increase the green time. .

On the other hand, since the blue time in the north-south direction and the blue time in the east-west direction cannot be increased or decreased at the same time, they are combined with negative weights so that they have opposite polarities. For example, since the vertical green time and the horizontal green time compete with each other, the neuron gv and the neuron gh are connected with a negative weight. Although not shown in FIG. 4 for all weights, two neurons are provided for all the intersections a to i in the same manner as the intersection g, and the speed deviation from each road is the same. It is designed to be input into.

Further, each neuron shown in FIG. 4 has a mutual active connection as shown by the solid line in FIG. 5 or has a mutual inhibitory connection as shown by the broken line. There is. For example, neuron gh and neuron d
v has a mutual active bond, and the neuron gh and the neuron gh have a mutual inhibitory bond.

Here, the active connection is to connect the neurons with positive weights so that a stable state is obtained when the outputs of the two neurons match. When two neurons are connected with a weight w12> 0, the energy E of this neural network is E = -w12u1
It is represented by u2.

In the above equation, u1 and u2 represent the outputs of the neuron 1 and the neuron 2, respectively.

Further, in the above equation, as the output value,
Assuming three possible values of -1, 0 and 1, the most stable point, that is, the state of low energy, is whether the combination (u1, u2) of the outputs of the two neurons is (1, 1). It is the time of (-1, -1). That is, the combination of the outputs of the neurons has the same sign. In this way, connecting each neuron with a positive weight is called active connection.

On the other hand, the connection when the weights of the two neurons are negative is called inhibitory connection. In this case, the most stable point, that is, the state of low energy is when the combination (u1, u2) of the outputs of the two neurons is (-1, 1) or (1, -1). That is, the combination of the outputs of the two neurons has a different sign. That is, connecting two neurons with negative weight is called inhibitory connection.

Since the stable solution of the neural network gives the minimum value of the energy function, if this energy function is made to correspond to the above equation (1), the following (12)
The formula is obtained.

[0086]

[Equation 12]

ΣΣ means adding for i and j.

In the equation (12), in order to select the three-valued logic of increasing or decreasing the green time of the traffic light or maintaining the current state, the judgment output that minimizes F is obtained. That is, the vertical judgment output U _V (j) is selected based on the deviation between the vertical actual speed and the target speed, and the horizontal judgment output U _V is selected based on the deviation between the horizontal actual speed and the target speed.
It is designed to select _H (j). Furthermore, to prevent vertical and horizontal blue time from increasing at the same time,
The mutual relationship in the vertical direction and the horizontal direction is selected by U _H (i) U _V (j).

In equation (12), when j is the index of the traffic signal at the opposite intersection facing the road i, the weight when the green time increase / decrease request is generated from the difference between the target speed and the actual speed is as follows. It is set according to the equation (13). That is, the weight for the vertical green time is expressed by the following equation.

[0090]

[Equation 13]

On the other hand, the weight of the green time in the horizontal direction is
It is expressed by the equation 4).

[0092]

[Equation 14]

Further, in order to prohibit a solution that simultaneously increases the vertical and horizontal green times at the same intersection, the following (1
Expressions 5) to (17) are set.

That is, when a solution that simultaneously increases vertical and horizontal green times at the same intersection occurs, the evaluation function value is significantly deteriorated (increased) in order to prevent unreasonable acceptance of requests. Negative weights are given between the vertical and horizontal neurons at the same intersection. Therefore, the symbol K shown below is set to a value such that the negative value becomes large.

[0095]

[Equation 15]

[0096]

[Equation 16]

[0097]

[Equation 17]

In each of the above equations, for the sake of simplicity, the influence of right-turning and left-turning vehicles can be ignored. Further, in the above formula, Wsi is set as a weight for the optimized road, K is set as an exclusive constraint of vertical and horizontal green time of the traffic signal, and M is set as a weight for determining a degree of correlating the adjacent traffic signal with a neutral traffic signal.

As the solution of the neural network of the above expression (15), the method by Hopfield (JJH) is used.
opfield et al. : Neuralcomp
automation of decisions in o
ptimization problems, Biol
organic Cybernetics, 52, pp.
141-152 (1985)) can be used.

Next, a concrete example using the neural network model will be described. First, a description will be given of a state in which the target speed values of the model shown in FIG. 4 are all 60 km / h and the speed situation is that of FIG. The numbers 32, 18 and 19 in FIG. 6 represent the identification numbers of the roads in FIG. 4, respectively.

In the situation as shown in FIG. 6, first, as an initial state, when the output of each neuron is zero, the evaluation function F as shown in FIG. 7 is obtained. That is, since the outputs of the neurons av to iv and h to ih are 0 and the determination outputs U _V (i) and U _H (i) of the equation (12) are 0, F = 0.
Becomes When the speed decrease of the road 32 is recognized in such a state, the deviation of the speed from the target value is notified from the intersection g to the neurons gv and dv. Here, this deviation signal is multiplied by a certain weight, and is input to the neuron gv as a value of 7, 0, for example. -7,0 is input to the neuron dv.

Next, when the controller 101 minimizes the evaluation function F, the values of each neuron become gv = 1, av = −1, dv = −1, as shown in FIG. It is determined that the vertical green time should be increased and the green times at the intersections a and d should be reduced. As a result, the value of the evaluation function F is reduced to -102. By repeating such processing, the speed of the road 32 gradually recovers as the traffic jam is eliminated.

Next, when the decrease in the speed of the road 18 is recognized, the deviation of the speed from the target value is notified from the intersection d to the neurons av and dv. As a result, the transportation network model shown in FIG. 4 is as shown in FIG. 9, and the evaluation function value at this time is −25 from the expression (12). Next, when the controller 101 minimizes the evaluation function F, the transportation network model becomes as shown in FIG. At this time, the values of the neurons are gv = 1, av = −1, and dv = 1, and the vertical green time of the intersection d is increased to obtain the intersections a and g.
It will be decided that the green time of the above should be reduced. As a result, the evaluation function value is reduced to -98. By repeating such processing, the speed of the road 18 gradually recovers.

On the other hand, when a decrease in the speed of the road 19 is recognized in the process of gradually recovering the speed of the road 18, the deviation from the target value of the speed is notified to the intersection d. At this time,
The transportation network model is as shown in FIG. 11, and the evaluation function value at this time is −84 from the equation (12). Next, when the controller 101 minimizes the evaluation function F, the transportation network model becomes as shown in FIG. At this time, the value of each neuron is av = −1, dv = 0, g
Since v = 1, it is determined that the vertical green time of the intersection a should be reduced and the vertical green time of the intersection g should be increased.
Further, the values of the neurons are dh = 0, eh = -1, and fh =
Since it is 1, it is determined that the horizontal green time of the intersection e should be reduced and the horizontal green time of the intersection f should be increased. As a result, the evaluation function value is reduced to -91.
In this case, a request for increasing the green time and a request for decreasing the green time are simultaneously generated at the intersection d, but these requests are finally arbitrated and it is determined that the green time is the current time. And if such a process is repeated, the road 1
The speed reduction of 8 and 19 gradually disappears.

As described above, in the present embodiment, when a pair of traffic lights provided at the same intersection have a conflict that increases the green time at the same time, the control for increasing the green time is given priority to a road with heavy traffic. Therefore, the green time of each traffic light can be adjusted according to the road conditions.

Next, an embodiment in which the deviation between the target traffic volume and the actual traffic volume is used as the evaluation value will be described with reference to FIG.

In this embodiment, the FQ shown in the following equation (18) is used as a function for evaluating the deviation of the target traffic volume and the actual traffic volume.
To obtain the data related to this function from the controller 101
To enter.

[0108]

[Equation 18]

Here, qsi, qri and qli are traffic volumes within the time interval T of the straight lane, the right lane and the left lane, respectively, and qsoi, qroi and qloi are the time intervals of the straight lane, the right lane and the left lane, respectively. It is the target traffic volume within T. Further, wsi, wry, and wli are weights that determine which of the lanes, straight ahead, right turn, and left turn, should be emphasized and evaluated. Hereafter, in order to simplify the explanation,
The case where only straight traffic is dominant will be described. Further, in the process using the above function, the green time of the signal of each road is determined by using the relationship between the green time of the signal of each road and the evaluation function FQ.

First, a signal from the traffic volume sensor Ki is fetched as road information of each road Pi and the straight-ahead traffic volume qsp (t) of each road is measured (step 900). When data relating to traffic volume is input to the controller 101, the evaluation function FQ is differentiated by the traffic volume of an arbitrary road in order to investigate how the evaluation function FQ is affected by the traffic volume of an arbitrary road. The processing is performed according to the following equation (19).

[0111]

[Formula 19]

On the other hand, the degree of change with the traffic volume qsp of each road when the green time grq of an arbitrary signal is changed is
It is represented by dqsp / dgrq. Therefore, the relationship between the evaluation function FQ and the green time of an arbitrary road is as follows (20)
It is obtained by the differentiation as shown in the formula.

[0113]

[Equation 20]

In the equation (20), if the value of the above equation is positive, the green time is reduced, and if the value of the above equation is negative, the green time is increased. And
Actually, the differential value of dqsp / dgrq can be calculated from the detection value detected by the traffic sensor Ki as shown below, and can be approximated by the following value of sens (p, q).

[0115]

[Equation 21]

Here, Δgrq is the amount of change in the green time from the previous time (before the time T) of the intersection q. ε is a small positive number.

When the value shown in the equation (21) is calculated based on the detection value detected by the traffic sensor Ki, the green time of each signal is changed according to the following equation (22) based on the calculated value. (Step 901).

[0118]

[Equation 22]

Here, β is a small positive number.

As the processing in step 901, for example, in the simplest case, the green time grq (t-T) of a certain previous signal is 40 seconds and the target traffic volume qsop of a certain road p is 80 vehicles / 5 minutes. , The actual traffic volume is 50 vehicles / 5 minutes, and if the previous blue time changed from 40 vehicles / 5 minutes to 50 vehicles / 5 minutes in 43 seconds, sens (p, q)
= (50-40) / (45-43) = 5.0.

Here, assuming that only road p deviates from the target traffic volume and wsp = 1.0, equation (22) is represented by equation (23).

[0122]

[Equation 23]

Here, if the value of β is 0.01, (2
Expression (3) is expressed by the following expression (24).

[0124]

[Equation 24]

When the actual blue time is calculated, it is determined whether the calculated blue time is within the range between the upper limit value and the lower limit value, and the blue time is within the range between the upper limit value and the lower limit value. Only when it is on, the signal time of the traffic light is adjusted according to the calculated green time (step 902). When the calculated green time is out of the range of the upper limit value and the lower limit value, the blue time is set according to the upper limit value or the lower limit value.

In the above embodiment, the green time is calculated according to the equation (22). However, as shown in the equation (6), a method of changing only the fixed time can be applied.

Also in the present embodiment, the flow of traffic over a plurality of intersections can be smoothed as in the case of using speed information as road information, and the influence of traffic congestion can be minimized. it can.

Next, an embodiment using the deviation between the target saturation and the actual saturation as the evaluation value will be described with reference to FIGS. 11 and 12.

In this embodiment, the degree of saturation is the ratio of the actual traffic volume to the maximum traffic volume on the road, and the maximum traffic volume on the road is q.
Let MAX, the actual traffic volume be q, and the saturation ρ, define the saturation ρ according to the following equation (25).

[0130]

[Equation 25]

Further, as a function for evaluating the deviation between the target saturation and the actual saturation, as shown in the following equation (26), FS
AT is used. Then, the data related to these functions are input to the controller 101.

[0132]

[Equation 26]

Here, ρsi, ρri and ρli are saturations within the time interval T of the straight lane, the right turn lane and the left turn lane, respectively, and ρsoi, ρroi and ρloi are the straight lane, the right turn lane and the left turn lane, respectively. It is the target saturation within the time interval T. Also, wsi, wri, wli go straight,
It is a weight that determines which of the lanes of the right turn and the left turn is to be emphasized and evaluated. In the following, for the sake of simplicity of description, a case where only straight-ahead traffic is dominant will be described. Further, in the processing using each of the above formulas, the green time of the signal having the smallest evaluation function value is determined by using the relationship between the green time of the signal of each road and the evaluation function.

First, as shown in FIG. 14, the detection outputs of the speed sensor Vi and the traffic amount sensor Ki are taken in as road information of each road Pi (step 120), and the saturation ρ and speed of each road are sequentially measured.

Here, in order to investigate how the above evaluation function influences the saturation ρ of an arbitrary road, the above evaluation function is differentiated by the saturation ρ of an arbitrary road. It is represented by.

[0136]

[Equation 27]

Here, the traffic volume q (vehicle / second) of the equation (25)
Is the vehicle space density k (vehicle / m) and the average speed v (m /
The maximum capacity is the road-specific capacity qMA.
Since it can be expressed by the product of X and the ratio (%) of the green time gr in one cycle of the signal, the saturation degree ρsp of the road p can be expressed by the following equation (28). That is, the relationship between the road saturation ρ and the green time gr of the signal can be understood by the equation (28).

[0138]

[Equation 28]

Here, grq is the ratio (split) of the straight green signal at the downstream intersection of the road p. Formula (25) ~
From the equation (28), the relationship between the evaluation function FSAT and the signal green time is represented by the following equation (29).

[0140]

[Equation 29]

Here, assuming that the saturation degree ρsp and the velocity vsp have a parabolic relationship as shown in FIG. 15, the saturation degree ρ is expressed by the following equation (30).

[0142]

[Equation 30]

Here, vspMAX is the maximum speed on the road p, and this value is set in advance.

In FIG. 15, the traffic volume qsp (vehicle / second)
Is related to the speed vsp (m / sec) and the running density ksp (vehicles / m) of the vehicle as shown in the following equation (31).

[0145]

[Equation 31]

Here, the vehicle density ks when driving the vehicle
It is assumed that p decreases linearly as the speed vsp increases, as shown in the following equation (31).

[0147]

[Equation 32]

Here, a and b are given positive constants.

Substituting the expression (32) into the expression (31),
The equation (31) is expressed by the following equation (33).

[0150]

[Expression 33]

Here, the values of vsp giving the minimum value of qsp (qsp = 0) are vsp = 0 and vsp = vspM.
It's AX. Therefore, the equation (33) is given by the following (34)
It is represented by a formula.

[0152]

[Equation 34]

Next, by substituting the equation (34) into the equation (33), the equation (35) is obtained.

[0154]

[Equation 35]

When vsp = vspMAX / 2, qs
If p takes its maximum value (maximum traffic volume) grqqspMAX, the following expression (36) is obtained from the expression (35).

[0156]

[Equation 36]

Therefore, the following expression (37) is obtained from the expression (36).

[0158]

[Equation 37]

From equations (35) and (37), the following (38)
The formula is obtained.

[0160]

[Equation 38]

The relationship of the equation (38) is shown in FIG. This relationship exists when the traffic volume is high and the vehicles are sparsely running, and when the vehicles are low-speed and densely running. For example, in FIG. 15, the speed is vspMAX.
When it is / 2, it shows that maximum traffic volume is given,
It shows that the traffic volume decreases above and below that.

Next, the relational expression (2) between the evaluation function FSAT and the green time of each signal is expressed by using the relational expression (30).
By transforming the equation (9), the following equation (39) is obtained.

[0163]

[Formula 39]

Here, in order to simplify the change of the average speed with respect to the change of the blue time, a fixed value β is set, and the assumption shown in the following equation (40) is set.

[0165]

[Formula 40]

In the equation (40), when β> 0, it means that if the green time of the signal is increased, the average speed vsp of the road p is also improved. Further, when -β <0, it means that when the green time of the signal is increased, the average speed vsp of the vehicle on the road p decreases. When it is 0, it means that even if the green time of the signal is changed, the average speed vsp of the vehicle on the road p does not change.

Here, in the expression (39), (1-2vsp
/ VspMAX) (1 / vspMAX) (dvsp / d
If the part of grq) is set to sens (p, q) as the sensitivity, a sequential algorithm for changing the green time of the signal in the direction of decreasing the evaluation function value is expressed by the following equation (41).

[0168]

[Formula 41]

Here, α is a minute positive constant.

Also, sens (p, q) is (39),
From the equation (40), it is represented by the following equation (42).

[0171]

[Equation 42]

Here, a specific example of actually adjusting the green time of the traffic light in step 121 will be described. First, it is assumed that the target target saturation is 0.9 and the maximum speed of all roads is 60 km / h. Also, formula (41)
It is assumed that all the equation weights wsp are 1.0. Further road p
All roads other than the above satisfy the saturation level of 0.9, but only the road p has the saturation level of 0.3 and the speed vsp of 10.
It is supposed to be 0 km / h. Furthermore, β = 9.0,
Let α = 1.0. By substituting these relationships into the equation (42), the sensitivity of the traffic signal q1 downstream of the road p is calculated by the following equation (43).
It is represented by a formula.

[0173]

[Equation 43]

Similarly, the sensitivity of the traffic signal q2 upstream of the road p is expressed by the following equation (44).

[0175]

[Equation 44]

Therefore, it is assumed that the green times grq1 and grq2 are 0.5 and 0.6, respectively (since this embodiment expresses the green time as a ratio within one signal cycle, the maximum value of the green time is Substituting these values into the equation (41), the following equations (45) and (46) are obtained.

[0177]

[Equation 45]

[0178]

[Equation 46]

From the above equations (45) and (46), the green time of the signal of the upstream traffic light p2 is reduced and the green time of the signal of the downstream traffic light q1 is increased.

When the green time according to the saturation ρ is calculated, it is determined whether the calculated blue time is within the range between the upper limit value and the lower limit value (step 122). When the calculated value is within the range between the upper limit value and the lower limit value, the signal time is adjusted according to the calculated green time.
On the other hand, when the calculated green time exceeds the range of the upper limit value or the lower limit value, the blue time is set according to the upper limit value or the lower limit value. After that, the next processing is performed according to the time width T, and the processing of steps 120 to 122 is periodically continued.

In the above embodiment, when the green time is changed, it is changed according to the equation (41). However, as shown in the equation (6), the green time can be adjusted according to a fixed value. .

As described above, also in this embodiment, the flow of traffic can be made smooth over a plurality of intersections, and the influence of traffic congestion can be suppressed to a minimum, as in the above-mentioned embodiments. .

Next, an embodiment for solving the traffic jam by the learning control will be described with reference to FIGS. 16 and 17.

In the present embodiment, it is considered that depending on the nature of the road, it may be difficult to clear the traffic congestion as compared to other roads unless the traffic congestion is cleared preferentially. Generally, learning control is adopted in consideration of the change over time.

Specifically, as shown in FIG. 16, weighting is set for the request set based on the detection signals detected by the speed sensor Vi for detecting road information, the traffic volume sensor Ki, and the congestion length sensor Li. To do. Then, as the weighting, the matrices WSV and WSH are initialized (step 130). For these values, it is appropriate to use the values obtained by the traffic simulation or the preliminary survey in advance, but it is not always necessary.

Next, the weight matrix set over the trial period is randomly changed by random numbers based on the initialized matrix (step 131). For example, if the trial period is seven days, seven sets of weight matrices randomly changed are prepared. The random changes in this case are
For example, the following formula (47) is used.

[0187]

[Equation 47]

Here, wrand (i, j) is w
It represents that (i, j) is randomly changed, and rand (-1, 1) is a uniform random number distributed from -1 to 1, and α is a constant.

Next, the number of days is set to 1, and the first set of weight matrices is taken out (steps 132 and 133). Then, based on the information obtained from each sensor, the traffic signal is controlled according to the traffic information of the day (step 134),
Based on these control results, for example, the total traffic time per day is recorded (step 135).

Such processing is repeated for the number of trial period days, and when the trial period ends (step 13
7) Then, the multiplication processing of the weight matrix group is executed based on the accumulated evaluation values (step 138). Here, selection is to delete the corresponding matrix from the matrix group, and multiplication is to increase the number of the corresponding matrix.

In the processing of this step, as shown in FIG. 17, the traffic signal control is performed using a weighting matrix that changes every day, and an evaluation value group using the information such as the total congestion time obtained as a result is obtained. . Then, according to the value of the evaluation value 141, selection and multiplication of the weight matrix are performed. For example, the matrix numbers p6 and p7 are selected as having bad evaluation values, and the matrix numbers p2 and p3 are multiplied as having high evaluation values. In this case, the matrix is grown so that the number of matrices is constant. Therefore, in this case, the same matrix is duplicated.

Then, the process proceeds to step 139 to perform mating processing. In this process, the weight matrix groups are randomly combined and multiplied. The multiplication process is performed, for example, by using two matrices as w (i, j), q
It is expressed by (i, j) and the following expression (48).

[0193]

[Equation 48]

The value obtained by the equation (48) is a stochastic operation value.

When the trial is repeated for the new weight matrix group resulting from this, the selective multiplication processing is performed again after the trial period. Here, as with the end of the first trial, selection and multiplication of the weight matrix group are performed based on the accumulated evaluation values.

In this way, the weight matrix group is further selected and multiplied based on the new evaluation value.

By continuously performing the above processing, the optimum weight matrix is gradually obtained, which becomes more suitable for the regional characteristics of the controlled object. For example, if the priority order regarding the signal time of the traffic signal is set according to the characteristic of the area, the signal time of each traffic signal can be adjusted according to the characteristic of the area.

[0198]

As described above, according to the present invention,
The speed, traffic volume, congestion length, and saturation are used as road information to optimize the signal times of multiple traffic lights by associating them with each other, so that the flow of traffic across multiple intersections can be smoothed and traffic congestion can be reduced. It can be suppressed to a minimum.

[Brief description of drawings]

FIG. 1 is an overall configuration diagram of a traffic signal control system showing an embodiment of the present invention.

FIG. 2 is a diagram for explaining a traffic flow relationship between a plurality of intersections.

FIG. 3 is a flowchart for explaining an operation when adjusting the green time based on a speed.

FIG. 4 is a diagram for explaining a transportation network model.

FIG. 5 is a diagram for explaining a connection state of neurons used in a traffic network model.

FIG. 6 is a diagram for explaining traffic conditions of a traffic network model.

FIG. 7 is a diagram showing an initial state of a transportation network model.

FIG. 8 is a diagram for explaining a state in which the speed of the road 32 has decreased.

FIG. 9 is a diagram showing a state when a speed decrease of the road 18 shown in the traffic network model is detected.

FIG. 10 is a diagram for explaining a state in which the decrease in speed of the road 18 is being resolved.

FIG. 11: Road 19 shown in the transportation network model
FIG. 6 is a diagram for explaining the speed decrease of FIG.

FIG. 12 is a diagram for explaining a state in which a decrease in speed of roads 18 and 19 is being resolved.

FIG. 13 is a flowchart for explaining the operation when adjusting the signal time based on the traffic volume.

FIG. 14 is a diagram for explaining the operation when adjusting the signal time based on the degree of saturation.

FIG. 15 is a characteristic diagram showing a relationship between traffic volume and speed.

FIG. 16 is a flowchart for explaining learning control of a weight matrix.

FIG. 17 is a diagram for explaining an evolution state of a weight matrix.

[Explanation of symbols]

 P1, P2, ... Pi Speed sensor K1, K2, ... Ki Traffic volume sensor L1, L2 ,.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Yoshiki Kobayashi 7-1-1 Omika-cho, Hitachi City, Ibaraki Hitachi Ltd. Hitachi Research Laboratory (72) Inventor Wataru Hamada ▲ Hiro ▼ Sugata Kanda, Chiyoda-ku, Tokyo 4-6, Hitachi Ltd.

Claims (24)

[Claims] 1. A speed detection process for detecting a speed of a vehicle traveling on a control target road between adjacent intersections of a road network, and a detection speed detected by the speed detection process and a setting for the control target road. Based on the speed deviation calculated by the speed deviation calculation processing and the speed deviation calculation processing for calculating the speed deviation from the average target traveling speed, the signal adjustment time for suppressing the speed deviation to zero is set at each intersection. A traffic signal control method comprising: a signal adjustment time calculation process for a traffic signal; and a signal time adjustment process for adjusting a signal time of each traffic signal on a controlled road according to the signal adjustment time calculated in the signal adjustment time calculation process. 2. A speed detection process for detecting a speed of a vehicle traveling on each of a plurality of control target roads with a road between adjacent intersections of a road network as one control target road, and the speed detection process. A speed deviation calculation process for calculating a speed deviation between each detected speed and an average target traveling speed set for each control target road, and each speed deviation based on each speed deviation calculated in the speed deviation calculation process Signal adjustment time calculation processing for calculating the signal adjustment time for each traffic signal at each intersection, and the signal time of each traffic signal on each controlled road according to the signal adjustment time calculated by the signal adjustment time calculation processing Traffic signal control method including signal time adjustment processing for adjusting traffic. 3. A travel time measurement process for measuring a travel time of a vehicle traveling on a control target road between adjacent intersections of a road network, and a travel time measurement process on the control target road based on a measurement value measured by the travel time measurement process. Speed calculation processing for calculating the actual travel speed of the vehicle, and speed deviation calculation processing for calculating the speed deviation between the calculated speed calculated in the speed calculation processing and the average target travel speed set for the control target road A signal adjustment time calculation process for calculating a signal adjustment time for suppressing the speed deviation to zero based on the speed deviation calculated in the speed deviation calculation process for the traffic signal at each intersection, and the signal adjustment time calculation process And a signal time adjusting process for adjusting the signal time of each traffic signal on the road to be controlled according to the signal adjusting time calculated in step 1. 4. A travel time measuring process for measuring a travel time of a vehicle traveling on each of a plurality of control target roads with a road between adjacent intersections of a road network as one control target road, and the travel time measuring process. Speed calculation processing for calculating the actual traveling speed of the vehicle on each control target road from the measured values, each calculated speed calculated by the speed calculation processing, and the average target set for each control target road The speed deviation calculation process for calculating each speed deviation from the traveling speed, and the signal adjustment time for suppressing each speed deviation to zero based on each speed deviation calculated in the speed deviation calculation process for the traffic signal at each intersection A signal adjustment time calculation process for calculating,
A traffic signal control method comprising: a signal time adjustment process of adjusting a signal time of each traffic signal of each control target road according to the signal adjustment time calculated by the signal adjustment time calculation process. 5. A speed detection process for detecting a speed of a vehicle traveling on a control target road between adjacent intersections of a road network, and a traffic congestion length detection process for detecting a traffic congestion length of a vehicle on the control target road. Traffic volume detection processing for detecting the traffic volume indicating the number of vehicles passing through the control target road within a certain period of time, and each detection value detected by the speed detection processing, the congestion length detection processing, and the traffic volume detection processing Based on the current signal time of each traffic signal on the control target road and the speed sensitivity calculation process for calculating the speed sensitivity indicating the degree of influence of the change in the signal time of each traffic signal on the average speed of the vehicle on the control target road, A speed deviation calculation processing for calculating a speed deviation between the detected speed detected by the speed detection processing and an average target traveling speed set for the control target road, and a speed deviation calculated by the speed deviation calculation processing. Each traffic light Of the speed evaluation process to evaluate in relation to the change of the signal time, suppresses the speed deviation to zero based on the evaluation value obtained by the speed evaluation process and the speed sensitivity calculated by the speed sensitivity calculation process. Signal adjustment time calculation processing for calculating the signal adjustment time for each traffic light at each intersection, and signal time adjustment processing for adjusting the signal time of each traffic light on the controlled road according to the signal adjustment time calculated by the signal adjustment time calculation processing A traffic signal control method including: 6. A speed detection process for detecting a speed of a vehicle traveling on a plurality of control target roads with a road between adjacent intersections of a road network as one control target road, and a traffic jam of vehicles on each control target road. Congestion length detection processing to detect the length, traffic volume detection processing to detect the traffic volume indicating the number of vehicles passing through each control target road within a fixed time, the speed detection processing and the congestion length detection processing and the Based on each detected value detected in the traffic volume detection process and the current signal time of each traffic light on each controlled road, the change in the signal time of each traffic light affects the average speed of vehicles on each controlled road. Speed sensitivity calculation processing for calculating the speed sensitivity indicating the degree, and speed deviation calculation processing for calculating the speed deviation between the detected speed detected by the speed detection processing and the average target traveling speed set for each control target road. And the speed A speed evaluation process that evaluates the speed deviation calculated by the deviation calculation process in relation to the change in the signal time of each traffic signal, an evaluation value obtained by the speed evaluation process, and the speed sensitivity calculated by the speed sensitivity calculation process. Based on the signal adjustment time calculation process for calculating the signal adjustment time for suppressing the speed deviation to zero for each traffic signal at each intersection, each control target road according to the signal adjustment time calculated by the signal adjustment time calculation process Traffic signal control method including signal time adjustment processing for adjusting the signal time of each traffic light. 7. The speed evaluation processing, when evaluating the relationship between each speed deviation and all other speed deviations, assigns each speed deviation a weight determined from the positioning of each controlled road on the transportation network. 6. A process for calculating an evaluation value for minimizing the sum of the weighted speed deviations is included.
Alternatively, the traffic signal control method according to item 6. 8. The process according to claim 1, wherein each process is executed periodically, and the signal adjustment time calculation process includes a process for calculating a signal adjustment time for stepwise suppressing the speed deviation to zero.
The traffic signal control method according to 4, 5, or 6. 9. A congestion length detection process for detecting a congestion length of a vehicle traveling on a control target road between adjacent intersections of a road network, and a congestion length and a control target road detected by the congestion length detection process. In order to suppress the congestion length deviation to zero based on the congestion length deviation calculation process for calculating the congestion length deviation from the average target congestion length set for the congestion length deviation and the congestion length deviation calculation process described above. Signal adjustment time calculation processing for calculating the signal adjustment time of the traffic signal of each intersection, and signal time adjustment processing for adjusting the signal time of each traffic signal of the controlled road according to the signal adjustment time calculated in the signal adjustment time calculation processing, Traffic signal control method including. 10. A congestion length detection process for detecting a congestion length of a vehicle traveling on each of a plurality of control target roads, with a road between adjacent intersections of a road network as one control target road,
In the congestion length deviation calculation process for calculating the congestion length deviation between each congestion length detected in the congestion length detection process and the average target congestion length set for each control target road, and the congestion length deviation calculation process. A signal adjustment time calculation process for calculating a signal adjustment time for suppressing each congestion length deviation to zero on the basis of the calculated congestion length deviation, and a signal adjustment calculated by the signal adjustment time calculation process. A traffic signal control method, comprising: a signal time adjustment process for adjusting the signal time of each traffic signal on each controlled road according to time. 11. The process according to claim 9, wherein each process is performed cyclically, and the signal adjustment time calculation process includes a process for calculating a signal adjustment time for stepwise suppressing the congestion length deviation to zero. Traffic signal control method. 12. A traffic volume detecting process for detecting a traffic volume indicating the number of vehicles passing through a control target road between adjacent intersections of a road network within a certain time, and the traffic volume detecting process. Based on the traffic amount deviation calculated by the traffic amount deviation calculation process for calculating the traffic amount deviation between the specified traffic amount and the average target traffic amount set for the control target road and the traffic amount deviation calculation process, Signal adjustment time calculation processing for calculating the signal adjustment time for suppressing the traffic deviation for each traffic signal at each intersection, and the signal of each traffic signal on the controlled road according to the signal adjustment time calculated by the signal adjustment time calculation processing A traffic signal control method including a signal time adjustment process for adjusting time. 13. A traffic volume detection that detects a traffic volume that indicates the number of vehicles passing through a plurality of control target roads within a certain time period with a road between adjacent intersections of a road network as one control target road. Processing, traffic volume deviation calculation processing for calculating a traffic volume deviation between each traffic volume detected by the traffic volume detection processing and the average target traffic volume set for each control target road, and the traffic volume deviation calculation The signal adjustment time calculation process for calculating the signal adjustment time for suppressing each traffic amount deviation to zero based on the traffic amount deviation calculated in the process for the traffic signal at each intersection, and the signal adjustment time calculation process A traffic signal control method comprising: a signal time adjustment process of adjusting the signal time of each traffic signal on a controlled road according to the signal adjustment time. 14. A traffic volume detection that detects a traffic volume that indicates the number of vehicles passing through a plurality of control target roads within a certain time period, with a road between adjacent intersections of a road network as one control target road. Sensitivity indicating the degree of influence of changes in the signal time of each traffic signal on the average traffic volume of each controlled road based on each traffic volume detected by the traffic volume detection processing and the current signal time of each traffic signal And a traffic volume deviation calculation process for calculating a traffic volume deviation between each traffic volume detected by the traffic volume detection processing and the average target traffic volume set for each control target road. And a traffic volume evaluation process that evaluates the traffic volume deviation calculated by the traffic volume deviation calculation process in relation to the change in the signal time of each traffic light, and the evaluation value and the traffic volume obtained by the traffic volume evaluation process. Calculated by the sensitivity calculation process Signal adjustment time calculation processing for calculating a signal adjustment time for suppressing each traffic deviation to zero for each traffic light at each intersection based on the detected traffic sensitivity, and the signal adjustment calculated by the signal adjustment time calculation processing A traffic signal control method comprising: a signal time adjustment process of adjusting the signal time of each traffic signal on a road to be controlled according to time. 15. The traffic volume evaluation process is determined from the positioning of each control target road on the traffic network when evaluating the relationship between each traffic volume deviation and all other traffic volume deviations. 15. The traffic signal control method according to claim 14, further comprising: a process of weighting and calculating an evaluation value for maximizing a sum of the weighted traffic volume deviations. 16. Each of the processes is executed cyclically, and the signal adjustment time calculation process includes a process of calculating a signal adjustment time for stepwise suppressing the traffic volume deviation to zero.
The traffic signal control method according to 3 or 14. 17. A traffic volume detection process for detecting a traffic volume indicating the number of vehicles passing through a control target road between adjacent intersections of a road network within a certain time, and the traffic volume detection process. Saturation calculation processing for calculating the saturation that indicates the ratio between the traffic volume and the maximum traffic capacity of the controlled road, and the saturation calculated by the saturation calculation processing and the average set for the controlled road Based on the saturation deviation calculated by the saturation deviation calculation processing for calculating the saturation deviation from the target saturation and the saturation deviation calculated by the saturation deviation calculation processing, the signal adjustment time for suppressing the saturation deviation to zero is set. Traffic signal control method including signal adjustment time calculation processing for a traffic signal at an intersection, and signal time adjustment processing for adjusting the signal time of each traffic signal on a controlled road according to the signal adjustment time calculated in the signal adjustment time calculation processing . 18. A traffic volume detection for detecting a traffic volume indicating the number of vehicles passing through a plurality of control target roads within a certain time period with a road between adjacent intersections of a road network as one control target road. Processing, saturation calculation processing for calculating the saturation indicating the ratio of each traffic detected in the traffic detection processing and the maximum traffic capacity of each control target road, each calculated in the saturation calculation processing Saturation degree deviation calculation processing for calculating the saturation degree deviation between the saturation degree and the average target saturation degree set for each control target road, and each saturation based on the saturation degree deviation calculated by the saturation degree deviation calculation processing Signal adjustment time calculation process for calculating the signal adjustment time for suppressing the degree deviation to zero for each traffic light at each intersection, and the signal of each traffic signal on each controlled road according to the signal adjustment time calculated by the signal adjustment time calculation process Time A traffic signal control method including signal time adjustment processing for adjusting each. 19. A traffic volume detection for detecting a traffic volume indicating a number of vehicles passing through a plurality of control target roads within a certain time period with a road between adjacent intersections of a road network as one control target road. Processing, a saturation degree calculation processing for calculating a saturation degree indicating a ratio of each traffic volume detected by the traffic volume detection processing and the maximum traffic capacity of each control target road, and a saturation calculated by the saturation degree calculation processing Saturation degree sensitivity calculation processing to calculate the saturation degree sensitivity indicating the degree of the influence of the change in the signal time of each traffic signal on the average saturation degree of each control target road based on the degree and the current signal time of each traffic signal, and Saturation degree deviation calculation processing for calculating the saturation degree deviation between each saturation degree calculated in the saturation degree calculation processing and the average target saturation degree set for each control target road, and the saturation degree deviation calculation processing Degree of saturation deviation for each signal Saturation evaluation processing to evaluate in relation to the change in signal time of the machine, each saturation degree based on the evaluation value obtained in the saturation evaluation processing and the saturation sensitivity calculated in the saturation deviation calculation processing Signal adjustment time calculation processing for calculating the signal adjustment time for suppressing the deviation to zero for the traffic light at each intersection, and the signal time of each traffic light on each controlled road according to the signal adjustment time calculated by the signal adjustment time calculation processing A traffic signal control method including a signal time adjustment process for adjusting the traffic signals. 20. When evaluating the relationship between each saturation deviation and all other saturation deviations, the saturation evaluation processing is determined from the positioning of each control target road on the traffic network for each saturation deviation. 20. The traffic signal control method according to claim 19, further comprising a process of weighting and calculating an evaluation value for maximizing the sum of the weighted saturation deviations. 21. Each of the processes is performed cyclically, and the signal adjustment time calculation process includes a process of calculating a signal adjustment time for stepwise suppressing the saturation deviation to zero.
8. The traffic signal control method according to 8 or 19. 22. The signal time adjusting process is a process of adjusting a ratio of a blue time to another signal time in each signal cycle when changing the signal time with the blue time, the red time and the yellow time as one signal cycle. Claims 1, 2, 3, 4, 5, including
6, 9, 10, 12, 13, 14, 17, 18 or 1
9. The traffic signal control method described in 9. 23. The traffic light paired with each traffic light as the information that affects the traffic light time of each traffic light on the controlled road, and the red time and the green time included in the traffic light are in the opposite relationship. The road information detection process for detecting road information related to the road information, and the signal adjustment time calculation process includes a process for calculating the signal adjustment time in consideration of the road information detected by the road information detection process. 3,
4, 5, 6, 9, 10, 12, 13, 14, 17, 18
Or the traffic signal control method described in 19. 24. Congestion time detection processing for detecting information on the congestion time of each control target road, congestion information storage processing for sequentially storing information on the congestion time detected by the congestion time detection processing, and congestion information storage processing. And a priority setting process for setting a priority related to the signal time for suppressing the traffic congestion based on the stored information regarding the traffic congestion for each traffic signal, and the signal adjustment time calculation process is set in the priority setting process. 6. A process for calculating the signal adjustment time in consideration of the priority order described above.
The traffic signal control method according to 3, 14, 18 or 19.