JP2000113365A - Traffic condition prediction simulation device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To set an optimum parameter to an online prediction system. SOLUTION: An input data memory 9 storing the respective traffic quantities of respective positions, a result data memory 10 storing a result traffic condition at every time, a parameter memory 8c storing various parameters which are operated/inputted, an input time designation means 20 designating the input time of the traffic quantities of the respective positions, a simulation operation part 11 applying a prediction operation to the traffic quantities stored in the input data memory 9 corresponding to the designated input time by using a prediction program and the various parameters stored in the parameter memory and obtaining a simulation prediction traffic condition and a comparison part 16 comparing the obtained simulation predicted traffic condition and the result traffic condition stored in the result data memory are installed.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an online traffic situation prediction system for predicting the traffic situation at each position on a road network in an online state, and more particularly to a traffic situation prediction simulation incorporated in the online traffic situation prediction system. Related to the device.

[0002]

2. Description of the Related Art For example, traffic conditions such as a vehicle traffic volume per unit time, a vehicle speed, a vehicle density, and a degree of congestion at each position on a complicated road network as shown in FIG. An online traffic situation prediction system that is used as basic data for traffic control work is configured, for example, as shown in FIG.

[0003] The traffic volume collecting device 2 includes a vehicle traffic volume, a vehicle speed, and a vehicle speed measured by a detector 1 disposed at each position on the road network shown in FIG. vehicle density, traffic volume I ₁ including the congestion degree _{like, I 2, I 3, ...} , I n
(Each suffix indicates information for each location)
The data is transmitted to the prediction calculation unit 3 at a constant cycle T.

In the prediction program memory 4, for example,
A known prediction program for predicting traffic conditions such as an input / output method, a neural network method, and a section density method is stored.

[0005] Further, in the parameter memory 5, various parameters for actually performing the prediction calculation using the prediction program are stored. In order to improve the prediction accuracy of the traffic condition in accordance with the change of the traffic flow accompanying the change of the road shape, various parameters stored in the parameter memory 5 need to be frequently changed. The setting section 6 is provided so that various parameters stored in the parameter memory 5 can be arbitrarily set and changed.

[0006] prediction computation unit 3, the traffic each traffic volume of each position on the road network from the amount collection device 2 are sequentially input at a predetermined period _{T I 1, I 2, I} 3, ..., predicted for I _n A prediction calculation is performed using the prediction program stored in the program memory 4 and the various parameters stored in the parameter memory 5, and after a lapse of a predetermined time Td such as 30 minutes from each input time t of the constant period T, for example. Predicted traffic situation J ₁ , J at each position on the road network at each time (t + Td)
_{2, J 3, ..., J} n ( each suffix indicating the information of each position) is sent to the traffic control center 7 sequentially determined.

The predicted traffic conditions J ₁ , J at each position
₂ , J ₃ ,..., J _n include a predicted traffic amount, a predicted vehicle speed, a predicted vehicle density, a predicted congestion degree, and the like at the corresponding position.

The traffic control center 7 calculates traffic congestion information based on the predicted traffic conditions J ₁ , J ₂ , J ₃ ,..., J _n at each position sequentially input from the prediction operation unit 3 of the online traffic condition prediction system. It is possible to output a display on an electric display panel or to implement various traffic regulations.

[0009]

However, the on-line traffic situation prediction system shown in FIG. 3 still has the following problems to be solved.

That is, as described above, each parameter stored in the parameter memory 5 needs to be set to an optimum value in order to improve the prediction accuracy of the traffic situation. However, in the online traffic situation prediction system, many types of parameters are required, and it is unclear which parameter should be corrected and how.

Next, specific parameters will be described with reference to FIGS. Here, the input / output method will be described as an example.

That is, in the online traffic situation prediction system which measures the traffic volume of each section A, B and C in the road network shown in FIG. 2 and predicts the traffic situation at each position, FIG.
As shown in (a), a process of flowing traffic volume from the upstream section (section A → section B → section C) is performed (upstream calculation). If the flow is excessive, as shown in FIG. Return the excessive flow amount (backup amount) (section C → section B)
→ Process (downstream operation) of section A) is performed.

In this case, when pouring as shown in FIG. 4A, the section traffic capacity is set as a parameter for limiting the pouring amount. In addition, Q (traffic volume) shown in FIG. 5 is used as a parameter used at the time of return shown in FIG.
-K (density) pattern is set.

In the upstream operation shown in FIG. 4A, when the section traffic capacity of section A as a parameter is QA and the section traffic capacity of section B as another parameter is QB, traffic from section A to section B is obtained. The quantity can flow up to QA. Further, it is possible that only the maximum QB flows from the section B to the section C.

In the downstream operation shown in FIG. 4B, when there is a backup amount from the section B to the section A, the backup amount BA is obtained, and when there is a backup amount from the section C to the section B, the backup amount BB is obtained. Is returned.

The QK pattern as a parameter used when calculating the backup amounts BA and BB is shown in FIG.
As shown in the figure, the traffic volume per unit time (vehicle traffic volume) Q
And the density K indicated by the number of vehicles existing at a unit distance on the road. That is, traffic Q and density K
Is small, the traffic Q increases as the density K increases. However, if the density K exceeds a certain limit, the inter-vehicle interval becomes narrower, the vehicle speed decreases, and the traffic Q decreases. , Traffic jams. Therefore, the QK pattern as this parameter is a parameter for determining the conditions for occurrence of traffic congestion in predicting traffic conditions.

What is needed in this QK pattern is the traffic congestion side (right curve). On its section when traffic volume Q ₁ is flow which means that there can be only a vehicle on a road to a density K _1.

In such an online traffic situation prediction system, by reducing the traffic capacity QA, QB in each section as a parameter shown in FIG. 4 (a), the flow of vehicles in each section is suppressed. A traffic jam phenomenon can be generated upstream from the suppressed section.

On the other hand, Q- as a parameter shown in FIG.
The same phenomenon can be obtained by reducing the curve on the traffic congestion side of the K pattern, and the traffic congestion phenomenon can be created.

As described above, by appropriately changing the parameters, even if the same traffic volume is input, it is possible to control the prediction of the traffic situation in which the traffic jam occurs.

However, in the conventional system shown in FIG.
The change of each of these parameters was made by the person in charge at Shiijo without confirming the predicted results after the change. For this reason, changing the parameters has been an extremely nervous operation. As a result, the parameter tuning operation is avoided, and a highly accurate prediction result has not been obtained.

The present invention has been made in view of such circumstances, and the parameters and input values are determined under the same environment as an online traffic situation prediction system that actually implements the traffic situation at each position on the road network online. It is an object of the present invention to provide a traffic situation prediction simulation apparatus capable of changing a traffic volume, performing a repetitive prediction calculation offline, and easily setting optimal parameters for an online traffic situation prediction system.

[0023]

A traffic condition prediction simulation apparatus according to the present invention has a prediction program, various parameters and a prediction operation unit incorporated therein, and the prediction operation unit operates on a road network which is sequentially input at a constant period. A prediction calculation is performed for each traffic volume at each position using a prediction program and various parameters, and a predicted traffic situation at each position on the road network at each time after a lapse of a predetermined time from each input time of a fixed cycle. It is built into the online traffic situation forecasting system, which is determined sequentially.

The traffic situation prediction simulation apparatus according to the present invention comprises: an input data memory for storing the traffic volume at each position on the road network which is input at a fixed cycle, with an input time added thereto; Result data memory for storing the actual traffic situation at each time at each position on the road network to which the input time is attached, a parameter memory for storing various parameters input by operation, and the input time of the traffic volume at each position Input time designating means for designating the traffic volume stored in the input data memory corresponding to the input time designated by the input time designating means. And performing a prediction calculation using the simulation prediction at each position on the road network at a time after a lapse of a predetermined time from the designated input time. A simulation operation unit for obtaining a traffic condition, and a comparison unit for comparing the simulation predicted traffic condition obtained by the simulation operation unit with the actual traffic condition at an input time after a lapse of a predetermined time from the designated input time stored in an actual data memory. And

In the traffic situation prediction simulation apparatus thus constructed, the traffic volume at each position inputted to the online traffic situation prediction system is written into the input data memory with the input time. The actual traffic situation at each time at each position is written to the actual data memory. As the prediction program for executing the prediction calculation, the same prediction program as the prediction program of the online traffic situation prediction system is used.

Therefore, for the same input traffic volume at each position, the prediction calculation can be repeatedly performed by changing the parameters. Since the actual traffic situation at each time is stored and held in the actual data memory, each time the simulation prediction operation is performed, the simulation prediction operation result is compared with the actual value so that the prediction operation result approaches the actual value. Just change the parameters.

Further, in the traffic situation prediction simulation apparatus incorporated in the online traffic situation prediction system of another invention, each traffic volume at each position on the road network operated and input with the designated input time is stored. A prediction operation is performed on the input data memory and the traffic volume with the designated input time stored in the input data memory using a prediction program and various parameters, and a time after a predetermined time has elapsed from the designated input time. And a simulation calculation unit for obtaining a simulation predicted traffic situation at each position on the road network.

In the traffic situation prediction simulation device configured as described above, the input data memory includes:
For example, each traffic volume at each position on the road network operated and input by the person in charge is stored.

The same prediction program and the same parameters as the prediction program of the online traffic situation prediction system are used for the prediction program and parameters for executing the prediction calculation.

Therefore, even in an emergency situation that cannot be implemented by the online traffic situation prediction system, the traffic situation at each position can be predicted in advance, and it is possible to prepare measures in advance when an abnormality occurs.

[0031]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing a schematic configuration of an online traffic situation prediction system in which a traffic situation prediction simulation apparatus according to one embodiment of the present invention is incorporated. The same parts as those of the conventional online traffic situation prediction system shown in FIG. 3 are denoted by the same reference numerals, and detailed description of the overlapping parts will be omitted.

The traffic volume collecting device 2 includes a vehicle traffic volume, a vehicle speed, and a vehicle speed measured at a fixed period T such as every five minutes by each detector 1 disposed at each position on the road network shown in FIG. vehicle density, traffic volume I ₁ including the congestion degree _{like, I 2, I 3, ...} , I n
(Each suffix indicates information for each location)
The data is transmitted to the prediction calculation unit 3 at a constant cycle T.

Further, the traffic volumes I ₁ , I ₂ , I _{3 ,.}
Are sent to the predictive operation unit 3 and written with time t in the input data memory 9 in time series.
Furthermore, the traffic volume _{I 1, I 2, I 3} , ..., I n is actual traffic conditions P ₁ at each time _{t, P 2, P 3,} ..., P
_{As n,} it is written to the performance data memory 10.

In the prediction program memory 4, for example,
A known prediction program for predicting traffic conditions such as an input / output method, a neural network method, and a section density method is stored. The prediction program stored in the prediction program memory 4 is supplied to the prediction operation unit 3 and the simulation operation unit 11.

Further, in the parameter memory 6, various parameters for actually performing the prediction calculation using the prediction program are stored. Specifically, the common parameters 8a commonly used by the prediction calculation unit 3 and the simulation calculation unit 11, the online parameters 8b used only by the prediction calculation unit 3, and the only used by the simulation calculation unit 11 are: The simulation parameter 8c is stored.

The operator can arbitrarily set and change the value of the online parameter 8b by operating the parameter setting section 6 such as a keyboard. Furthermore, the operator can arbitrarily set and change the value of the simulation parameter 8c by operating the simulation parameter setting unit 12 such as a keyboard.

The prediction calculation unit 3 receives one
Each traffic volume at each position on the road network which is sequentially input at a fixed period T
I₁, I_Two, I_Three, ..., I_nForecast program
Prediction program and parameter memo stored in memory 4
Common parameters 8a stored in the library 8 and online
The prediction calculation is performed using the
A predetermined time T such as 30 minutes from each input time t of the cycle T
Each position on the road network at each time (t + Td) after elapse of d
Traffic situation J₁, J_Two, J_Three, ..., J
_n(Each suffix indicates the information of each position)
To the prediction value file 13 once. This predicted value file
Predicted traffic situation J written in file 13₁, J _Two, J_Three,
…, J_nIs read out and sent to the traffic control center 7.

The predicted traffic conditions J ₁ , J at each position
₂ , J ₃ ,..., J _n include a predicted traffic amount, a predicted vehicle speed, a predicted vehicle density, a predicted congestion degree, and the like at the corresponding position.

The traffic control center 7 calculates traffic congestion information based on the predicted traffic conditions J ₁ , J ₂ , J ₃ ,..., J _n at each position sequentially inputted from the predicted value file 13 of the online traffic condition prediction system. Display output on the lightning board and implement various traffic regulations.

In contrast to such an online traffic situation prediction system for predicting a basic traffic situation in an online state, the traffic situation prediction simulation apparatus of the embodiment is
It has the following configuration and operates.

(1) Each traffic volume I ₁ , I ₂ , I at each position at each input time t stored in the input data memory 9
_3, ..., each movement amount of each position in the input time t designated by the input time specifying section 14 of the I _n I _1, I _2,
I _3, ..., only I _n is sent to the simulation unit 11.

The simulation unit 11, the traffic volume I ₁ of each position on the road network at time t designated inputted from the input data memory _{9, I 2, I 3,} ..., I n
, A prediction operation is performed using the prediction program stored in the prediction program memory 4 and the common parameter 8a and the simulation parameter 8c stored in the parameter memory 8, and for example, 30 minutes from the input time t.
Simulated traffic conditions R ₁ , R ₂ , R ₃ ,..., R _n at each position on the road network at a time (t + Td) after a lapse of a predetermined time Td such as minutes (each suffix indicates information of each position). It is sequentially obtained and written into the predicted value file 15.

The simulated predicted traffic conditions R ₁ , R ₂ , at each position written in the predicted value file 15
R ₃ ,..., R _n are sent to the comparison unit 16. Comparison section 16
The simulation predicted obtained by simulation unit 11 traffic situation _{R 1, R 2, R 3} , ..., R n and the actual data memory 10 to the stored specified input time t from the time of input after a predetermined time Td elapses ( The actual traffic conditions P ₁ , P ₂ , P ₃ ,..., P _{n at (} t + Td) are compared. And
The comparison result Za is displayed and output to the comparison result output unit 17.

[0045] In this case, traffic I ₁ particular input time _{t, I 2, I 3,} ..., can be repeated simulations predicted by specifying I _n.

(2) Each traffic volume I ₁ , I ₂ , I ₃ ,..., I inputted at a constant period T from the traffic volume collection device 2.
_n can be sent to the prediction operation unit 3 and input directly to the simulation operation unit 11 in real time via the input data memory 9. in this case,
The same traffic I ₁ and the prediction computation unit 3 and the simulation unit _{11, I 2, I 3,} ..., will be carried out prediction operation on I _n.

Then, the predicted traffic conditions J ₁ , J ₂ ,
J _3, ..., J _n and simulation predicted traffic condition R ₁ of the simulation unit 11 stored in the predicted value file _{15, R 2, R 3,} ..., are compared in comparator 18 and R _n. Then, the comparison result Zb is output to the comparison result output unit 17.
Display output to

In this case, the parameters 8b and 8c used
Only different prediction calculations are performed, and it is possible to grasp the contribution ratio of the parameter when there is a difference between the calculated predicted traffic conditions.

(3) Further, the comparison unit 19
Predicted traffic situation J obtained by the prediction calculation unit 3₁, J
_Two, J_Three, ..., J_nIs stored in the performance data memory 10
Input time after a lapse of a predetermined time Td from each input time t
Actual traffic situation P of (t + Td) ₁, P_Two, P_Three, ..., P
_nCompare with Then, the comparison result Zc is compared with the comparison result.
Display output to the output unit 17.

In this case, it is possible to easily verify how much the online traffic condition predicted value matches the actually measured value.

(4) For the input data memory 9
The traffic volume I collected by the traffic volume collection device 2₁,
I_Two, I_Three, ..., I_n Not the traffic input device.
Input time t input via the device 20
Traffic volume H₁, H_Two, H_Three, ..., H _n Can be written
It is.

In this case, the operator sets the same parameters as the online parameters 8b in the parameter memory 8 as the simulation parameters 8c.

Then, the simulation operation unit 11
Traffic H ₁ of each position the operator has arbitrarily designated, H _2, H
_3, ..., the input time after a predetermined time Td elapses H _n (t +
Td) Simulation predicted traffic conditions R ₁ , R ₂ , R
₃ ,..., R _n are obtained.

In the traffic situation prediction simulation apparatus thus constructed, the traffic volumes I ₁ , I ₂ , I ₃ ,... At each position on the road network which are sequentially input to the online traffic situation prediction system at a constant period T. with I _n is input to the prediction computation unit 3, the input time t is written into the input data memory 9 are assigned.

The actual traffic conditions P ₁ , P ₂ , P ₃ ,..., P _n at each time t at each position are stored in the actual data memory 1.
Written to 0. Then, the simulation operation unit 11
The same prediction program as the prediction program for performing the prediction calculation in the prediction calculation unit 3 of the online traffic situation prediction system is used as the prediction program 4 for executing the prediction calculation.

Therefore, as a traffic situation prediction simulation device, the same input traffic volume I ₁ ,
I _2, I _3, ..., with respect to I _n, the simulation parameters 8c sequentially changed, can be carried out repeatedly simulation prediction calculation.

The actual traffic conditions P ₁ , P ₂ ,
Since P ₃ ,..., P _n are stored and held in the actual data memory 10, every time the simulation prediction calculation is performed, the simulation prediction traffic conditions R ₁ , R ₂ , R ₃ ,.
, R _n and the actual traffic conditions P ₁ , P ₂ , P ₃ ,..., P _n may be compared, and the parameter 8c may be changed so that the prediction calculation result approaches the actual value.

As a result, the optimal simulation parameter 8c can be determined in a short time, and the determined optimal simulation parameter 8c may be set as the online parameter 8a of the online traffic situation prediction system.

As described above, the prediction accuracy of the online traffic situation prediction system can always be maintained at a high level.

Further, the person in charge stores the traffic volumes H ₁ , H ₂ , H ₃ ,
..., it can be arbitrarily set the H _n. The same prediction program and the same parameters 8c as the prediction program of the online traffic situation prediction system are used as the prediction program and the simulation parameters 8c for executing the simulation prediction calculation.

Therefore, the simulation prediction at each position on the road network is performed using the traffic volume at each position on the road network when heavy rain, heavy snow, a major accident, etc. are assumed, which rarely occur on the actual road network. Traffic condition ₁ ,
_{R 2, R 3, ...,} R n is easily obtained.

Therefore, even in an emergency situation that cannot be implemented by the online traffic situation prediction system, the traffic situation at each position can be predicted in advance, so that it is possible to prepare measures in advance when an abnormality occurs.

[0063]

As described above, in the traffic situation prediction simulation apparatus of the present invention, the traffic situation at each position on the road network is actually performed online under the same environment as the online traffic situation prediction system. For the same input traffic volume, it is possible to carry out off-line prediction calculations by changing parameters. Therefore, optimal parameters can be easily set for the online traffic situation prediction system.

The input traffic volume can be arbitrarily set by the operator. Therefore, it is possible to make a prediction assuming a situation that cannot be implemented by the online traffic situation prediction system, and the application range of the prediction can be greatly expanded.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic configuration of an online traffic situation prediction system in which a traffic situation prediction simulation device according to an embodiment of the present invention is incorporated.

FIG. 2 is a diagram showing a road network to be predicted;

FIG. 3 is a block diagram showing a schematic configuration of a conventional online traffic situation prediction system.

FIG. 4 is a diagram illustrating an example of parameters used in a prediction calculation;

FIG. 5 shows one of parameters used in the prediction calculation.
Diagram showing -K pattern

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Detector 2 ... Traffic volume collection device 3 ... Prediction calculation part 4 ... Prediction program memory 6 ... Parameter setting part 7 ... Traffic control center 8 ... Parameter memory 9 ... Input data memory 10 ... Actual data memory 11 ... Simulation calculation part 12: simulation parameter setting unit 13, 15, prediction value file 14: input time designation unit 16, 18, 19 ... comparison unit 17: comparison result output unit 20: traffic volume input device

Claims (2)

[Claims] 1. A prediction program, various parameters, and a prediction operation unit are incorporated therein. The prediction operation unit is configured to execute the prediction program and the prediction operation for each traffic volume at each position on a road network sequentially input at a constant period. An online traffic situation prediction system that performs a prediction operation using various parameters and sequentially obtains a predicted traffic situation at each position on the road network at each time after a lapse of a predetermined time from each input time of the fixed cycle. An integrated traffic situation prediction simulation device, comprising: an input data memory for storing, with an input time, each traffic volume at each position on the road network input at the constant cycle; An actual data memory for storing the actual traffic conditions at each position on the road network at each time with an input time, and storing various parameters input by operation Parameter memory, input time specifying means for specifying the input time of the traffic volume at each position, and for the traffic volume stored in the input data memory corresponding to the input time specified by the input time specifying device, A prediction calculation is performed using the prediction program and various parameters stored in the parameter memory, and a simulation prediction traffic situation at each position on the road network at a time after a predetermined time has elapsed from the designated input time. A simulation operation unit to be obtained, and a comparison unit that compares the simulation predicted traffic condition obtained by the simulation operation unit with the actual traffic condition at an input time after a lapse of a predetermined time from the designated input time stored in the actual data memory. Traffic condition prediction simulation device provided with. 2. A prediction program, various parameters, and a prediction operation unit are incorporated therein. The prediction operation unit is configured to execute the prediction program and the prediction operation for each traffic volume at each position on the road network which is sequentially input at a constant period. An online traffic situation prediction system that performs a prediction operation using various parameters and sequentially obtains a predicted traffic situation at each position on the road network at each time after a lapse of a predetermined time from each input time of the fixed cycle. An integrated traffic situation prediction simulation device, comprising: an input data memory for storing each traffic volume at each position on a road network input with a designated input time; and a designated input stored in the input data memory. A prediction calculation is performed on the traffic volume with time using the prediction program and the various parameters, and a predetermined calculation is performed from the designated input time. Traffic situation prediction simulation device and a simulation unit for obtaining the simulation predicted traffic condition at each position on between the road network of the time after a lapse.