JP2009093425A - Traffic stream simulation method and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve the processing efficiency of traffic stream simulation while securing necessary accuracy. <P>SOLUTION: A traffic stream simulation method manages a position of a mobile object that moves on a road at each time and simulates the movement of the mobile object by repeating the processing of calculating the position of each mobile object after a specific time step. A range subject to simulation is divided into several ranges and a value of the time step is varied by the region. Then in the case that the mobile object enters a range with a different time step after one time step, it is preferable to calculate a time that corresponds to the time step of the range which the mobile object has entered and its position at that time. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a technique for simulating the movement of a moving body moving on a road.

  Conventionally, a traffic flow simulation technique for simulating the movement of a vehicle using a computer (information processing apparatus) has been studied. There are several methods for simulating traffic flow, but there is a method of simulating the movement of each vehicle and calculating the position after a minute time step. In this way, the simulation method for simulating the movement of each vehicle is similar to a general particle system simulation such as molecular dynamics, after a minute time step according to an equation that defines the movement of the vehicle (particle) at each time point. Seeking position and condition.

In such a traffic flow simulation that simulates the movement of each vehicle, it is possible to execute a highly accurate simulation by making time steps fine. However, miniaturization of the time step results in an increase in the amount of calculation and requires a great amount of calculation time.
Japanese Patent Laid-Open No. 2004-258889 JP 2005-292945 A

  A traffic flow simulation as described above may be used to simulate the occurrence of an accident. In such a case, in an area (for example, an intersection) to be focused on as an accident occurrence location, it is necessary to perform a simulation with the smallest possible (short) time step to reproduce a micro operation. On the other hand, the vehicle entering the area of interest does not need to be simulated in such a fine time step, and only needs to be able to reproduce a macro motion such as how much vehicle travels on which road.

  As described above, when a simulation is performed in order to pay attention to the movement of the vehicle in a specific area, it is necessary to simulate the movement of the vehicle other than this area. However, there is a problem that the amount of calculation becomes enormous if the movement is simulated with fine time steps for all the vehicles.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to improve processing efficiency while ensuring necessary accuracy in traffic flow simulation.

  In order to solve the above-described problems, a traffic flow simulation method according to the present invention is characterized by simulating the movement of a moving object using different time step values according to a region to be simulated.

  More specifically, the traffic flow simulation method according to the present invention manages the position of each moving body moving on the road at each time, and repeats the process of calculating the position of each moving body after a predetermined time step. A traffic flow simulation method for simulating the movement of a moving body, characterized in that a region to be simulated is divided into a plurality of regions, and the time step value is different for each region.

  In this way, the value of the time step is changed according to the region, the value of the time step is decreased in the region where the operation needs to be simulated in detail, and the value of the time step is increased in the region where rough movement is to be simulated. According to this method, in a region where a detailed operation needs to be simulated, a detailed operation is simulated by reducing the time step value, and the operation is simulated in consideration of the global situation related to the surrounding region. Therefore, it is possible to perform a simulation with high accuracy. At the same time, for the area around the area of interest, the processing time can be shortened by taking a large time step value and reducing the amount of calculation.

  When the simulation is performed by the method as described above, there is a possibility that inconsistency may occur at the joint portion in a region having different time steps. Therefore, in the traffic flow simulation method according to the present invention, when a position after a time step corresponding to a region is calculated for a moving body located in a certain region, the time step is greater than the region in which the moving body was originally located. When entering a short region, it is preferable to perform the following processing. That is, the time when the moving body enters the region having a short time step, the time that matches the short time step, and the position at that time are obtained. Here, the “time matching the time step” refers to a time at which the elapsed time on the simulation is an integral multiple of the time step value.

  Various methods can be considered for obtaining the time that matches the time step and the method of obtaining the position at that time. For example, when it is found that the vehicle enters an area with a short time step, the operation may be simulated again with the short time step from the original position. Alternatively, the position in a short time step may be obtained by interpolation from the original position and the position after the long time step, and the time when the vehicle enters the adjacent area may be obtained.

  Then, when the time and position of entering the region with a short time step are obtained, it is preferable to obtain the next position in this short time step increment thereafter.

  In this way, even when the moving body enters a region having a short time step from a region having a long time step, the operation can be simulated without causing inconsistency.

  In addition, the traffic flow simulation method according to the present invention calculates a time step for a moving body located in a certain area after the time step corresponding to the area, compared to the area where the moving body was originally located. When entering a short region, it is preferable to perform the following processing. That is, the time when the moving body enters the region having a long time step, the time that matches the long time step, and the position at that time are obtained.

  Various methods can be considered for obtaining the time that matches the time step in this case and the method of obtaining the position at that time. For example, the simulation of the operation may be continued in short time step increments until the elapsed time matches the long time step. Alternatively, when entering an area with a long time step, the time until the time that matches the long time step is obtained, and the operation when moving the area continues during that time, and the next time step is met. You may obtain | require the position of the mobile body in time.

  Then, when entering a region with a long time step, if the time and the position at that time are found, the subsequent position is obtained in increments of this long time step.

  In this way, even when the moving body enters a region having a long time step from a region having a short time step, the operation can be simulated without causing inconsistency.

  Note that the present invention can also be understood as a traffic flow simulation method including at least a part of the above processing, or a program for realizing the method. Each of the above means and processes can be combined with each other as much as possible to constitute the present invention. Moreover, this invention can be grasped | ascertained as an information processing apparatus (traffic flow simulation apparatus) which performs said traffic flow simulation method.

  According to the present invention, it is possible to realize a traffic flow simulation with improved processing efficiency while ensuring necessary accuracy.

  Exemplary embodiments of the present invention will be described in detail below with reference to the drawings.

  The traffic flow simulation apparatus according to the present embodiment simulates the traffic of a vehicle by individually calculating the position at each time for each vehicle. Here, the area to be simulated is divided into several areas. The simulation is performed with fine (short) time steps in areas that require highly accurate simulation, and in the surrounding areas. Simulation with coarser (longer) time steps.

  The traffic flow simulation device 1 according to the present embodiment is configured by a normal computer (information processing device) as a hardware configuration, and includes a CPU (central processing unit), a main storage device such as a memory, and an auxiliary storage device such as a hard disk device. And an input device such as a keyboard and a display device such as a display.

  In FIG. 1, the functional block diagram of the traffic flow simulation apparatus 1 which concerns on this embodiment is shown. The traffic flow simulation apparatus 1 functions as a condition input unit 2, a simulation execution unit 3, a result storage unit 4, and a result display unit 5 when a CPU executes a program stored in a memory. Note that some or all of these functional units may be realized by a dedicated chip.

  When a condition is input to the condition input unit 2, the traffic flow simulation device 1 calculates the behavior of the vehicle such as the moving speed, position, and moving direction of the vehicle that the simulation executing unit 3 simulates based on the set condition. The state of the vehicle such as the position and speed at each time calculated by the simulation execution unit 3 is stored in the result storage unit 4. Based on each time stored in the result storage unit 4, the simulation execution unit 3 further calculates the state of the vehicle after a predetermined time step. In this way, the state of the position, speed, etc. after the predetermined time step is repeatedly calculated, and the movement of the vehicle is simulated. The simulation result stored in the result storage unit 4 is displayed on a display device or the like by the result display unit 5.

  Information related to the shape of the road such as the road length, width, number of lanes, and degree of climbing of the road to be simulated is input to the condition input unit 2. In addition, the condition input unit 2 receives information about the vehicle such as the number of vehicles to be simulated and the movement route. The condition input unit 2 also receives a time step (time increment) for performing a simulation.

Here, with reference to FIG. 2, the time step in each area of the road to be simulated will be described. Area A ₃ is an overall area of the road to be simulated (Figure 2 (a)). In this traffic flow simulator 1, aims at simulating the accident in one intersection in the entire area A _3. Thus, defining the intersections of interest as a target area _{A 1} (FIG. 2 (c)). Then, a part of the periphery thereof, to define a peripheral area _{A 2} (Figure 2 (b)). Here, the focus area A _1, a simulation with 10 ms increments. In the peripheral area _{A 2,} the simulation in 100 ms increments. Then, the entire area _{A 3,} the simulation in one second (1000 milliseconds) increments. In this way, the time step Δt _j is determined for each arbitrary area A _j . In addition, although the value of each said time step is an illustration, it defines so that (DELTA) _{tj + 1} may be divisible by (DELTA) _tj . Here, three time steps are prepared and three areas are defined correspondingly, but the number of time steps and areas can be any number of 2 or more.

  Next, details of the simulation processing performed by the traffic flow simulation apparatus 1 will be described with reference to FIG. 3, FIG. 4 and the pseudo program code of Table 1. FIG. 3 is a flowchart showing the flow of the entire simulation process performed by the simulation execution unit 3, and FIG. 4 is a flowchart showing details of the position update process (S14) in FIG. Table 1 shows these processes in pseudo program code.

First, the overall flow will be described with reference to FIG. As an initial process, the simulation execution unit 3 stores the initial states C _i of all the vehicles in the result storage unit 4 (S2, code first line). That is, the vehicle state C _i such as the vehicle position, moving direction, speed, and the like at the simulation start time T _S is received from the condition input unit 2 and stored. In the figure, only the time and position are shown as the vehicle state C _i , and description of other states is omitted.

After the initialization process, the simulation execution unit 3 updates the state (position) of each vehicle at the finest time step (Δt ₁ ) (S4, a loop of lines 3 to 46). At each time (t = k × Δt ₁ ), the following processing is repeated for each vehicle (i = 1 to N) (S6, code 5 to 45 line loop).

First, it is determined whether the time _Tcur of the vehicle state data C _i stored in the result storage unit 4 is later than the current time t (S8, line 6 of the code). In the case where the time T _cur of the vehicle state data C _i is later than the current time t (S8-YES), process the next vehicle without performing the processing for the vehicle (code line 8). On the other hand, the time _{T cur} of the vehicle state data _{C i} may not later than the current time t (S8-NO), i.e. if _{T cur} is equal to t acquires the area where the vehicle i is located (S10, code 13th line). When the area where the vehicle i is positioned and A _a, time step Delta] t _a corresponding to that area is obtained. Then, calculates the position of the vehicle i after Delta] t _a from the current time t (S12, code line 14). In addition, what is necessary is just to employ | adopt a well-known method in the traffic flow simulation as the method of calculating the position after micro time by simulating the motion of a vehicle. Then, based on the position of the calculated time t + Delta] t _a, and updates the position _{C i} of the vehicle i (S14, codes 16-43 line). Although described in detail later in this position update process is performed different processing depending on whether to enter the different areas after Delta] t _a.

It is determined whether all the vehicles have been processed at time t (S16), and if there are any unprocessed vehicles, the process returns to step S6. On the other hand, processing of all of the vehicle returns to the step S4 if the ends, the time after Delta] t _1, the above processing is repeated. In this way, the simulation is executed in increments of Δt ₁ , and the process is terminated when the end time _TE is reached.

Before describing the details of the position update process in step S14, a specific example of the process so far will be described with reference to FIG. Note that FIG. 5 does not consider movement across areas. FIG. 5A shows the position of the vehicle at t = 100 milliseconds. Here, it is assumed that there are two areas A ₁ and A ₂ as areas. Time step Delta] t ₁ of the area _{A 1} is 10 ms, time step Delta] t ₂ areas _{A 2} is 100 ms. Two vehicles X and Y are located in areas A ₂ and A ₁ , respectively. In the figure, white circles represent the position of the vehicle at the current time (t = 100).

  First, for t = 100, the processes of steps S6 to S16 are performed.

Since the time of the latest data of the vehicle X is t = 100 at time t = 100, a negative determination is made in step S8, and the position at the next time step is calculated. Since the vehicle X is located in the area _{A 2,} time step Delta] t ₂ (= 100 milliseconds after) that is, the position at t = 200 ms is calculated. The calculated position is indicated by a black circle.

At time t = 100, since the time of the latest data of the vehicle Y is t = 100, a negative determination is made in step S8, and the position of the vehicle Y at the next time step is similarly calculated. Since the vehicle Y is located in the area _{A 1,} the position at time step Delta] t ₁ (= after 10 ms) ie t = 110 is calculated.

In this way, the process at t = 100 seconds is completed. Next, after Δt ₁ , that is, at t = 110, the processes of steps S6 to S16 are performed. The positional relationship of each vehicle at t = 110 is shown in FIG.

  Since the time of the latest data of the vehicle X is t = 200 at t = 110, an affirmative determination is made in step S8, and the position of the vehicle X at the next time is not calculated. On the other hand, since the time of the latest data of the vehicle Y is t = 110, a negative determination is made in step S8, and the position at the next time step t = 120 is calculated.

  Similarly, at t = 120 to 190, only the vehicle Y calculates and updates the position at the next time step. Then, at t = 200 (FIG. 5C), which is equal to the time of the latest data of the vehicle X, the positions at the next time step are calculated for both the vehicles X and Y.

At each time, the position after a time step of a certain vehicle is obtained based on the state of the surrounding vehicle (position, traveling direction, speed, etc.) at that time. Here, for example, the vehicle (Y) in the area A ₁ has state information at that time, such as t = 110, but the vehicle (X) in the area A _{2 with} a coarser time step than that. Does not have state information at that time. In such a case, the vehicle in the area A _1, a method for calculating the position after time steps are considered the following methods.

The first method is a method of ignoring the influence of a vehicle in an area whose time step is coarser than that area when calculating the position of the next time for the vehicle in a certain area. That is, only the influence of vehicles in the same area and an area with a finer time step than the area (area A _j to areas A _{1 to} A _j ) is considered. According to this method, the above problems can be solved.

  In the second method, when the position of the next time is calculated for vehicles in a certain area, the state at that time is calculated by interpolation (interpolation) for vehicles in an area whose time step is coarser than that area. This is a method using this. For example, at t = 110 shown in FIG. 5B, the position of the vehicle X at the time is not obtained. However, the position and the like of the vehicle X are obtained at times before and after t = 110 (t = 100 and 200). Therefore, the position at t = 110 is estimated by interpolation from the position at t = 100 and 200, and this is used. Although the calculation process is larger than that in the first method, the simulation can be performed with high accuracy because the influence of the vehicle in the area with rough time steps can be taken into consideration. In addition, since the estimation by interpolation has a much smaller calculation amount than the actual simulation, the processing efficiency is improved as compared with the simulation with fine time steps for all areas.

As described above, by changing the time step for updating the position information for each area, it is possible to improve the efficiency due to the omission of the calculation amount, but there is a possibility that inconsistency may occur at the joint portion of the areas having different time steps. This will be described with reference to FIG. In Figure 6, the vehicle X at time t = 200, are located in an area _{A 2.} Since the time step of the area A ₂ is Δt ₂ = 100 milliseconds, the position of the vehicle X at t = 300 is calculated after t = 200. Here, the position at t = 300 is calculated based only on the situation at time t = 200. As shown in FIG. 6, when the vehicle X enters the area A ₁ with finer time steps at t = 300, the position of the vehicle X is affected by the vehicle Y in the area A ₁ at t = 210 to 290. The position is determined without consideration. This is, within the area A _1, and contrary to the purpose of running a simulation in small time step, it is necessary to deal with the time of entering the different areas of the time step. Although FIG. 6 shows a case where the time step enters an area with a fine time step from an area with a coarse time step, it is necessary to deal with a case where the time step shifts from an area with a fine time step to an area with a coarse time step.

  In step S14 of the flowchart of FIG. 3, the position update process is changed depending on whether the calculated position after the predetermined time step is in the same area or in a different area. FIG. 4 is a flowchart showing details of the position update process in step S14.

First, in step S1400, determination is made whether (code 17,21,32 line) calculated position when the t + Delta] t _a calculated in step S12 is located which area. Area after movement, if it is identical to the original area A _a, the process proceeds to step S1402, updates the calculated position as is (code line 19).

In t + Δt _a, if you enter the small area _{A a-1} of the time step, the process proceeds to step S1404. In step S1404 from S1410 (code 21 to 30 line), the time between the time t to t + Delta] t _a, for each time step _{Δt a-1 (Δt a)} / (Δt a-1) is divided, the vehicle Among times when i enters the area A _a-1 , a time satisfying t + j × Δt _a-1 (where j = 1,..., Δt _a / Δt _a-1 ) is obtained (such time Is referred to as a time that matches the time step Δta _-1 ). Position of the vehicle at time _{t + j × (Δt a /} Δt a-1) may be determined by interpolation (interpolation) from the position at time t and t + Delta] t _a. Then, the position information of the vehicle i is updated as C _i (T _cur = t + j × Δt _a−1 , P _cur = calculated position).

This will be specifically described with reference to FIG. Vehicle X is located in the area _{A 2} at time t = 200. At time t = 200, the next time step, that is, the position at t = 300 is calculated. The position at t = 300 is the area A ₁ with a finer time step than the area A ₂ . Therefore, the position at t = 200 and the position at t = 300 are divided by Δt ₂ / Δt ₁ = 100/10 = 10 to obtain the time when the vehicle entered the area A ₁ . Then, it can be seen that enters fourth, that is, the area _{A 1} at t = 240. Therefore, the position information of the vehicle X is updated at t = 240. In addition, after t = 240, the vehicle X is also updated every time step Δt ₁ = 10 milliseconds.

Next, in step S1400, it calculates a position when the t + Delta] t _a calculated in step S12, the case entering the coarse time step area _{A a + 1.} In this case, the process proceeds from step S1400 to step S1414. In step S1414, the position and the like at time t + Δta _{+ 1} that matches the time step of the area Aa _{+ 1} after the transition are obtained (codes 33 to 41). Specifically, from the time t + Delta] t _a to time t ₊ Δt _a + _1, assuming that moves at the same speed, the moving distance from the time t to t + Delta] t _a, if multiplied by directly Δt _{a + 1} / Δt _a, obtains the position at time t ₊ Δt _a + ₁ be able to. Then, the position information of the vehicle _i is updated as C _i (T _cur = t + Δta _{+ 1} , P _cur = calculated position) (S1416).

This will be specifically described with reference to FIG. Vehicle X is located in the area _{A 1} at time t = 340. At time t = 340, the position at the next time step, ie t = 350, is calculated. The position at t = 350 is the area A ₂ having a coarser time step than the area A ₁ . Therefore, the position of the vehicle X at t = 400, which is the time after t = 350 and matches the time step Δt ₂ = 100 milliseconds, is calculated. Here, assuming that the movement from t = 340 to 350 continues for the subsequent 50 milliseconds, the movement distance is further multiplied by 5, and the position at t = 400 is calculated.

  As described above, the position information update process in step S14 is performed. As described above, even when shifting to an area having a different time step value, the position information at the time matching the time step corresponding to the area after the transition is calculated, so inconsistency is also detected at the time of area transition. The simulation can be continued without waking up.

  Thus, the traffic flow simulation apparatus 1 according to the present embodiment can reduce the amount of calculation by taking a large time step in the area around the area of interest. Moreover, since the movement of the vehicle in the peripheral area of the area of interest is also simulated, the operation of the vehicle in the area of interest can be executed with high accuracy.

It is a figure which shows the functional block of the traffic flow simulation apparatus which concerns on this embodiment. It is a figure explaining the area | region used as the object of simulation, and the time step in each area | region. It is a flowchart which shows the flow of a traffic flow simulation process. It is a flowchart which shows the flow of a detailed process of a position update process. It is a figure explaining basic control of traffic flow simulation processing. It is a figure explaining the problem in the junction part of the area from which a time step differs. It is a figure explaining a position update process when approaching into an area with a fine time step from an area with a rough time step. It is a figure explaining a position update process at the time of approaching into an area with a rough time step from an area with a fine time step.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Traffic flow simulation apparatus 2 Condition input part 3 Simulation execution part 4 Result memory | storage part 5 Result display part

Claims (6)

A traffic flow simulation method for simulating the movement of a moving body by managing the position of each moving body moving on a road at each time and repeating the process of calculating the position of each moving body after a predetermined time step,
The area to be simulated is divided into multiple areas,
The time step value is different for each region.
A traffic flow simulation method characterized by that. The traffic flow simulation method according to claim 1,
When the position after the time step corresponding to the area is calculated for a moving body located in a certain area, if the mobile body enters an area having a shorter time step than the area where the moving body was originally located,
The time when the moving object enters the area where the time step is short and matches the short time step and the position at that time are calculated, and after that time, the next position is calculated for each time step of the entered area. To
A traffic flow simulation method characterized by that. The traffic flow simulation method according to claim 1 or 2,
When the position after the time step corresponding to the area is calculated for a mobile object located in a certain area, if the mobile object enters an area having a longer time step than the area where the mobile object was originally located,
The time when the moving body enters the area with a long time step and the time that matches the long time step and the position at that time are obtained, and after that time, the next position is calculated for each time step of the entered area. To
A traffic flow simulation method characterized by that. Position management means for managing the position of the moving body moving on the road at each time;
Motion simulation means for calculating the position of each moving body after a predetermined time step;
A traffic flow simulation device that simulates the movement of the moving body by repeating the calculation of the position of each moving body by the motion simulation means,
The area to be simulated is divided into multiple areas, and the time step value corresponding to each area is different.
The motion simulating means calculates, for each moving body, a position after a time step according to a region where the moving body is located.
A traffic flow simulation device characterized by that. The traffic flow simulation device according to claim 4,
When the position after the time step corresponding to the area is calculated for a moving body located in a certain area, if the mobile body enters an area having a shorter time step than the area where the moving body was originally located,
The motion simulating means obtains a time when the moving body enters an area where the time step is short and matches the short time step, and a position at that time, and after that time, every time step of the area where the mobile object enters To calculate the next position,
A traffic flow simulation device characterized by that. The traffic flow simulation device according to claim 4 or 5,
When the position after the time step corresponding to the area is calculated for a mobile object located in a certain area, if the mobile object enters an area having a longer time step than the area where the mobile object was originally located,
The motion simulating means obtains a time when the moving object enters an area where the time step is long and matches the long time step, and a position at that time, and after that time, every time step of the area where the mobile object enters To calculate the next position,
A traffic flow simulation device characterized by that.

Images