WO2022180901A1

WO2022180901A1 - Simulation method and device

Info

Publication number: WO2022180901A1
Application number: PCT/JP2021/034336
Authority: WO
Inventors: 啓明中田
Original assignee: 日立Astemo株式会社
Priority date: 2021-02-24
Filing date: 2021-09-17
Publication date: 2022-09-01
Also published as: DE112021006082T5; JPWO2022180901A1

Abstract

Provided are a simulation method and device which, when reproducing on a simulation a scenario for evaluating a vehicle that operates autonomously, can easily enable, through natural behavior that could occur in an actual traffic environment, a state which satisfies initial conditions of the scenario such as position, speed, and acceleration at a start point of the scenario. A predicted time (predicted scenario start time) at which a given vehicle (host vehicle 200) arrives at a scenario start position, which is a position designated at the scenario start point, is established. On the basis of the predicted time and the position and speed of a designated moving object, a movement plan for the moving object is formulated. A simulation that includes the given vehicle and the moving object is started using the formulated movement plan. It is determined at a given timing during execution of the simulation whether there is a change in the predicted time. If the predicted time has changed (there is a change in the predicted time), the movement plan for the moving object is updated according to the change in the predicted time.

Description

Simulation method and apparatus

The present invention relates to a simulation method and apparatus used for vehicle evaluation.

In recent years, technology development for autonomous driving of vehicles has been carried out as driving support technology and autonomous driving technology. These technologies need to respond to various situations that may occur during operation, and a great many evaluations are required to ensure safety. However, it is becoming difficult to conduct all evaluations in running tests using actual vehicles. Therefore, it is essential to carry out evaluation by simulation instead of running tests using actual vehicles.

As an evaluation method when a vehicle operates autonomously, there is a flow of identifying scenarios that require evaluation and conducting tests using the identified scenarios. To conduct a test using a scenario, it is necessary to determine the parameters that specifically define the scenario, and then reproduce the experiment that satisfies the conditions. That is, even when performing a simulation, in order to reproduce a scenario that satisfies the conditions, means for instructing the behavior of the vehicle to be evaluated and its surrounding vehicles is required.

　Patent Document 1 discloses a method using a scenario description language as a means of instructing the behavior of the vehicle to be evaluated and the surrounding vehicles. In the scenario description language, it is possible to define a condition and have the execution of some action or some other conditional decision wait until the condition is met.

Japanese Patent Publication No. 2020-502615

In order to reproduce a scenario for evaluating an autonomously operating vehicle on a simulation, traffic participants (hereafter referred to as moving objects and moving bodies However, at the start of the scenario, the initial conditions of the scenario, such as position, velocity, and acceleration, must be met. Furthermore, since the target vehicle continuously checks the surrounding conditions before the scenario starts, the surrounding traffic participants go through the natural behavior that can occur in the actual traffic environment and reach the initial condition of the scenario. There is a need to. Moreover, since the vehicle to be evaluated operates autonomously, its behavior itself can change depending on the situation of surrounding traffic participants. Therefore, the time at which the vehicle to be evaluated reaches the position where the scenario starts, that is, the scenario start time also changes, so it is difficult to predetermine the movement of surrounding traffic participants fixedly according to the scenario.

In principle, it is possible to define some conditions and change and adjust the parameters that define the speed and acceleration of surrounding vehicles when the conditions occur in order to respond to changes in the scenario start time. However, considering the relationship between many moving objects, it is necessary to define the conditions for generating the desired state at the scenario start position and the parameter change contents when the conditions occur without making each moving object move unnaturally. is very difficult and difficult to adjust.

The present invention has been made in view of the above circumstances, and when reproducing an evaluation scenario on a simulation, the positions and behaviors of surrounding traffic participants at the start of the scenario can be easily reproduced from the parameters at the start of the scenario. To provide a simulation method and apparatus capable of making the behavior of each traffic participant in the surroundings leading to the start of a scenario look natural.

In order to solve the above problems, a simulation method according to the present invention is a simulation method for simulating vehicle behavior, comprising: position and speed of the target vehicle at the start of a scenario for verifying the behavior of the target vehicle; , specifying the position and speed of a moving object existing in the vicinity of the target vehicle, obtaining a predicted time at which the target vehicle will reach the scenario start position, which is the specified position at the start of the scenario, and determining the target vehicle; adjusts the movement of the moving object by limiting the speed change range per specified time so that the position and speed of the moving object become specified values at the predicted time when the will reach the scenario start position. and

Further, a simulation apparatus according to the present invention is a simulation apparatus for realizing a simulation method for simulating vehicle behavior, and includes the position and speed of the target vehicle at the start of a scenario for verifying the behavior of the target vehicle, and specifying the position and speed of a moving object existing around the target vehicle, obtaining a predicted time at which the target vehicle will reach the scenario start position, which is the specified position at the start of the scenario, and The movement of the moving object is adjusted by restricting a speed change range per specified time so that the position and speed of the moving object become specified values at the predicted time when the scenario start position is reached. do.

According to the present invention, when reproducing an evaluation scenario on a simulation, the positions and behaviors of the surrounding traffic participants at the start of the scenario can be easily reproduced from the parameters at the start of the scenario. The behavior of traffic participants can also be made natural.

Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

4 is a flowchart showing an example of a simulation execution procedure; Schematic diagram showing an example of a road used for simulation execution. The figure which shows the example of a self-vehicle driving definition. Schematic diagram showing an example of definition of a scenario start condition. FIG. 4 is a diagram showing an example of speed plans (movement plans) of other vehicles; FIG. 5 is a diagram showing an example of updating speed plans (movement plans) of other vehicles; The figure which shows the example of the relationship between the time of the own vehicle, and a mileage. FIG. 2 is a configuration diagram of an apparatus that executes an example of a method of generating an acceleration command for another vehicle; The figure which shows the structural example of a simulation apparatus. FIG. 10 is a diagram showing an example of the relationship between each process of the simulation execution procedure in FIG. 1 and each part of the simulation apparatus in FIG. 9;

Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. In the following description, the evaluation target vehicle is expressed as the own vehicle. A surrounding traffic participant (moving object, moving object) may use more specific expressions such as other vehicles, bicycles, and pedestrians. In addition, it is assumed that the own vehicle has a driving support function and an automatic driving function to be evaluated, and the driving support function and the automatic driving function are collectively referred to as a driving control function.

[Simulation method]
A procedure for implementing an evaluation method (simulation method) by simulation according to the present invention will be described with reference to FIG. An evaluation method (simulation method) according to the present invention is a method for simulating vehicle behavior.

(Own vehicle travel definition creation S110)
First, the own vehicle traveling to be evaluated is defined in own vehicle traveling definition creation S110. For example, when evaluating the road shown in FIG. When the 200 evaluates an event occurring at the scenario start position Le (201), the travel of the own vehicle 200 is defined by the method shown in FIG.

As shown in FIG. 3, the definition of running the own vehicle is divided into one or more steps, and each step is specified with a period for executing the step, a target speed, and a target lateral position. As the period, it is conceivable to specify the travel distance, the arrival at the specified travel position, the duration of the step, or the like. If a more complicated definition is desired, a specified condition, such as continuing up to a specified speed or higher, may be used. The contents of each step shown in FIG. 3 will be described later.

(Own vehicle running timing acquisition S120)
In own vehicle traveling timing acquisition S120, the relationship between when and where the own vehicle 200 is traveling from the start of the simulation (that is, the relationship between the time and position or traveling distance of the own vehicle 200 of the vehicle to be evaluated) is acquired. . For example, the simulation is executed in a state where only the own vehicle 200 is defined as a moving body to be handled in the simulation (in other words, the simulation is executed by running the own vehicle 200 alone as the evaluation target vehicle), and the time and the own vehicle 200 The relationship between the position and travel distance of is obtained in advance. Here, the time is virtual time handled in the simulation world. Hereinafter, unless otherwise specified, "time" refers to virtual time handled in the simulation world, and "time" is also time handled in the simulation world.

When executing the simulation in the own vehicle travel timing acquisition S120, pre- and post-processing necessary for executing the simulation, such as initialization of the simulation, is also performed. The simulation is performed under the same conditions as the evaluation scenario, including road conditions, except that moving bodies other than the own vehicle 200 are excluded from the simulation. However, there may be a case where surrounding traffic participants are not included in the simulation even if they are stationary objects. For example, it is conceivable that parked vehicles that have little effect on the running of the own vehicle 200 are not included in the simulation.

(Movement planning of surrounding traffic participants S130)
In S130, the movement plan of the surrounding traffic participants, movement plan (hereinafter sometimes referred to as speed plan or travel plan) of each moving body other than the own vehicle 200 related to the scenario, that is, the position and orientation at the start of the simulation. , formulate the relationship between the time, velocity and direction at the time of simulation execution. At this time, the simulation time at which the own vehicle 200 passes the scenario start position Le (201) obtained in the own vehicle travel timing acquisition S120 (hereinafter referred to as the scenario start time for verifying the behavior of the own vehicle 200 of the evaluation target vehicle, Alternatively, a movement plan is formulated so that each moving object has a value specified as the state (position, speed, etc.) at the start of the scenario at the estimated scenario start time. In other words, by referring to the relationship between the time and the position of the own vehicle 200 obtained by running the own vehicle 200 as the evaluation target vehicle alone in the own vehicle travel timing acquisition S120, the own vehicle 200 starts the scenario. The simulation time at which the vehicle 200 reaches the position Le (201) is calculated (determined). It adjusts the movement (also called running or behavior) of the moving body so that it becomes the value specified as the state of time (position, speed, etc.). The specified values can be position, velocity, acceleration, direction, etc., but it is also possible to add other elements, such as adding an arm swing condition in the case of a pedestrian.

Below, a simulation including the own vehicle 200 and each moving object is started using the movement plan formulated in S130 for the movement plan formulation of the surrounding traffic participants. From simulation execution initialization S140 to simulation result recording S180, a simulation of a scenario (vehicle evaluation scenario) to be actually evaluated is executed. As the travel definition for the own vehicle 200 in this simulation, the same travel definition as the travel definition used in the own vehicle travel timing acquisition S120 is used.

(Simulation execution initialization S140)
In simulation execution initialization S140, various initializations necessary for simulation are performed. This initialization also includes setting the initial states of the own vehicle 200 and surrounding traffic participants.

The range from simulation 1-step execution S150 to simulation end determination S170 is a portion that is repeatedly executed when executing the simulation.

(Simulation 1 step execution S150)
The simulation 1-step execution S150 is a process of advancing the time by the increment time of the simulation. Here, as the time advances, a process of updating all the parameters representing the physical phenomenon to be simulated corresponding to the advanced time is performed. In order to simulate the physical phenomena related to vehicle running while maintaining a certain degree of accuracy, we assume that the increment time is about 1 ms. However, it is conceivable to change the increment time according to the purpose.

The parameters that represent the physical phenomena of each surrounding traffic participant follow the respective movement plans as much as possible. decide. However, it is also possible to make adjustments prioritizing collision avoidance with the own vehicle 200 and other traffic participants, and then control so as to follow the movement plan only when there is no priority event.

(Movement plan update of surrounding traffic participants S160)
In step S160 for updating the movement plan of the surrounding traffic participants, the scenario start time is re-predicted (estimated) from the running position of the own vehicle 200 at the simulation time (scenario start time), and each surrounding traffic participant is updated at the predicted scenario start time. Review (update) the movement plan of each surrounding traffic participant so as to satisfy the behavior (specified position, speed, etc.) at the start of the scenario. In other words, based on the running position of the vehicle 200 at the simulation time (scenario start time), the estimated time at which the vehicle 200 will reach the scenario start position, which is a specified position at the scenario start time, is calculated again. Each peripheral traffic participant has a specified position and velocity at the predicted time when 200 reaches the scenario start position. Coordinate the movement of traffic participants. Furthermore, in other words, it is determined whether the predicted time has changed from the running position of the own vehicle 200 at the simulation time (scenario start point), and the predicted time has changed (the predicted time has changed). In this case, the movement plan of each peripheral traffic participant is updated according to the change in the predicted time, and the movement of each peripheral traffic participant is adjusted.

The vehicle 200 may change its behavior (position, speed, etc.) from the travel obtained in the vehicle travel timing acquisition S120 due to changes in the circumstances of surrounding traffic participants, so it is necessary to re-predict the scenario start time. In addition, if feedback control is simply applied to the speed of surrounding traffic participants according to changes in the scenario start time, the behavior (position, speed, etc.) of surrounding traffic participants will become unnatural, and the behavior at the start of the scenario will not change. Since it may become impossible to satisfy the conditions, the movement plan is reviewed, and deviations from the movement plan are compensated for by feedback control.

(Simulation end determination processing S170)
The simulation end determination process S170 determines the end of the simulation by detecting the elapse of a predetermined period of time after the start of the scenario or the arrival of the vehicle 200 at a predetermined position. If the simulation is not to be ended, the process returns to the simulation 1 step execution S150 to continue the execution of the simulation.

In determining the end of simulation execution, whether before or after the start of the scenario, the vehicle 200 collides (contacts) with something such as a surrounding traffic participant, which is a moving object existing in the vicinity of the vehicle 200, or when there is a surrounding traffic participant. For example, a situation in which a collision (contact) occurs between surrounding traffic participants, or a situation in which the speed of the own vehicle 200 or surrounding traffic participants is outside the expected range (deviates from the predetermined speed range). It may also be determined that the simulation is finished when a case where the evaluation fails or the evaluation result is NG is detected, such as when the evaluation continues for a certain period of time or more. Even if the conditions at the start of the scenario are not satisfied at the start of the scenario, it may be considered as an evaluation failure, and the simulation may be determined to end. However, it is difficult to completely match conditions such as position, speed, and acceleration at the start of the scenario due to various factors. That is, when the own vehicle 200 passes the scenario start position, which is the position specified at the start of the scenario, the state of the own vehicle 200 and surrounding traffic participants is checked, and the conditions at the start of the scenario are determined to be within a predetermined range or more. Even when deviation occurs, the evaluation may be considered to have failed, and the simulation may be determined to end.

By adding such a judgment to the simulation end judgment processing S170, it is possible to reduce the useless simulation execution time after the evaluation result NG is determined, and the own vehicle 200 stops due to circumstances such as collision avoidance, and the simulation does not end. can avoid the problem. Such functions are important for automatic and continuous execution of a large number of simulations.

(Simulation result recording S180)
In the simulation result record S180, it is recorded whether the evaluation result of the simulation is NG, whether there is an evaluation-unacceptable case, or whether the evaluation result is OK. At the time of recording, a log of parameters representing various states of the own vehicle 200 and surrounding traffic participants at the end of the simulation, and main parameters such as position and speed during the simulation execution period are also recorded, and used later. It can also be used for analysis.

(Specific example of self-vehicle travel definition creation S110)
An example of the travel definition method of the own vehicle 200 will be described with reference to FIG. 3, assuming that the vehicle 200 travels on the road shown in FIG. Note that the contents, values, etc. of each step in FIG. 3 are an example, and the present invention is not limited to this.

Each step in FIG. 3 is normally executed sequentially according to the order of the step numbers. However, it is conceivable that when a specific condition specified in advance occurs, a direct transition to the specified step is made. In each step, execution continues for the period specified by "Period". The “period” can be expressed using time, running distance, running position, and the like.

At each step, the vehicle 200 is instructed to aim for the values specified by the target speed and target lateral position as long as there is no problem. The target lateral position is indicated using the travel lane and the deviation from the lane center within the travel lane. It is assumed that when an event occurs that takes priority over the target, such as the risk of collision or compliance with traffic rules, the own vehicle 200 will give priority to responding to those events.

In the example shown in FIG. 3, the vehicle 200 is expected to run as follows.

First, step 1 corresponds to running on the merging road 460 from the position Le0 (211) at the start of the simulation. run. However, the driving control function may judge that driving at 40 km/h is not appropriate due to a sharp curve, etc., and may decelerate.

After completing the confluence 460 and entering the acceleration lane 470, accelerate to 100 km/h according to the instructions in step 2. At this time, as a result of judging that the operation control function is unavoidable, such as hindrance to merging with the main line 410 due to the influence of vehicles running on the main line, etc., it is possible to run at a speed other than 100km / h. do not have.

After traveling the acceleration travel distance in step 2, step 3 is entered, and the host vehicle 200 moves to the lane of the main line 410 designated by the target lateral position and to the lateral position within the lane over five seconds. Lane identification information and a lateral position offset within the lane are used to specify the lateral position. Here, the lanes are designated as Lane A and Lane B in order from the right side in the direction of travel, and the left side in the direction of travel is treated as the positive direction for the offset of the lateral position within the lane.

When adjusting the lateral position to the target, it is assumed that the vehicle 200 performs steering control so as to match the target in a specified period in principle. However, the host vehicle 200 may delay the start of the lane change by the operation control function according to the situation of the destination lane. It doesn't matter if it slows down or accelerates. Also, there may be a case where the lateral movement cannot be completed within the specified period due to the relationship with the surroundings. It is also possible to use a method of adding attributes to the target lateral position, specifying the maximum steering angle and required travel time as separate attributes, and performing steering control according to the values indicated by the attributes regardless of the period specified in this step. I do not care.

After completing the movement to the destination lane and the horizontal position within the lane, in step 4, travel within the lane to the acceleration/deceleration start position immediately before the evaluation scenario. When it is determined that the operation control function is unavoidably stopped due to the relationship with other vehicles while traveling in the lane, the vehicle may decelerate or accelerate from the target speed.

When step 4 ends, step 5 is entered and the host vehicle 200 attempts to accelerate to the target speed of 120 km/h toward the start of the scenario. This does not mean that the host vehicle 200 is required to reach 120 km/h by the start of the scenario, but simply sets the target to 120 km/h. Therefore, it may not reach 120km/h when the scenario start position is reached, and in order to avoid contact with other vehicles, the driving control function may run at a speed of 120km/h or less. . If the speed has not reached 120 km/h when the vehicle 200 reaches the scenario start position, 120 km/h is also set as the target speed in step 6, so the vehicle 200 continues to accelerate. Scenarios during acceleration can also be evaluated. If the target speed in steps 5 and 6 specifies a speed lower than the speed in the final state in step 4, the scenario in which the host vehicle 200 is decelerating can be evaluated.

Up to step 5 corresponds to the preparation run before the start of the evaluation scenario. By the preparatory running up to step 5, the position, speed, and acceleration corresponding to the conditions at the start of the scenario for verifying the behavior of the own vehicle 200 can be adjusted.

When the own vehicle 200 reaches the scenario start position Le (201) specified in the period of step 5, the process proceeds to step 6, and running corresponding to the scenario is started. In step 6, the target velocity continues to be specified in step 5, and the target lateral position is changed. That is, as long as there is no problem in running, the own vehicle 200 tries to move to a position 5 cm to the right from the center of lane A in 5 seconds.

When step 6 is completed, step 7 specifies the same state as step 6. If the target lateral position is not changed from the previous step and the lateral movement of the vehicle 200 is not completed for some reason, the period specified in step 6 is not affected and the period specified in step 6 is applied. Continue the pace of lateral movement.

In addition to securing time as a continuation period for unfinished lateral movement, step 7 also has the meaning of a period for confirming whether stable driving can be performed even after scenario execution is completed. Even if the scenario itself ends normally, a dangerous situation such as a remarkably narrow gap between vehicles is possible at the end of the scenario, and there is a possibility of collisions with nearby traffic participants. Continue to check the status after completion of execution for a while.

(Concrete example of movement plan formulation S130 for surrounding traffic participants)
An example of formulating a movement plan for surrounding traffic participants will be described. Assuming the scenario start conditions shown in FIG. 4, FIG. 5 shows an example of a movement plan for other vehicles (other vehicles) as surrounding traffic participants. In the moving plan formulation example shown in FIG. 5, only the vertical direction (direction along the road) plan is shown as the speed plan, but if necessary, it is also possible to add a horizontal position plan or a direction plan with respect to the track. be done. In this example, the target is up to the start of the scenario, and only the lane in which the other vehicle is traveling is considered without considering the lateral position offset. After the scenario starts, the surrounding traffic participants will reproduce the behavior according to the acceleration, deceleration, lane change, etc. defined in the scenario. A movement plan different from that of the vehicle may be used according to the behavior characteristics of each of the surrounding traffic participants. For example, it is conceivable to prepare a travel plan for pedestrians, a travel plan for bicycles, and the like.

By using a velocity plan as a movement plan, in other words, by formulating a movement plan based on velocity, the movement distance or position can be calculated by first-order integration, and the acceleration can be calculated by first-order differentiation. In other words, there is an advantage in that it is easy to plan considering the acceleration, and it is easy to guarantee the continuity of the speed.

In the scenario shown in FIG. 4, the scenario starts when the vehicle 200 reaches the position Le (201), and the vehicle 200 at the start of the scenario has a velocity Ve (204) and an acceleration Ae (206). A parked vehicle 360 is stopped on the road shoulder at position Lp (361) on the road (see also FIG. 2). Specifying the position Le ( 201 ) of the vehicle 200 at the start of the scenario is necessary to define the positional relationship with the parked vehicle 360 . Also, when the curvature and gradient of the road are changing, where on the road the scenario starts is important for evaluation under desired road conditions.

Other vehicle X (310) and other vehicle Y (320) exist as surrounding traffic participants. At the start of the scenario, another vehicle X (310) travels in lane A behind host vehicle 200 by a distance Df1 (312) at speed Vf1 (314) and acceleration Af1 (316). Another vehicle Y (320) travels in lane B ahead of host vehicle 200 by a distance Df2 (322) at speed Vf2 (324) and acceleration Af2 (326).

Fig. 5 shows an example of a speed plan for other vehicles to be formulated. A speed plan needs to be formulated for each of the other vehicle X (310) and the other vehicle Y (320). An example of formulating a speed plan will be described with reference to FIG.

As shown in FIG. 5, the speed plan is determined in the order of constant speed (time Tp0 (530) to Tp1 (531)), acceleration change (time Tp1 (531) to Tp2 (532) period) and constant acceleration (time Tp2 (532) to Tssi (550) period). Acceleration change is treated as the behavior of the period required for the vehicle to change the acceleration as a quadratic function, that is, assuming that the acceleration changes linearly.

The time Tssi (550) is the (predicted) time obtained as the time when the vehicle 200 reaches the scenario start position Le (201) in the vehicle travel timing acquisition S120, and is the estimated scenario start time at the beginning of the simulation. becomes.

Calculation of velocities Vfp0 (511), Vfp1 (512), Vfp2 (513) and times Tp0 (530), Tp1 (531), and Tp2 (532) to define changes in velocity 510 of other vehicle X (310) Is required. Since the speed Vf (515) is the speed at the start of the scenario, it is equal to the speed Vf1 (314) (FIG. 4) in the speed plan of the other vehicle X (310).

　Calculation of the velocities required to define the change in velocity 510 is performed in order from the initial scenario start prediction time Tssi (550) in the backward direction.

First, considering the acceleration or deceleration travel time of the other vehicle X (310) which is considered sufficient for evaluating own vehicle 200, time Tp2 (532) is determined so as to be before time Tssi (550). do. The time from time Tp2 (532) to time Tssi (550) may be fixed to 5 seconds, 10 seconds, or the like.

Once the time Tp2 (532) is determined, the acceleration Af (516) at time Tssi (550), that is, the amount of change in speed, is defined as the acceleration Af1 (316) (FIG. 4) from the scenario start conditions, so the speed Vfp2 (513 ) is found. If the speed Vfp2 (513) is out of the assumed range as the traveling speed of the other vehicle X (310) (deviates from the predetermined speed range), the time Tp2 (532) is brought closer to the time Tssi (550). to adjust the speed Vfp2 (513) to be within the expected range.

After calculating the time Tp2 (532) and the speed Vfp2 (513), the time Tp1 (531) and the speed Vfp1 (512) are calculated. First, in order to calculate time Tp1 (531), the time from time Tp1 (531) to time Tp2 (532) is determined. This time is determined by proportional calculation, table reference, etc., based on the acceleration desired to be changed in the relevant period. Alternatively, it is conceivable to use a fixed value more simply. If the time from time Tp1 (531) to time Tp2 (532) is obtained, time Tp1 (531) can be calculated from time Tp2 (532).

　In order to calculate the speed Vfp1 (512), a function indicating the relationship between the time and the speed during the period between the time Tp1 (531) and the time Tp2 (532) is obtained. Since the relationship between time and speed during the period from time Tp1 (531) to time Tp2 (532) is defined as a quadratic function, the function indicating the relationship can be determined by calculating the 0th to 2nd order coefficients. These coefficients are obtained under the condition that the velocity and acceleration are changed continuously at time Tp1 (531) and time Tp2 (532). Note that the acceleration is 0 at time Tp1 (531) and Af (516) at time Tp2 (532), and that time Tp1 (531) and time Tp2 (532) have already been calculated.

It can be seen that the velocity Vfp1 (512) can be calculated by Equation 1 below from the function showing the relationship between the time and the velocity during the period from time Tp1 (531) to time Tp2 (532).

As shown in Equation 1, the speed Vfp1 (512) is subject to restrictions in relation to the speed Vfp2 (513) depending on the acceleration Af (516). If Af (516) is 0, then Vfp1 (512) must equal velocity Vfp2 (513). When Af (516) is positive (acceleration), the speed Vfp1 (512) is always smaller than the speed Vfp2 (513). must be made larger. If the speed Vfp1 (512) is out of the assumed range as the traveling speed of the other vehicle X (310) (deviates from the predetermined speed range), the time Tp1 (531) is set to be within the assumed range. It is necessary to either approach Tp2 (532), or determine that such a scenario will not occur and treat it as non-evaluation.

The speed Vfp0 (511) is equal to the speed Vfp1 (512) because the period from time Tp0 (530) to time Tp1 (531) is constant speed.

For time Tp0 (530), the traveled distance up to time Tssi (550) calculated based on the speed plan is the traveled distance to the scenario start position Le (201) on the road used in the simulation, such as the road shown in FIG. Then, the traveling distance from the position (the position of the other vehicle at the start of the simulation) L0 (301) to the position (scenario start position) Le (201) is calculated to be equal.

Depending on the scenario start conditions, time Tp0 (530) may be later than time Tp1 (531) or time Tp2 (532). In such a case, the traveling distance during the period from time Tp0 (530) to time Tssi (550), which can be calculated from the speed plan from time Tp1 (531) to time Tssi (550), is calculated from the position L0 (301) to the position Le ( 201), the time Tp0 (530) should be calculated so as to match the traveled distance.

When executing the simulation, the other vehicle corresponding to the speed plan is caused to start traveling from position L0 (301) at time Tp0 (530). The speed of the other vehicle at the start of travel is set according to the speed plan, not in the stopped state.

Depending on the scenario start conditions, the time Tp0 (530) may be negative, that is, the time before the start of the simulation. In such a case, the time Tp0 (530) is set to time 0, the traveled distance during the period from time 0 to time Tssi (550) is calculated from the speed plan, and from the scenario start position Le (201) The other vehicle may start traveling from a position ahead of the travel distance to cope with the situation.

If the other vehicle has a speed from the beginning at the start of the simulation, the own vehicle 200 suddenly detects the other vehicle that is running at the start of the simulation, and if the operation control function of the own vehicle 200 is in an unexpected situation, the own vehicle 200 A correct evaluation of the car 200 can be difficult. In order to avoid such a situation, a road such as that shown in FIG. In other words, the own vehicle 200 can travel from a road different from the road on which the other vehicles (the other vehicle X (310) and the other vehicle Y (320)) existing around the own vehicle 200 exist. By merging with the road (driving lane) on which other vehicles (other vehicle X (310) and other vehicle Y (320)) exist, the other vehicle may be caused to travel outside the external sensing range of own vehicle 200. FIG. Also, by merging the own vehicle 200 and other vehicles from different roads, it is possible to reproduce a problem due to the speed difference between the own vehicle 200 and the other vehicles at the start of the simulation, or a scenario in which other vehicles are arranged in front of and behind the own vehicle 200. problem can be dealt with.

(Concrete example of update S160 of movement plan of surrounding traffic participants)
An example of the movement plan update method performed in the movement plan update S160 of the peripheral traffic participants will be described with reference to FIG. 6 in association with the speed plan (movement plan) shown in FIG.

　Fig. 6 shows the update of the movement plan at time Tc (565). At time Tc (565), the estimated scenario start time is advanced by time dT (568) from the initial scenario estimated start time Tssi (550), estimated from the running condition of the vehicle 200, and the updated scenario estimated start time Tssu (555) is reached. It shows the situation that has become (changed). Note that the change in the predicted time is determined at arbitrary timing during execution of the simulation. In such a case, initial speed plan 551 is generally advanced by time dT (568) and updated to updated speed plan 556 in response to the aforementioned change in predicted time.

Due to the update of the speed plan, the planned travel distance until the start of the scenario changes to the area indicated by the area 575, and a deviation occurs between the travel distance actually required before the start of the scenario, but the deviation is eliminated. By performing separate feedback control and giving an offset to the running speed, an area 570 is added as a compensating running distance. In other words, the result of updating the movement plan of the other vehicle is a plan that is shifted in time (in the direction of the time axis) from the formulated movement plan, and based on the result of updating the movement plan of the other vehicle, The speed plan is updated by making adjustments to compensate for the excess or deficiency of the distance to the position of the other vehicle, and the behavior of the other vehicle is readjusted. Note that the update scenario start predicted time Tssu (555) may be delayed, so the area 570 may work negatively (in the direction of decreasing the speed). Also, during the simulation, the speed plan is updated each time one step is executed, and the actual mileage required to start the scenario decreases as well. Relationships also change from time to time. Therefore, since the traveling distance for compensation (corresponding to the area 570) also changes sequentially, the feedback control amount also changes sequentially.

An example of a method for updating the predicted scenario start time will be explained using FIG. 7 while corresponding to FIG. FIG. 7 shows a relationship 710 between the time acquired in the own vehicle traveling timing acquisition S120 and the traveling distance of the own vehicle 200. As shown in FIG.

At time Tc (565) (FIG. 6), if the travel distance of own vehicle 200 is Lec (307), it is estimated from the relationship shown in FIG. 7 that own vehicle 200 is in a state corresponding to time Tcp (567). . In this case, the advance time dT (568) (FIG. 6) of the updated scenario predicted start time Tssu (555) (FIG. 6) with respect to the initial scenario predicted start time Tssi (550) is from time Tcp (567) to time Tc (565). is estimated to be the value after subtracting

Fig. 8 shows an example of how to generate acceleration commands for other vehicles. A movement plan is formulated using a speed plan, but in order to make the behavior of the other vehicle realistic, the behavior of the other vehicle is instructed using the acceleration. Separately, after limiting the speed according to the assumed performance of other vehicles, by limiting the acceleration range (speed change range per specified time), it is possible to suppress extreme speed changes, It can be made to behave realistically. This is the same for surrounding traffic participants (bicyclists, pedestrians, etc.) other than other vehicles, although the range of restrictions is different. It is conceivable to limit not only the angular velocity but also the angular acceleration of the directions of surrounding traffic participants, including other vehicles, as necessary to keep the behavior within a realistic range. Also, even for the same type of peripheral traffic participants, the range of restrictions may be changed as needed.

The scenario start time prediction unit 650 calculates the predicted update scenario start time Tssu (555) by the method shown using FIG. The time-shifted speed plan formulation unit 660 shifts the entire speed plan in the direction of the time axis in accordance with the update scenario start prediction time Tssu (555) as shown in FIG.

The planned acceleration calculation unit 680 calculates the acceleration at the current time with respect to the time-shifted speed plan. If the velocity plan is defined by a function, the acceleration can be calculated by analytically differentiating the velocity plan. If the speed plan is a list of speed values, it is necessary to calculate the differential value by numerical calculation based on the speed values around the current time. In this embodiment, since the velocity plan is handled in a form that can be expressed as a function, the acceleration can be calculated by analytical differentiation.

The planned acceleration calculator 680 calculates the acceleration based on the speed plan that matches the updated scenario start predicted time Tsu (555), and the speed plan matches the speed and acceleration that are the conditions at the start of the scenario. Therefore, basically, if the other vehicle is controlled so as to match the result of the planned acceleration calculation unit 680, the speed and acceleration of the other vehicle can satisfy the conditions at the start of the scenario.

The planned distance calculation unit 670 until the predicted scenario start time calculates the traveled distance from the current time to the updated scenario start predicted time Tsu (555) using the time-shifted speed plan. That is, the speed in the range from the current time to the update scenario start prediction time Tssu (555) is integrated. More specifically, the area corresponding to region 575 in FIG. 6 is obtained.

In order to obtain the traveled distance from the current time to the update scenario start predicted time Tssu (555), it is necessary to integrate the function indicating the speed plan. Numerical approximation may be performed.

The actual remaining distance calculation unit 610 until the start of the scenario calculates the traveling distance (remaining distance) from the current position of the other vehicle corresponding to the speed plan to the position where the other vehicle should be at the start of the scenario.

A subtractor 620 calculates the difference between the result of the actual remaining distance calculator 610 up to the start of the scenario and the result of the planned distance calculator 670 up to the predicted scenario start time, and the output of the subtractor 620 is corrected by the correction acceleration calculator 630. Calculate the acceleration for PID control or the like may be used to calculate the correction acceleration. However, since the input is the distance difference, it needs to be adjusted to output the value in the form of acceleration. Also, the input is the remaining distance to the target scenario start position, and since the pace of decrease in the remaining distance increases as the speed increases, it is necessary to pay attention to the sign.

The effect of the correction acceleration appears in the speed of the other vehicle corresponding to the speed plan in the form of area 570 in FIG. There is Note that the shape of the area 570 is merely an example, and varies depending on the method of mounting the correction acceleration calculator 630 and the parameters used for calculation.

The output of the planned acceleration calculator 680 and the output of the correction acceleration calculator 630 are added by the adder 640 to obtain the target acceleration, and the acceleration limiter 690 limits the acceleration range. By limiting the acceleration, that is, the range of speed change per specified time, it is possible to prevent the absolute value of the correction acceleration calculation unit 630 from becoming too large and the acceleration or deceleration from becoming unrealistic. Here, the specified time may be a time corresponding to one step of the simulation.

The output of the acceleration limiter 690 becomes the acceleration command 695, which becomes the acceleration request for the model of the other vehicle corresponding to the speed plan.

In order to avoid a collision, the other vehicle model may not follow the acceleration command 695 depending on the surrounding conditions of the other vehicle. Feedback is applied so that the position that should exist at the start of the scenario (scenario start position) is gradually reached.

Normally, as the scenario start time approaches, the prediction accuracy of the scenario start time increases, and the difference between the output of the actual remaining distance calculation unit 610 until the scenario start and the output of the planned distance calculation unit 670 until the predicted scenario start time becomes smaller. The change in predicted update scenario start time Tssu (555) is also reduced. Therefore, the correction amount to be included in the speed plan and the output of the correction acceleration calculator 630 are both small, and the difference between the speed and acceleration relative to the speed plan is also small.

[Simulation device]
FIG. 9 shows an example of a simulation apparatus configured using the evaluation method (simulation method) described above. FIG. 10 shows an example of the relationship between each process of the simulation execution procedure in FIG. 1 and each part of the simulation apparatus in FIG.

The simulation device 900 is configured as a computer including a processor such as a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), HDD (Hard Disk Drive) and other memories. Each function of the simulation device 900 is implemented by the processor executing a program stored in the ROM. The RAM stores data including intermediate data for calculations by programs executed by the processor.

The simulation apparatus 900 of this embodiment implements the simulation method for simulating the vehicle behavior described above. As shown in FIG. , a surrounding moving object behavior model unit 830, a vehicle behavior comparison unit 840, a scenario condition setting unit 910, a vehicle driving plan formulation unit 920, a driver model unit 930, a vehicle driving control model unit 940, a vehicle behavior model unit 950, It includes an environment reproduction unit 960 , a peripheral sensor model unit 970 , an environment data provision unit 980 , a behavior recording unit 990 and a result confirmation unit 995 . The peripheral moving object initial movement plan formulation unit 810, the peripheral moving object movement plan updating unit 820, the peripheral moving object behavior model unit 830, and the own vehicle behavior comparison unit 840 are related to the control of the peripheral moving object control unit 800. constitutes

(Scenario condition setting unit 910)
The scenario condition setting unit 910 takes in the scenario starting conditions as shown in FIG. 4 and the behavior instructions of the own vehicle 200 and each surrounding traffic participant after the start of the scenario from an external source such as a file into the simulator.

(Vehicle travel plan formulation unit 920, environmental data provision unit 980)
Based on the information taken in by the scenario condition setting unit 910, the own vehicle travel plan formulation unit 920 formulates a travel plan for the own vehicle 200 from the information about the own vehicle 200. Parked vehicles that do not move at all) are constructed by the environmental data providing unit 980 as environmental data including roads.

(Peripheral moving object initial movement plan formulation unit 810)
A surrounding moving object initial movement plan formulating unit 810 formulates a movement plan at the start of the simulation based on information about surrounding moving objects including other vehicles. However, in order for the surrounding moving object initial movement plan formulation unit 810 to formulate a movement plan, the estimated scenario start time provided by the behavior recording unit 990 (Predicted arrival time, which is the time acquired in own vehicle traveling timing acquisition S120 and stored in the behavior recording unit 990) is required. Since the operation is unnecessary in the simulation of the situation in which it is not performed, the operation is not performed when the simulation in the own vehicle travel timing acquisition S120 is performed.

(Driver model section 930)
Based on the travel plan formulated by the own vehicle travel plan formulation unit 920 , a command corresponding to the travel plan corresponding to the time is supplied to the driver model unit 930 and the own vehicle driving control model unit 940 . The contents that the driver of the own vehicle 200 should operate in terms of the travel plan are supplied to the own vehicle driving control model unit 940 via the driver model unit 930 . The reason why there is a route directly connecting the own vehicle driving plan formulation unit 920 to the own vehicle driving control model unit 940 is that the evaluation of the own vehicle driving control model unit 940 corresponds to information that cannot be input by the driver. For example, this path makes it possible to deal with cases where the driver cannot directly input values such as the target speed, and cases where it is desired to set values for evaluation even if they cannot be set from the driver.

(Own vehicle driving control model unit 940)
The host vehicle driving control model unit 940 is a model that controls driving support and automatic driving, and is a main evaluation target of the simulation device 900 . The vehicle driving control model unit 940 controls the vehicle 200 according to the judgment based on the surrounding conditions of the vehicle 200 obtained from the surrounding sensor model unit 970 and the instructions output from the vehicle driving plan formulation unit 920 and the driver model unit 930. Instructions necessary for driving, such as acceleration, braking, and steering, are issued to the own vehicle behavior model section 950 .

(Self-vehicle behavior model unit 950)
The own vehicle behavior model unit 950 calculates the behavior of the own vehicle 200 based on the instructions from the own vehicle driving control model unit 940 and the road information (such as road surface conditions and gradients) from the environment reproduction unit 960 .

(Environment reproduction unit 960)
The environment reproduction unit 960 obtains information on the road, objects around the road, and parked vehicles obtained from the environment data providing unit 980, behavior information of the own vehicle 200 obtained from the own vehicle behavior model unit 950, and information from the surrounding moving object behavior model unit 830. The virtual space in the simulator is reproduced based on the obtained information on the behavior of surrounding moving objects.

(Peripheral sensor model unit 970)
The information of the virtual space reproduced by the environment reproduction unit 960 is provided to the peripheral sensor model unit 970, and the peripheral sensor model unit 970 uses the peripheral sensors (camera, radar, lidar, Ultrasonic waves, GNSS, etc.) are simulated, and information corresponding to the outputs of these peripheral sensors is provided to the own vehicle driving control model unit 940 . The peripheral sensor model unit 970 can be a main evaluation target by the simulation device 900, although it may be limited to a part.

(Behavior recording unit 990)
Information about the virtual space reproduced by the environment reproducing unit 960 is also provided to the behavior recording unit 990, and the behavior recording unit 990 stores information necessary for evaluation, such as position information and behavior information of the own vehicle 200 and surrounding moving objects. Record in series. At this time, it is conceivable to record more detailed information when some event such as collision (contact) occurs.

(Result confirmation unit 995)
The information recorded by the behavior recording unit 990 is provided to the result confirmation unit 995, and is used for extraction of evaluation result NG cases, evaluation failure cases, and the like.

Note that the result confirmation unit 995 of this embodiment also has a function of confirming abnormal behavior, so this function is also used to determine abnormal termination. The behavior recording unit 990 is also necessary for end determination in order to confirm that all the information for the required period has been acquired.

In this embodiment, the behavior recording unit 990 performs the progress during the simulation, and the final confirmation result (whether the own vehicle has completed running normally, or whether the simulation is performed due to an abnormality such as contact with another vehicle or emergency braking) completed, etc.) is performed by the result confirmation unit 995 .

(Peripheral moving object initial movement plan formulation unit 810)
The peripheral moving object initial movement plan formulating unit 810 performs processing corresponding to the contents of the aforementioned peripheral traffic participant movement plan formulating S130 (see also FIGS. 4 and 5).

(Peripheral moving object movement plan update unit 820, own vehicle behavior comparison unit 840)
The output of the surrounding moving object initial movement planning unit 810 and the position information acquired by the behavior recording unit 990 from the own vehicle 200 traveling alone obtained from the own vehicle behavior comparing unit 840 are simulated including surrounding moving objects. Based on the comparison result of the position information of the own vehicle 200 in the situation, the peripheral moving object movement plan update unit 820 updates the movement plan according to the change in the predicted time (updating scenario start predicted time). Issue behavior instructions (instructions to adjust behavior) to objects. This update of the movement plan corresponds to the process of updating the movement plan of the surrounding traffic participant S160 described above, and has the contents described with reference to FIGS. 5, 6, 7 and 8.

(Peripheral moving object behavior model unit 830)
The peripheral moving object behavior model unit 830 follows the behavior instruction from the peripheral moving object movement plan updating unit 820, and also considers the situation of the virtual space obtained from the environment reproduction unit 960. , to move. The result of the movement is provided to the environment reproducing section 960 and used to update the situation of the virtual space.

It should be noted that the own vehicle driving control model unit 940 may be actual software or an electronic control device that controls driving support and automatic driving to be evaluated. The peripheral sensor model unit 970 may also be partial, but it is also conceivable to use actual software or an electronic control device that configures the peripheral sensor to be evaluated.

By using the method and apparatus shown in the present embodiment, when simulating the automatic driving function and the driving support function in accordance with the evaluation scenario, the situation leading to the evaluation scenario start state is automatically set from the initial conditions of the evaluation scenario. In addition, moving objects such as surrounding vehicles can be caused to move realistically. Therefore, it is possible to easily perform an evaluation based on an evaluation scenario after setting the internal state of the own vehicle 200 to be evaluated to a state that can actually occur before the start of the evaluation scenario.

[Installation example on actual vehicle]
The method and apparatus shown in this embodiment are installed in an actual vehicle or the like to prevent the positional relationship between the own vehicle 200 and surrounding traffic participants from becoming a high-risk situation in an actual traffic environment. Utilization is also considered. In other words, high-risk positions, speeds, and acceleration relationships are registered in advance in a system installed in the vehicle, and if such a situation can be predicted, the vehicle 200 can transmit the risk to surrounding traffic participants. It is conceivable that a change in the speed plan will be requested so that the speed is low, and that the surrounding traffic participants will judge whether the request is appropriate and then review the speed plan.

[Effect]
As described above, the simulation method of the present embodiment is a simulation method for simulating the behavior of the target vehicle (self-vehicle 200). The position and speed, as well as the position and speed of a moving object existing around the target vehicle, are specified, and the predicted time at which the target vehicle reaches the scenario start position, which is the specified position at the start of the scenario ( (Predicted scenario start time) is obtained, and the position and speed of the moving object are set to the specified values at the predicted time when the target vehicle reaches the scenario start position. ), restricting the speed change range (acceleration range) per specified time to adjust the movement of the moving object (updating the movement plan of the moving object).

Also, based on the predicted time and the specified position and speed of the moving object, a movement plan of the moving object is formulated, and the planned movement plan is used to include the target vehicle and the moving object. start the simulation, determine whether the predicted time has changed at any timing during the execution of the simulation, and if the predicted time has changed (the predicted time has changed), the predicted time The movement plan of the moving object is updated in accordance with the change of .

Further, the simulation apparatus of the present embodiment is a simulation apparatus that realizes a simulation method for simulating vehicle behavior. The position and speed of the target vehicle at the time point and the position and speed of a moving object existing around the target vehicle are specified, and the target vehicle is placed at the scenario start position, which is the specified position at the start time of the scenario. is predicted to arrive at the target vehicle (predicted scenario start time), and at the predicted time at which the target vehicle arrives at the scenario start position, the position and speed of the moving object are set to specified values (the moving object is the The movement of the moving object is adjusted by limiting the speed change range (acceleration range) per specified time (the moving plan of the moving object is updated).

Further, a computer (processor) formulates a movement plan for the moving object based on the predicted time and the specified position and speed of the moving object, and uses the formulated movement plan to generate the target vehicle. and starting a simulation including the moving object, determining whether the predicted time has changed at any timing during execution of the simulation, and if the predicted time has changed (the predicted time has changed) Then, the movement plan of the moving object is updated according to the change in the predicted time.

In other words, first, the relationship between the time and the position of the evaluation target vehicle (self-vehicle 200) in the period from the start of simulation execution, which is a time somewhat before the start of the scenario, to the start of the scenario, is determined by driving the evaluation target vehicle alone in the simulation. It is acquired as a reference relationship by a method such as Then, in accordance with the timing when the vehicle to be evaluated passes the scenario start position, the behavior of each traffic participant in the surrounding area is planned so that the movement of each traffic participant, which is composed of surrounding vehicles, etc., will be in the desired state at the start of the scenario. formulate When evaluating a scenario on a simulation, the behavior plan of each traffic participant in the vicinity is updated as appropriate based on the reference relationship between the time and position of the vehicle being evaluated and the actual position of the vehicle being evaluated, and the updated behavior is obtained. Feedback control is performed on the behavior of each traffic participant so that each traffic participant reaches the respective position where the scenario should be at the start of the scenario when the vehicle to be evaluated passes the scenario start position according to the plan. When feedback control is performed, the parameter change range (speed change range per specified time) is restricted so that the behavior of each traffic participant does not become unnatural.

According to this embodiment, when reproducing an evaluation scenario on a simulation, the positions and behaviors of surrounding traffic participants at the start of the scenario can be easily reproduced from the parameters at the start of the scenario. The behavior of each traffic participant can also be made natural.

In other words, a scenario for evaluating a vehicle that operates autonomously is reproduced in a simulation. ), at the start of the scenario, the state that satisfies the initial conditions of the scenario, such as position, velocity, and acceleration, can be easily achieved in the form of natural behavior that can occur in an actual traffic environment.

It should be noted that the present invention is not limited to the above-described embodiments, and includes various modifications. For example, the above-described embodiments have been described in detail in order to explain the present invention in an easy-to-understand manner, and are not necessarily limited to those having all the described configurations. In addition, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. Moreover, it is possible to add, delete, or replace a part of the configuration of each embodiment with another configuration.

In addition, each of the above configurations, functions, processing units, processing means, etc. may be realized in hardware, for example, by designing a part or all of them with an integrated circuit. Moreover, each of the above configurations, functions, etc. may be realized by software by a processor interpreting and executing a program for realizing each function. Information such as programs, tables, and files that implement each function can be stored in storage devices such as memory, hard disks, SSDs (Solid State Drives), or recording media such as IC cards, SD cards, and DVDs.

In addition, control lines and information lines indicate what is considered necessary for explanation, and not all control lines and information lines are necessarily indicated on the product. In practice, it may be considered that almost all configurations are interconnected.

S110 Create own vehicle travel definition, S120 Acquire own vehicle travel timing, S130 Formulate movement plan of surrounding traffic participants, S140 Initialize execution of simulation, S150 Execute one step of simulation, S160 Movement plan of surrounding traffic participants Update, S170...Simulation end determination process, S180...Simulation result recording, 200...Self-vehicle, 201...Scenario start position Le of own vehicle, 204...Self-vehicle speed Ve at scenario start, 206...Self-vehicle at scenario start Acceleration Ae 211 Position of own vehicle at simulation start Le0 301 Position of other vehicle at simulation start L0 (provided that it can be placed furthest from the scenario start position in terms of movement plan) 307 Present vehicle Vehicle position Lec, 310 Other vehicle X, 312 Relative distance of other vehicle X from own vehicle at scenario start Df1, 314 Velocity of other vehicle X at scenario start Vf1, 316 Other vehicle at scenario start Acceleration of X Af1, 320 Other vehicle Y, 322 Relative distance of other vehicle Y from own vehicle at scenario start Df2, 324 Velocity of other vehicle Y at scenario start Vf2, 326 Other vehicle at scenario start Acceleration Af2 of Y, 360 Parked vehicle 361 Parked vehicle position Lp 410 Main line 460 Merging road 470 Acceleration lane 510 Other vehicle speed plan 511 Other vehicle running start speed Vfp0 512... Speed Vfp1 when the other vehicle enters the acceleration change period, 513... Speed Vfp2 when the other vehicle enters the constant acceleration period, 515... Speed Vf when the scenario of the other vehicle starts, 516... When the scenario of the other vehicle starts Acceleration Af, 530... Time when the other vehicle starts running Tp0, 531... Time when the other vehicle enters the acceleration change period Tp1, 532... Time when the other vehicle enters the constant acceleration period Tp2, 550... Predicted initial scenario start time of the other vehicle Tssi, 551... Initial speed plan (movement plan) of other vehicle, 555... Predicted update scenario start time Tssu of other vehicle, 556... Updated speed plan (movement plan) of other vehicle, 565... Current simulation time Tc, 567 ... Time Tcp corresponding to the position obtained from the relationship between the own vehicle and the travel distance obtained in advance, 568 ... Advance time dT with respect to the initial scenario start prediction time, 570 ... Area representing the corrected travel distance, 575 ... Updated speed plan ( 610 ... Actual remaining distance calculation unit until scenario start time 620 ... Subtractor 630 ... Correction acceleration calculation unit 640 ... Adder , 650... Scenario start time prediction unit 660... Time-shifted speed plan formulation unit 670... Planned distance calculation unit to predicted scenario start time 680... Planned acceleration calculation unit 690... Acceleration limit unit 695... Acceleration command, 710 -- Pre-obtained relationship between own vehicle and traveling distance 800 -- Peripheral moving object control unit 810 -- Peripheral moving object initial movement plan formulating unit 820 -- Peripheral moving object movement plan updating unit 830 -- Peripheral moving object behavior model unit , 840 Own vehicle behavior comparison unit 900 Simulation device 910 Scenario condition setting unit 920 Own vehicle driving plan formulation unit 930 Driver model unit 940 Own vehicle driving control model unit 950 Own vehicle behavior Model part 960... Environment reproduction part 970... Peripheral sensor model part 980... Environment data provision part 990... Behavior recording part 995... Result confirmation part

Claims

A simulation method for simulating vehicle behavior, comprising:
Designating the position and speed of the target vehicle at the start of a scenario for verifying the behavior of the target vehicle, and the position and speed of moving objects existing around the target vehicle;
obtaining a predicted time at which the target vehicle will reach the scenario start position, which is the specified position at the start of the scenario;
Adjusting the movement of the moving object by restricting a speed change range per specified time so that the position and speed of the moving object become specified values at the predicted time when the target vehicle reaches the scenario start position. A simulation method characterized by:
In the simulation method according to claim 1,
formulating a movement plan of the moving object based on the predicted time and the specified position and speed of the moving object, and using the formulated movement plan, performing a simulation including the target vehicle and the moving object; and start
It is determined at any timing during execution of the simulation whether or not the predicted time has changed, and if the predicted time has changed, the movement plan of the moving object is updated according to the change in the predicted time. A simulation method characterized by:
In the simulation method according to claim 1,
In order to calculate the predicted time at which the target vehicle reaches the scenario start position, the target vehicle is run alone and a simulation is executed, and the time and position of the target vehicle obtained by executing the simulation. A simulation method, characterized by referring to the relationship of
In the simulation method according to claim 2,
A simulation method, wherein the movement plan of the moving object is determined based on speed, and the movement plan of the moving object is planned so that the speed changes continuously.
In the simulation method according to claim 2,
a result of updating the moving plan of the moving object is a plan obtained by shifting the determined moving plan in terms of time;
Adjusting the behavior of the moving object by adding an adjustment that compensates for a movement distance that is excessive or insufficient with respect to the distance to the position of the moving object at the start of the scenario based on the update result of the movement plan of the moving object. A simulation method characterized by:
In the simulation method according to claim 1,
The simulation method, wherein the target vehicle is made to join the road on which the moving object exists from a road different from the road on which the moving object exists around the target vehicle.
In the simulation method according to claim 1,
A simulation method, wherein the simulation is terminated when at least one of the target vehicle and moving objects existing around the target vehicle deviates from a predetermined speed range.
In the simulation method according to claim 1,
A simulation method, wherein the simulation is terminated when the target vehicle comes into contact with a moving object existing in the vicinity of the target vehicle.
In the simulation method according to claim 1,
A simulation method, wherein the simulation is terminated when a moving object existing around the target vehicle comes into contact with another moving object existing around the target vehicle.
In the simulation method according to claim 1,
When the target vehicle passes the scenario start position, which is the specified position at the start of the scenario, the states of the target vehicle and moving objects existing around the target vehicle are confirmed, and the scenario is started. A simulation method, comprising the step of terminating the simulation when the conditions at the point in time deviate from the conditions by a predetermined range or more.
A simulation device for realizing a simulation method for simulating vehicle behavior,
Designating the position and speed of the target vehicle at the start of a scenario for verifying the behavior of the target vehicle, and the position and speed of moving objects existing around the target vehicle;
obtaining a predicted time at which the target vehicle will reach the scenario start position, which is the specified position at the start of the scenario;
Adjusting the movement of the moving object by restricting a speed change range per specified time so that the position and speed of the moving object become specified values at the predicted time when the target vehicle reaches the scenario start position. A simulation device characterized by:
In the simulation device according to claim 11,
formulating a movement plan of the moving object based on the predicted time and the specified position and speed of the moving object, and using the formulated movement plan, performing a simulation including the target vehicle and the moving object; and start
It is determined at any timing during execution of the simulation whether or not the predicted time has changed, and if the predicted time has changed, the movement plan of the moving object is updated according to the change in the predicted time. A simulation device characterized by: