Simulation system, simulation program, and simulation method

Abstract

To provide a simulation system that can reproduce vehicle deceleration more realistically. [Solution] This simulation system comprises a reaction time processing unit that calculates the reaction distance based on the reaction time of a human or machine, including the time it takes to perceive and recognize the occurrence of a dangerous event and the time from that recognition to the start of emergency braking of the vehicle, and the speed immediately before the start of emergency braking; and a braking processing unit that calculates the braking distance from the start of emergency braking to the stopping of the vehicle. The stopping distance until the moving vehicle comes to a stop due to emergency braking is determined by adding the reaction distance calculated by the reaction time processing unit and the braking distance calculated by the braking processing unit.

Description

[Technical Field]

[0001] The present invention relates to a simulation system, a simulation program, and a simulation method. [Background technology]

[0002] It is not possible to conduct experimental research on traffic accidents because it would endanger subjects. Therefore, there is a need for the development of simulators that can accurately reproduce actual human behavior and vehicle behavior. The inventors have developed simulation software suitable for traffic flow called "Land Scape Simulator (LSS)" (see Patent Documents 1 and 2). This technology is completely different from conventional simulators that treat traffic flow as an overall flow, and is a groundbreaking simulator that employs a "multi-agent model" in which each vehicle in the field behaves with individuality. Each vehicle proceeds while independently making judgments based on its respective settings, such as forward visibility, surrounding sounds, road conditions, and traffic signal conditions. With a unique algorithm that runs multiple time axes in parallel, it has the high capability to simultaneously deploy and process an overwhelming number of vehicles. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] International Publication Number WO2021 / 054267 [Patent Document 2] International Publication Number WO2021 / 039845 [Overview of the project] [Problems that the invention aims to solve]

[0004] The inventors are conducting research to reproduce rear-end collisions using LSS (Long Stake System). This research is expected to contain a wealth of information that will be useful for the development of autonomous driving and advanced driver-assistance systems (ADAS), as it should encompass both human-side and vehicle-side issues.

[0005] Even if LSS is implemented as is, rear-end collisions will not occur. This is because the simulator driver fully fulfills their duty to pay attention to the road ahead. Therefore, to replicate a real human driver, they are intentionally made to "look away" from the road. This is done by stopping the signal input for forward visibility for a certain period of time. It is being shown that this can reproduce the same situation as "looking away" from the road, and thus reproduce rear-end collisions. However, it has also become clear that the number of accidents reproduced in the simulator is far greater than the actual number of accidents.

[0006] Vehicle deceleration is classified into "normal deceleration" and "rapid deceleration." Since LSS was not originally developed as a traffic accident simulator, it only implements an algorithm for "normal deceleration," which takes into account the distance to the vehicle in front and the deceleration of the vehicle behind it.

[0007] The object of the present invention is to provide a simulation system, a simulation program, and a simulation method that can reproduce vehicle deceleration more realistically. [Means for solving the problem]

[0008] The simulation system according to the present invention comprises a reaction time processing unit that calculates the reaction distance using the reaction time of a human or machine, which includes the time it takes to perceive and recognize the occurrence of a dangerous event and the time from that recognition to the start of emergency braking of the vehicle, and the speed immediately before the start of emergency braking, and a braking processing unit that calculates the braking distance from the start of emergency braking to the stopping of the vehicle, and the stopping distance until the moving vehicle stops due to emergency braking is determined by adding the reaction distance calculated by the reaction time processing unit and the braking distance calculated by the braking processing unit.

[0009] Vehicle deceleration, such as that of automobiles, is classified into "normal deceleration" and "rapid deceleration." Rapid deceleration as an accident avoidance action is assumed to be determined by the sum of the reaction time of humans or machines (sensors or computers) (perception → recognition → action) and the braking time of the vehicle. In this invention, an algorithm for rapid deceleration due to emergency braking of a vehicle is introduced, and this rapid deceleration algorithm is based on the sum of the reaction time of humans or machines and the braking time of the vehicle, taking into account arbitrary human or machine characteristics and arbitrary performance characteristics. Therefore, by introducing such a rapid deceleration algorithm in addition to the normal deceleration algorithm, vehicle deceleration can be reproduced more realistically.

[0010] The simulation system according to the present invention includes a state transition unit that transitions the vehicle from a driving state until it is stopped by emergency braking, to a coasting state from the perception of the occurrence of the dangerous event until emergency braking is initiated, to a braking state from the start of emergency braking until the vehicle stops, and to a stopped state. Furthermore, the state transition unit transitions to the coasting state when the stopping distance becomes greater than the distance between vehicles, and transitions to the driving state or the stopped state when the stopping distance becomes less than or equal to the distance between vehicles or when the vehicle stops.

[0011] In this invention, by incorporating this rapid deceleration algorithm into a "multi-agent model," it becomes possible to simulate a highly realistic traffic environment in which drivers with diverse personalities drive vehicles with varying performance characteristics. The results of these simulations will greatly contribute to the application of autonomous driving algorithms and the development of advanced driver-assistance systems (ADAS), such as automatic braking.

[0012] In the simulation system according to the present invention, the first step time, which is the time interval for processing to calculate the braking distance, is shorter than the second step time for the entire simulation.

[0013] To “accurately determine” the occurrence of an accident, extremely high computational accuracy is required. In the present invention, by making the first time interval, which is the time interval of the process for calculating the braking distance, shorter than the second time interval of the entire simulation, it is possible to precisely calculate the process related to the braking time of the vehicle during “rapid deceleration” without increasing the calculation time of the entire simulation.

Advantages of the Invention

[0014] According to the present invention, the deceleration of a vehicle can be reproduced more realistically.

Brief Description of the Drawings

[0015] [Figure 1] FIG. 1 is a block diagram showing the configuration of a simulation system according to an embodiment of the present invention. [Figure 2] FIG. 2 is an explanatory diagram of “rapid deceleration” of an automobile according to an embodiment of the present invention. [Figure 3] FIG. 3 is a diagram showing an example of the state transition of an automobile using “rapid deceleration” according to an embodiment of the present invention and an example of a normally traveling automobile. [Figure 4] FIG. 4 is a state transition diagram in the operation of “rapid deceleration” according to an embodiment of the present invention.

Modes for Carrying Out the Invention

[0016] Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[0017] FIG. 1 is a block diagram showing the configuration of a simulation system according to an embodiment of the present invention.

[0018] The simulation system 1 shown in FIG. 1 is obtained by applying the present invention to the aforementioned “Land Scape Simulator (LSS)” (see Patent Documents 1 and 2).

[0019] This simulation system 1 has a normal deceleration processing unit 10.

[0020] LSS uses the Optimal Velocity Model (OV model) to calculate vehicle acceleration.

[0021] As shown in equation (1), the acceleration of a vehicle is a superposition of an "acceleration component" (positive component) and a "deceleration component" (negative component).

[0022] a=acceleration+deceleration (1) Here, a: acceleration of the vehicle, acceleration: acceleration component, deceleration: deceleration component.

[0023] In the normal deceleration processing unit 10, the "deceleration component" simulates "normal deceleration," and the vehicle speed (U) and the distance between vehicles (D) are used. inter ), and using information (f1, f2, f3) from the driver / vehicle's distance sensors, the calculation is performed as shown in equation (2).

[0024] deceleration=β(U+(f1+f2+f3)(1+k3 / (k3+D inter )) (2) Here, deceleration: deceleration component, β: braking coefficient (the larger the value, the stronger the braking), U: vehicle speed, f1: change in distance between vehicles (as the distance between vehicles decreases, the driver's psychology tends to slow down), f2: difference between the acceleration of the vehicle in front and your own acceleration (as the vehicle in front begins to slow down, the driver's psychology tends to slow down), f3: difference between the speed of the vehicle in front and your own speed (vehicles with high ground clearance may appear to approach rapidly (illusion, etc.)), k3: speed difference coefficient (the larger the value, the stronger the braking against the slower vehicle in front), D inter This refers to the distance between vehicles.

[0025] This simulation system 1 further includes a rapid deceleration processing unit 20.

[0026] This rapid deceleration processing unit 20 includes a coasting processing unit 21 and a braking processing unit 22.

[0027] The reaction time processing unit 21 calculates the reaction distance using the reaction time of the human or machine, which includes the time it takes to perceive and recognize the occurrence of a dangerous event and the time from that recognition to the start of the emergency braking of the vehicle, and the speed immediately before the start of the emergency braking.

[0028] The braking processing unit 22 calculates the braking distance from the start of emergency braking until the vehicle comes to a complete stop.

[0029] The rapid deceleration processing unit 20 determines the stopping distance required for a moving vehicle to come to a complete stop by applying emergency brakes by adding the reaction distance calculated by the reaction distance processing unit 21 and the braking distance calculated by the braking processing unit 22.

[0030] "Sudden deceleration" as an accident avoidance action involves applying maximum braking force, largely unaffected by external inputs. Furthermore, individual differences in human behavior and differences in vehicle performance come into play in a complex manner. Braking time is assumed to be determined by the sum of the human reaction time (perception → recognition → action) and the vehicle's braking time, and the human reaction time is considered to be t reaction So, the braking time of a car is t brake This is expressed by taking into account arbitrary human characteristics and arbitrary performance characteristics.

[0031] The "sudden deceleration" of a car will be explained based on Figure 2. Assuming that the tires 101 of car 100 lock up during "sudden deceleration," the work done by the force of "sudden deceleration" and the frictional force can be calculated as shown in equations (3, 4).

[0032] F = μmg (3) W=FS=μSmg (4) Here, F is the force of "rapid deceleration," μ is the coefficient of friction, m is the mass of the car (100), g is the acceleration due to gravity, W is the work done by the frictional force, and S is the braking distance.

[0033] Furthermore, the kinetic energy of car 100 when it begins to "suddenly decelerate" is shown in equation (5).

[0034] K=(mU0 2 ) / twenty five) Here, K is the kinetic energy when the vehicle 100 starts "rapid deceleration", m is the mass of the vehicle 100, and U0 is the initial speed of the vehicle 100 during "rapid deceleration".

[0035] From the law of conservation of energy, the work done by the frictional force is the same as the kinetic energy when the vehicle 100 starts "rapid deceleration", so the braking distance is obtained as shown in Equation (6).

[0036] S = U0 2 / 2μg (6) Here, S is the braking distance, μ is the coefficient of friction, g is the acceleration due to gravity, and U0 is the initial speed of the vehicle 100 during "rapid deceleration".

[0037] Implement two deceleration algorithms, the "normal deceleration" algorithm (normal deceleration processing unit 10) and the "rapid deceleration" algorithm (rapid deceleration processing unit 20), in the LSS, and usually set the former (normal deceleration processing unit 10) to operate under specific conditions and the latter (rapid deceleration processing unit 20) to operate. This condition setting is also an important component of this simulation system 1.

[0038] The driver of each vehicle 100 will perform "rapid deceleration" when it is determined that "a dangerous event has occurred". The judgment formulas are shown in Equations (7a, 7b, 8). The criterion for activating rapid deceleration is defined by the "distance between the vehicle and the vehicle ahead".

[0039] If D inter <α(U0 2 / 2μg) Brake ON (7a) If D inter ≧α(U0 2 / 2μg) Brake OFF (7b) f(U0)=α(U0 2 / 2μg) (8) Here, D inter : the distance between the vehicles, α: the coefficient for reproducing the behavior of the driver during braking, μ: the coefficient of friction, g: the acceleration due to gravity, and U0: the initial speed of the vehicle 100 during "rapid deceleration".

[0040] (U0 in Equation (8)2 The term ( / 2μg) is the theoretical value of the shortest braking distance at which a vehicle 100 traveling at an initial speed U0 can stop when the brakes are applied with maximum force (Equation (6)). This value is determined solely by the friction coefficient between the road surface and the tires and the vehicle speed, and does not depend on the vehicle weight or performance. It is known that humans have an intuitive line at which they feel they can stop, and they do not approach the vehicle in front of them beyond that line. This distance is set as the threshold for activating "rapid deceleration". In other words, if the approach exceeds the threshold, the driver judges it to be dangerous and activates "rapid deceleration". As described later, "rapid deceleration" is a brake that is activated with maximum force. A coefficient α is introduced here to reproduce the driver's personality and behavioral characteristics. If α > 1, the driver is conservative or cautious and does not get too close to the vehicle in front. If α < 1, the driver becomes aggressive or lacks caution and does not hesitate to approach the vehicle in front.

[0041] This simulation system 1 further includes a state transition unit 30. The state transition unit 30 transitions to a coasting state when the corrected braking distance, which is obtained by correcting the shortest braking distance that a moving vehicle 100 can stop when the brakes are applied with the maximum pedal force applied as described above, using a coefficient to reproduce the driver's actions during braking, becomes greater than the distance to the vehicle in front. It then transitions to a driving state or a stopped state when the corrected braking distance becomes less than or equal to the distance to the vehicle in front, or when the vehicle 100 comes to a stop.

[0042] Here, the driver independently decides whether to "continue" or "deceleration" based on the distance to the vehicle ahead. If "deceleration" is chosen, the driver decides whether to "deceleration normally" or "deceleration rapidly". The state transition unit 30 transitions the state of the vehicle 100 that is using "deceleration rapidly".

[0043] The upper part of Figure 3 shows an example of the state transitions of car 100 using "rapid deceleration". Car 100, in the "driving" state, detects a hazard ahead. By "coasting" during this time, the human reaction time is reproduced. This includes the time from perception (seeing the vehicle ahead), recognition (judging that a collision is imminent) to action (pressing the brake pedal). Setting this time longer will reproduce the behavior of elderly drivers or drivers who take time to make decisions. The brakes are pressed and "braking" begins. When braking with maximum force, theoretically the braking force is determined only by the friction coefficient between the road surface and the tires. By introducing a coefficient here as well, road conditions (weather such as rain and snow, and pavement conditions) and the driver's braking force are reproduced. Changing this coefficient will reproduce rainy weather or female drivers or elderly drivers with weaker braking force. After this, car 100 "stops" and then restarts "driving".

[0044] LSS employs an "SCS (Complex System Simulation Platform)" that simultaneously calculates parallel events with different time step sizes. The high capability of LSS to simultaneously process the movements of a large number of cars 100 at high speed is supported by this SCS. Simulating "sudden deceleration" involves calculating an event with a large time change, requiring a fine time step size. On the other hand, as shown in the lower part of Figure 3, the other cars 100 that are driving normally are events with small time changes. If the process of making the time step size of the entire simulation finer every time "sudden deceleration" occurs in even one of the many cars 100 would incur enormous computational costs and become unmanageable. In the simulation system 1 according to this embodiment, as shown in Figure 3, the time step size of the entire simulation (Power_Δt) and the time step size of "sudden deceleration" (Brake_Δt) are set separately, and a method is adopted that precisely calculates "sudden deceleration" without increasing the computation time of the entire simulation. Extremely high computational accuracy is required to "accurately determine" the occurrence of an accident. LSS solves this problem by simultaneously computing parallel events with different time step sizes.

[0045] The rapid deceleration algorithm (state transition section 30) is summarized in Figure 4. The "rapid deceleration" operation is represented by four state transitions. The initial speed U0 of the car 100 during "rapid deceleration" is a value calculated by LSS and is not pre-set, so the ON / OFF state of "rapid deceleration" is represented by the speed U of the car 100 at each instant (Equation (9)).

[0046] f(U) = α(U) 2 (2μg) (9) Here, α is a coefficient for reproducing the driver's actions during braking, μ is the coefficient of friction, g is the acceleration due to gravity, and U is the speed of car 100.

[0047] As shown in equation (10), in the case of "rapid deceleration," the stopping distance is the sum of the reaction distance and the braking distance.

[0048] S stop =S reaction +S brake (10) Here, S stop :Stopping distance, S reaction :Empty running distance, S brake This is the braking distance.

[0049] As shown in Figure 4, car 100 is traveling at speed U (driving). If the distance between vehicles is less than f(U), the brakes are applied, but braking is delayed and the speed U of car 100 remains constant (coasting). At this time, the speed of car 100 is the initial speed U0 of car 100 during rapid deceleration. Subsequently, the brakes are applied and the speed decreases (braking). In the following state, car 100 comes to a stop (stopped) and restarts.

[0050] In the numerical simulation in Simulation System 1, the phenomenon is calculated discretely for each step time Δt. A new recurrence relation was constructed to introduce "rapid deceleration" in LSS (Equations (11) to (17)). When "rapid deceleration" is activated, if the initial speed of the vehicle is U0, the recurrence relations for "coasting" and "braking" in Figure 4 are as follows. Dry running If t≦t0U=U0(11) an =a n-1 =0 (12) U n =U n-1 =U0(13) S n =U n Δt = U0Δt (14) braking a n = -μg (15) U n =U n-1 -μgΔt (16) S n =U n Δt=(U n-1 -μgΔt)Δt (17) Here, t is the time from when danger is perceived until the vehicle stops, t0 is the time until emergency braking begins, U is the speed of car 100, U0 is the initial speed of car 100 during rapid deceleration, a is the acceleration of car 100, n is the calculation step (from n=1 to n), S is the distance, Δt is the step time, μ is the coefficient of friction, and g is the acceleration due to gravity.

[0051] In the simulation system 1 according to this embodiment, it is assumed that "sudden deceleration" as an accident avoidance action is determined by the sum of the human reaction time (perception → recognition → action) and the braking time of the automobile 100, and arbitrary human characteristics and arbitrary automobile characteristics are reproduced. Furthermore, by incorporating this algorithm into a multi-agent model, it becomes possible to simulate a traffic environment that is extremely close to reality, in which drivers with various personalities drive automobiles 100 with various performance characteristics.

[0052] Furthermore, "accurately determining" the occurrence of an accident requires extremely high computational precision. LSS solves this problem by simultaneously calculating parallel events with different time step sizes.

[0053] The present invention is not limited to the embodiments described above, and can be implemented in various modified forms, with the scope of such implementation also falling within the technical scope of the present invention.

[0054] In the embodiments described above, the premise was that the driver of the automobile was a human being, but the present invention is not limited thereto and can also be applied to autonomous driving and the like.

[0055] Furthermore, although the above embodiments were described using an automobile as an example of a vehicle, the present invention is not limited thereto and can of course be applied to other vehicles besides automobiles.

[0056] The rapid deceleration algorithm according to the present invention is expected to have applications in software (traffic flow, traffic accident analysis, safety measures) and ADAS fields. For example, it can be used to perform traffic accident simulations in ADAS development. This helps to determine the time required for ADAS responses to prevent traffic accidents. [Explanation of Symbols]

[0057] 1. Simulation System 10 Normal deceleration processing unit 20 Rapid deceleration processing unit 21. Free-running section 22 Braking Processing Unit 30 State transition section 100 automobiles

Claims

1. A reaction time processing unit calculates the reaction distance using the reaction time of a human or machine, which includes the time from the time it takes to perceive and recognize the occurrence of a dangerous event and the time from that recognition to the start of the vehicle's emergency braking, and the speed immediately before the start of the emergency braking. A braking processing unit that calculates the braking distance from the start of the emergency braking to the stopping of the vehicle. It is equipped with, The stopping distance required for the vehicle to come to a complete stop due to emergency braking is determined by adding the reaction distance calculated by the reaction distance processing unit and the braking distance calculated by the braking distance processing unit. Simulation system.

2. A simulation system according to claim 1, A state transition unit that transitions the state of the vehicle while it is in motion until it comes to a stop due to emergency braking, from the state of motion, through the state of coasting from the perception of the occurrence of the dangerous event to the start of emergency braking, through the state of braking from the start of emergency braking until the vehicle comes to a stop, and through the state of stopping. A simulation system equipped with the following features.

3. A simulation system according to claim 2, The state transition unit transitions to the coasting state when the corrected braking distance, which is obtained by correcting the braking distance that the moving vehicle can stop when the brakes are applied with maximum force using a coefficient to reproduce the driver's actions or the operation of the machine during braking, becomes greater than the distance to the vehicle in front. It also transitions to the driving state or the stopped state when the corrected braking distance becomes less than or equal to the distance to the vehicle in front, or when the vehicle comes to a stop. Simulation system.

4. A simulation system according to any one of claims 1 to 3, When the braking distance is denoted as S, S=U 0 2 / 2μg Here, μ: coefficient of friction, g: acceleration due to gravity, U 0 : Initial speed of a vehicle that decelerates rapidly Calculated by Simulation system.

5. A simulation system according to any one of claims 1 to 4, The first step time, which is the time interval for the process of calculating the braking distance, is shorter than the second step time of the entire simulation. Simulation system.

6. A simulation system according to claim 5, The braking distance S is calculated using the following recurrence relation. S ｎ =U ｎ Δt=(U ｎ－１ -μgΔt)Δt Here, U ０ : Initial speed of the vehicle undergoing rapid deceleration, n: Calculation step (from n=1 to n), Δt: First step time, μ: Coefficient of friction, g: Gravitational acceleration Simulation system.

7. A step of calculating the reaction distance using the reaction time of a human or machine, which includes the time from the time it takes to perceive and recognize the occurrence of a dangerous event and the time from that recognition to the start of emergency braking of the vehicle, and the speed immediately before the start of emergency braking. A step of calculating the braking distance from the start of the emergency braking to the stopping of the vehicle, The step of determining the stopping distance until the vehicle in motion comes to a complete stop due to emergency braking is to add the calculated reaction distance and the calculated braking distance. A simulation program that causes a computer to execute a simulation.

8. The reaction time of a human or machine, including the time it takes to perceive and recognize the occurrence of a dangerous event and the time from that recognition to the initiation of the vehicle's emergency braking, and the speed immediately before the initiation of the emergency braking are used to calculate the reaction distance. The braking distance from the start of the emergency braking to the stopping of the vehicle is calculated. The stopping distance required for the vehicle to come to a complete stop due to emergency braking is determined by adding the calculated reaction distance and the calculated braking distance. Simulation method.

Images