Travel evaluation device for mobile body or mobile body model, and travel evaluation program
The progress evaluation device and program address the inefficiency of repeated simulations by calculating and reflecting operator behavior in evaluation patterns, allowing for accurate and efficient human-like operation simulation of moving objects or models.
Patent Information
- Application Number
- PCT/JP2024/016429
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-30
- Filing Date
- 2024-04-26
- Publication Date
- 2024-12-05
AI Technical Summary
Existing methods for evaluating the progress of moving objects or models require repeated simulations to achieve human-like operations, leading to high processing loads and inefficient evaluation processes.
A progress evaluation device and program that calculates feature amounts reflecting operator behavior by comparing target and actual operation patterns, allowing for the adjustment of evaluation patterns to reflect human-like operations, thereby advancing the moving object or model based on these characteristics without repeated simulations.
Enables efficient evaluation of moving objects or models by simulating human-like operations, reducing processing load and improving evaluation accuracy without the need for multiple simulations.
Smart Images

Figure JP2024016429_05122024_PF_FP_ABST
Abstract
Description
Progress evaluation device and progress evaluation program for a mobile body or a mobile body model CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is based on Japanese Application No. 2023-088805, filed on May 30, 2023, the contents of which are incorporated herein by reference.
[0002] The present invention relates to a progress evaluation device and a progress evaluation program that use a computer or the like to advance a moving body or a moving body model and perform progress evaluation of the moving body or a moving body model.
[0003] A technology is known in which an operator model of a moving object is constructed, and the moving object or the moving object model is advanced using the operator model by a computer or the like, and the progress is evaluated. For example, in the manufacture and sale of vehicles such as automobiles, it is necessary to measure and display fuel economy and exhaust gas emissions when the vehicle is driven in a specific driving pattern (mode) defined by a country or region. The driving mode is represented, for example, by a graph showing the relationship between the time elapsed since the start of driving and the vehicle speed that should be reached at that time. This vehicle speed that should be reached is sometimes called a target vehicle speed.
[0004] When the moving body is a vehicle and the operator is a driver, the operator model is called a driver model. The fuel economy test described above is performed by running the vehicle or vehicle model using the driver model. The driver model, for example, simulates the operation of the accelerator pedal and brake pedal by the driver, and controls the vehicle speed of the vehicle or vehicle model.
[0005] In order to properly conduct tests related to fuel economy, etc., it is preferable to construct a driver model that performs human-like operations. For example, if the operations performed by the driver model deviate from human-like operations, such as operating the accelerator pedal or brake pedal too suddenly for a human to perform, it becomes difficult to properly evaluate the fuel economy performance of the vehicle.
[0006] Patent Literature 1 describes a driving simulation device that uses reinforcement learning to build a driver model that performs human-like pedal operation. Specifically, the driving simulation device automatically sets the gains of the driver model by running a vehicle model multiple times while changing the gain values of the driver model and evaluating the changed gain values based on a reward value. The gain values are evaluated not only by a vehicle speed reward function that evaluates vehicle speed tracking ability, but also by an accelerator reward function that evaluates the smoothness of accelerator pedal operation and a brake reward function that evaluates the smoothness of brake pedal operation.
[0007] JP 2014-115168 A
[0008] In the driving simulation device described in Patent Document 1, in order to adopt a highly rated gain value as the gain value of the driver model, it is necessary to repeatedly perform a driving simulation of the vehicle model while changing the gain value of the driver model, which poses a problem of a large processing load.
[0009] In view of the above, an object of the present disclosure is to provide a technology for evaluating the progress of a moving object or the like by performing human-like operations using a computer or the like, without repeatedly running simulations.
[0010] The present disclosure provides a progress evaluation device that progresses a moving object or a moving object model that models the moving object while changing a predetermined progress operation parameter based on a predetermined evaluation pattern, and performs a progress evaluation of the moving object or the moving object model. The progress evaluation device includes: a feature calculation unit that calculates a feature amount that indicates a characteristic of the operation behavior of the operator based on a difference between a target pattern that indicates a change in a target value of the progress operation parameter and an actual operation pattern that indicates a change in an actual value of the progress operation parameter when the operator actually operates the moving object in accordance with the target pattern; a pattern change unit that changes the evaluation pattern based on the feature amount calculated by the feature calculation unit, and changes the evaluation pattern to a feature reflection pattern that reflects the characteristic of the operation behavior of the operator; and a progress control unit that progresses the moving object or the moving object model based on the feature reflection pattern when performing the progress evaluation.
[0011] According to the above-described progress evaluation device, the feature calculation unit calculates a feature reflecting the characteristics of the operator's operation behavior. The feature is calculated based on the difference between a target pattern indicating changes in the target value of the progress operation parameter and an actual operation pattern indicating changes in the actual measured value of the progress operation parameter when the operator actually operates the moving object according to the target pattern. The pattern change unit then changes the evaluation pattern to a feature-reflecting pattern based on the feature. By changing the evaluation pattern based on the feature, a feature-reflecting pattern reflecting the characteristics of the operator's operation behavior can be obtained. The progress control unit causes the moving object or moving object model to progress based on the feature-reflecting pattern. By controlling the progress of the moving object or moving object model based on the feature-reflecting pattern (e.g., as a command value) rather than the evaluation pattern, the moving object or moving object model can progress as if it were actually operated by the operator. As a result, the progress of the moving object or moving object model can be evaluated by performing human-like operations using a computer or the like, without repeatedly running simulations.
[0012] The present disclosure can also be provided as a progress evaluation program for progressing the moving body or a moving body model modeled on the moving body while changing a predetermined progress operation parameter based on a predetermined evaluation pattern, and performing a progress evaluation of the moving body or the moving body model. This program causes a computer to execute the following steps: a feature calculation step of calculating a feature amount indicating a characteristic of the operation behavior of the operator based on a difference between a target pattern indicating a change in a target value of the progress operation parameter and an actual operation pattern indicating a change in an actual value of the progress operation parameter when the operator actually operates the moving body in accordance with the target pattern; a pattern change step of changing the evaluation pattern based on the feature amount calculated in the feature calculation step, and changing the evaluation pattern to a feature reflecting pattern that reflects the characteristic of the operation behavior of the operator; and a progress control step of progressing the moving body or the moving body model based on the feature reflecting pattern when performing the progress evaluation.
[0013] The above and other objects, features, and advantages of the present disclosure will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, in which Fig. 1 is a diagram illustrating a progress evaluation device according to an embodiment, Fig. 2 is a block diagram illustrating a feature amount calculation unit and an operator model construction unit according to an embodiment, Fig. 3 is a flowchart of progress evaluation processing according to an embodiment, Fig. 4 is a diagram illustrating calculation of feature amounts according to an embodiment, Fig. 5 is a diagram illustrating a transfer function as an example of a pattern modification unit according to an embodiment, Fig. 6 is a diagram illustrating a neural network as an example of a pattern modification unit according to an embodiment, Fig. 7 is a diagram illustrating an evaluation pattern according to an embodiment, and Fig. 8 is a diagram illustrating an evaluation pattern and a feature reflecting pattern according to an embodiment.
[0014] (Embodiment) As shown in FIG. 1 , a progress evaluation device 1 according to a first embodiment includes an interface 2, a communication unit 3, a display unit 4, a progress evaluation unit 10, and a storage unit 30. The progress evaluation device 1 calculates the behavior of a moving body or a moving body model based on an operator model, thereby evaluating its progress. Here, progress evaluation refers to evaluating the progress of a moving body or a moving body model, and includes a real-machine test in which the moving body is actually moved and evaluated, and a progress simulation in which the moving body model is run on a computer and evaluated. The progress evaluation device 1 can be applied to general progress evaluation systems for moving bodies and moving body models, and is particularly suitable for use in driving simulations in model-based development (MBD), in which design and development are carried out while verifying the system by modeling and simulating the controls and controlled objects that constitute the system, and for real-machine tests for automatically steering a vehicle. Examples of driving simulations in model-based development (MBD) include MILS (Model In the Loop Simulation), SILS (Software In the Loop Simulation), and HILS (Hardware In the Loop Simulation). In this embodiment, a case will be described in which a driving simulation of a vehicle model is executed when the moving body is a vehicle and the operator is the driver of the moving body.
[0015] The interface 2 accepts user operations and inputs various information to the progress evaluation unit 10. For example, the interface 2 accepts an instruction to execute a simulation process from the user and inputs the accepted instruction to the progress evaluation unit 10. An example of the interface 2 is an operation acceptance device including a keyboard, a mouse, etc.
[0016] The communication unit 3 is connected to a predetermined network (not shown) and transmits and receives information to and from an external information processing device (not shown). Examples of the communication unit 3 include a wired LAN and a wireless LAN. In the present embodiment, a case where a progress evaluation of a mobile body model is performed is described as an example. However, in the case of a real-machine test in which a mobile body (e.g., an actual vehicle) is operated, a progress evaluation device may be provided outside the mobile body, and the communication unit 3 may be configured to allow the mobile body to progress via communication.
[0017] The display unit 4 is, for example, a liquid crystal display, and displays various information to the user using characters, figures, etc. The display unit 4 displays the results of the driving simulation.
[0018] The storage unit 30 is, for example, a semiconductor memory element such as a flash memory, or a storage device such as a hard disk, an optical disk, etc. Note that the storage unit 30 is not limited to the above types of storage devices, and may be a RAM, a ROM, etc.
[0019] The storage unit 30 stores an operator model 31, a mobile body model 32, and feature quantities 33. In this embodiment, a driver model is stored as the operator model 31, and a vehicle model is stored as the mobile body model 32. The vehicle model is a model for simulating the internal and external states of the vehicle. The internal states include the engine state, powertrain state, and tire state. The external states include the vehicle's position, traveling direction, speed, and vehicle type. The driver model in this embodiment is a model for simulating the driver's operation of the accelerator pedal and brake pedal and controlling the vehicle speed of the vehicle or vehicle model. For example, a fuel economy test is performed by having the driver model drive the vehicle or vehicle model according to an evaluation pattern, which is a specific driving pattern defined for the test. A driving simulation for a fuel economy test is an example of a progress evaluation, and the form of the operator model can be changed as appropriate depending on the content of the progress evaluation. For example, the driver model may be a model for calculating the driver state and driving operation state. In this case, the driver state includes the planned driving route and the cognition and judgment state. The driving operation state includes the state of the steering wheel, the state of the accelerator, and the state of the brake. The feature amount 33 indicates the characteristics of the operation behavior of the operator, and is calculated by the feature amount calculation unit 11, which will be described later.
[0020] As shown in FIGS. 1 and 2 , the progress evaluation unit 10 includes a feature calculation unit 11 and an operator model construction unit 20. The progress evaluation unit 10 is primarily configured using a well-known microcomputer (MCU) including a CPU, ROM, RAM, flash memory, and the like. For example, the CPU executes a progress evaluation program installed in the storage unit 30 or ROM to realize the functions of the feature calculation unit 11, the operator model construction unit 20, and the like. The functions provided by the MCU may be provided by software stored in a physical memory device and a computer that executes the software, software alone, hardware alone, or a combination thereof. For example, when the MCU is implemented by electronic circuits, which are hardware, the functions may be provided by digital circuits including multiple logic circuits or analog circuits. For example, the MCU executes a program stored in a non-transient physical recording medium that serves as its own storage device. The program includes, for example, a progress evaluation processing program described below. Execution of the program results in the execution of a method corresponding to the program. The storage device may be, for example, a non-volatile memory. The programs stored in the storage device provided in the microcomputer and the storage unit 30 can be updated via a network such as the Internet.
[0021] The feature calculation unit 11 calculates feature amounts indicating characteristics of the operator's operation behavior. Specifically, the feature amounts of the operator's operation behavior refer to the operator's habits and tendencies in operation, such as the timing and speed of depressing the brake pedal or accelerator pedal. The feature calculation unit 11 calculates feature amounts based on a difference between a target pattern indicating changes in a target value of a progress operation parameter and an actual operation pattern indicating changes in an actually measured value of the progress operation parameter when the operator actually operates the moving object according to the target pattern. The feature calculation unit 11 may be configured to calculate all feature amounts that can be extracted based on a comparison between the target pattern and the actual operation pattern, or may be configured to calculate a portion of the extractable feature amounts (e.g., calculate feature amounts that significantly contribute to the difference between the target pattern and the actual operation pattern).
[0022] Specific examples of the feature quantity include a feature quantity related to speed, a feature quantity related to steering, a feature quantity related to operation timing, etc. Specific examples of the feature quantity related to speed include a vehicle speed, acceleration, deceleration, etc. Specific examples of the feature quantity related to steering include a steering angle, angular velocity, and each acceleration, etc. Specific examples of the feature quantity related to operation timing include operation timing of the brake, accelerator, turn signal lever, steering wheel, etc.
[0023] The travel operation parameters in the target pattern and the actual operation pattern are set according to the feature quantity to be extracted and the travel operation parameters in the evaluation pattern described below. Examples of travel operation parameters include the speed and acceleration / deceleration of the moving object, the accelerator and brake operation amounts and operation timings, and the steering amount and steering timings. For example, if the evaluation pattern is expressed by the relationship between the time elapsed since the start of vehicle travel and the vehicle speed to be reached at that time, i.e., expressed by the time change in the speed of the moving object, the target pattern and the actual operation pattern are also expressed by the time change in the speed of the moving object. It is preferable to set the change in the target value of the travel operation parameter in the target pattern according to the feature quantity to be extracted. For example, if the acceleration / deceleration of the moving object is to be extracted as a feature quantity, it is preferable to set a target pattern that accelerates and decelerates the speed of the moving object. The actual operation pattern can be obtained, for example, by actually measuring the travel operation parameters using a sensor or the like when the operator actually operates the moving object according to the target pattern. If the moving object is a vehicle, the actual measured values of the vehicle's speed and acceleration / deceleration can be obtained from a speed sensor, an acceleration / deceleration sensor, or the like. The actual measured values of the accelerator and brake operation amount and operation timing can be obtained from the brake pedal, accelerator pedal, etc. The actual measured values of the steering amount and steering timing can be obtained from a steering angle sensor, etc.
[0024] In this embodiment, the feature amount calculation unit 11 calculates the feature amount based on a target pattern and an actual operation pattern represented by a time change in the speed of the moving object, where the speed of the moving object is set as the travel operation parameter. In this case, the feature amount includes at least one of the following: a delay time of the actual operation pattern relative to the target pattern, a following speed, an overshoot at a speed peak, a speed difference, an acceleration / deceleration difference, and a difference in the acceleration / deceleration change amount. By extracting such feature amounts, the feature amount calculation unit 11 can appropriately reflect the characteristics of the operator's operation behavior in a feature reflection pattern, which will be described later. For example, as shown in FIG. 2 , the feature amount calculation unit 11 obtains a target speed pattern Vf(t) as the target pattern and an actual operation speed pattern Vr(t) as the actual operation pattern, and calculates a feature amount f, which indicates the characteristics of the operator's operation behavior, based on the difference between the target speed pattern Vf(t) and the actual operation speed pattern Vr(t). As shown in FIG. 4 , the target speed pattern Vf(t) is a pattern representing a time change in the target speed that the vehicle should reach. The actual operation speed pattern Vr(t) is a pattern that shows the change over time of the actual measured value of the speed of the moving body when the operator actually operates the moving body in accordance with the target speed pattern Vf(t).
[0025] The feature quantity preferably includes at least one-dimensional or two-dimensional feature quantities, and further includes three-dimensional or higher-dimensional feature quantities. For example, a speed difference is a one-dimensional feature quantity, an acceleration / deceleration difference is a two-dimensional feature quantity, and a difference in acceleration / deceleration change amount is a three-dimensional feature quantity. By extracting three-dimensional or higher-dimensional feature quantities in addition to one-dimensional or two-dimensional feature quantities, the features of the operator's operation behavior can be more appropriately reflected in a feature reflection pattern, which will be described later.
[0026] The feature calculation unit 11 may be configured to calculate an acceleration feature in a time domain in which the moving object accelerates and a deceleration feature in a time domain in which the moving object decelerates, separately. Taking into consideration the case in which the features of the operation behavior of the operator differ between when accelerating and when decelerating, the features of the operation behavior of the operator can be more appropriately reflected in the feature reflection pattern described below.
[0027] The operator model construction unit 20 constructs an operator model 31 that models the tendency of the moving object's moving operation by the operator. When performing a moving object progress evaluation, the operator model 31 acquires a predetermined evaluation pattern and uses the operator model 31 to move the moving object or a moving object model 32 while changing a predetermined moving operation parameter based on the evaluation pattern. In this embodiment, as shown in FIG. 2 , the operator model construction unit 20 sets the moving operation parameter to the speed of the moving object model 32 and constructs a driver model as the operator model 31.
[0028] The driver model is formed as a control rule when the moving body model 32 is virtually driven by a computer. In this embodiment, as shown in Fig. 2 , the driver model operates the accelerator pedal and brake pedal of the moving body model 32 based on the evaluation speed pattern Ve(t) input to the operator model construction unit 20, thereby controlling the speed of the moving body model 32.
[0029] The operator model construction unit 20 includes a pattern change unit 21 and a progress control unit 22. The pattern change unit 21 changes the evaluation pattern input to the driver model into a feature-reflecting pattern based on the feature value f calculated by the feature value calculation unit 11. The feature-reflecting pattern can be said to be a pattern that reflects the characteristics of the operator's operating behavior. As shown in FIG. 2 , the driver model is constructed as an operator model in which the pattern change unit 21 changes the input evaluation speed pattern Ve(t) into a feature-reflecting speed pattern Va(t), and the progress control unit 22 calculates values of the accelerator pedal and brake pedal operation amounts to be output to the mobile body model 32 using the feature-reflecting speed pattern Va(t) as a command vehicle speed. Based on the output values of the accelerator and brake pedal operation amounts, the driving state of the mobile body model 32 can be calculated on a computer to perform a driving simulation.
[0030] The evaluation pattern is typically expressed as a progression control parameter, such as the speed of the moving object, that is changed according to time, progression distance, etc. The evaluation pattern may be any pattern used to evaluate the progression of a moving object or a moving object model, and may be defined by a standard or may be defined individually for experimentation or development. Examples of evaluation patterns defined by standards include evaluation speed patterns related to vehicle fuel consumption rate tests, such as the WLTC (Worldwide Harmonized Light Duty Driving Test Cycle) mode and the JC08 mode. The evaluation speed pattern Ve(t) shown in FIG. 2 represents the evaluation speed pattern Ve(t) defined by the WLTC mode. The pattern change unit 21 changes the evaluation speed pattern Ve(t) to a feature-reflecting speed pattern Va(t) based on the feature value f.
[0031] The pattern modification unit 21 may be configured to calculate the feature reflection pattern by modifying the evaluation pattern using a transfer function set based on the feature value f, as shown in Fig. 5. As the transfer function, for example, a combination of dead time and an n-th order lag low-pass filter can be suitably used.
[0032] The pattern modification unit 21 may be configured to calculate the feature reflection pattern by modifying the evaluation pattern through machine learning such as a neural network using the feature amount f as learning data.
[0033] The progress control unit 22 controls the progress of the moving body or the moving body model 32 based on the feature reflection pattern. For example, as shown in FIG. 2 , the progress control unit 22 calculates accelerator pedal and brake pedal operation amount values to output to the moving body model 32 the feature reflection speed pattern Va(t) acquired from the pattern modification unit 21 as a command speed. Then, the progress control unit 22 calculates Vo(t), which is the speed when the moving body model 32 is traveling, based on the output accelerator and brake pedal operation amount values. The progress control unit 22 sets a proportional gain, a differential gain, and an integral gain using this speed Vo(t) as a feedback value for Va(t) as the command speed.
[0034] Fig. 3 is a flowchart of progress evaluation executed by the progress evaluation device 1. The processing shown in the flowchart in Fig. 3 is realized by the CPU constituting the progress evaluation unit 10 executing a progress evaluation program installed in the ROM, and is executed repeatedly at predetermined intervals.
[0035] In step S101, the target speed pattern Vf(t), the actual operation speed pattern Vr(t), and the evaluation speed pattern Ve(t) are obtained. For example, the target speed pattern Vf(t), the actual operation speed pattern Vr(t), and the evaluation speed pattern Ve(t) are obtained as shown in FIGS.
[0036] In step S102, a feature value f is calculated based on the target speed pattern Vf(t) and the actual operation speed pattern Vr(t) acquired in step S101. Step S102 corresponds to a feature value calculation step. Three types of feature values f1, f2, and f3 will be exemplified as the feature value f. For example, as shown in FIG. 4, when the target speed pattern Vf(t) and the actual operation speed pattern Vr(t) are acquired, the delay time of the actual operation speed pattern Vr(t) relative to the target speed pattern Vf(t) is represented by f1. More specifically, the delay time f1 can be calculated from the difference between time t1 and time t2, at which the speed values are the same (f1=t2−t1).
[0037] The following speed f2 of the actual operation speed pattern Vr(t) relative to the target speed pattern Vf(t) can be calculated based on the difference between the speed change f2f in the target speed pattern Vf(t) at a predetermined speed and the speed change f2r in the actual operation speed pattern Vr(t). In FIG. 4, the speeds v3 and v5 are equal, and (t4-t3) and (t6-t5) are equal. The speed change f2f can be calculated by f2f = (v4-v3) / (t4-t3), and the speed change f2r can be calculated by f2r = (v6-v5) / (t6-t5). The following speed f2 can be expressed, for example, by the ratio (f2r / f2f) of the speed change f2f to the speed change f2r.
[0038] The overshoot amount at the speed peak of the actual operation speed pattern Vr(t) relative to the speed peak of the target speed pattern Vf(t) is represented by f3, which can be calculated by f3=v7-v8.
[0039] The target speed pattern Vf(t) is set to a pattern that accelerates and then converges to a substantially constant speed so that the above-mentioned feature quantities f1, f2, and f3 can be obtained. In the target speed pattern Vf(t), a speed peak is formed between the time region of acceleration and the time region where the speed converges to a substantially constant speed while accelerating and decelerating thereafter, so that the overshoot quantity f3 can be extracted.
[0040] In step S103, the evaluation speed pattern Ve(t) acquired in step S101 is changed to a feature-reflecting speed pattern Va(t) based on the feature value f calculated in step S102. Step S103 corresponds to a pattern changing step. The feature-reflecting speed pattern Va(t) is a speed pattern obtained by changing the evaluation speed pattern Ve(t) so that the feature value f is reflected in the evaluation speed pattern Ve(t). To put it simply, for example, the feature-reflecting speed pattern Va(t) is changed to one that is delayed by a time corresponding to the delay time f1 relative to the evaluation speed pattern Ve(t). Furthermore, for example, the speed peak of the feature-reflecting speed pattern Va(t) is changed to one that exceeds the speed peak of the evaluation speed pattern Ve(t) by an amount corresponding to the overshoot amount f3.
[0041] In this embodiment, the feature-reflecting speed pattern Va(t) is calculated by modifying the evaluation speed pattern Ve(t) using a transfer function set based on the extracted multiple feature amounts f. More specifically, as shown in Fig. 5, by applying a transfer function configured from dead time and an n-th order lag low-pass filter to the evaluation speed pattern Ve(t), the feature-reflecting speed pattern Va(t) can be calculated by linking the characteristic constants (damping coefficient s and characteristic frequencies a0 to an-1) of the n-th order lag low-pass filter shown in Fig. 5 to the feature amount f.
[0042] In step S103, the feature-reflecting speed pattern Va(t) may be calculated by machine learning using a neural network (NN), as shown in Fig. 6. Numerical values a0 to an associated with the feature value f are input to the input layer of the neural network shown in Fig. 6, and the feature-reflecting speed pattern Va(t) is obtained from the output layer. By increasing the number of intermediate layers that form between the input layer and the output layer, it is possible to calculate a feature-reflecting speed pattern Va(t) that more accurately reflects the extracted feature value f, which contributes to improving the accuracy of the driving simulation described below.
[0043] FIG. 7 shows a portion of the evaluation speed pattern in the WLTC mode as the evaluation speed pattern Ve(t). FIG. 8 shows an enlarged view of the evaluation speed pattern Ve(t) from time 0 to t11 in FIG. 7, and also shows a feature-reflecting speed pattern Va(t) obtained by modifying the evaluation speed pattern Ve(t) by reflecting the feature value f. By using the technique using the transfer function shown in FIG. 5 or a machine learning technique such as the neural network shown in FIG. 6 to reflect the feature value f in the evaluation speed pattern Ve(t), the evaluation speed pattern can be modified to the feature-reflecting speed pattern Va(t) as shown in FIG. 8. As shown in FIG. 8, the feature-reflecting speed pattern Va(t) is calculated assuming that there is a delay time and an overshoot amount compared to the evaluation speed pattern Ve(t), and it can be seen that the feature value f is reflected.
[0044] In step S104, the moving body model 32 is controlled to move forward using the characteristic-reflecting speed pattern Va(t) estimated in step S103 as a command speed, and a running simulation is performed as a progress evaluation. Step S104 corresponds to a progress control step. Since the characteristic-reflecting speed pattern Va(t) reflects the feature value f that indicates the characteristics of the operator's operating behavior, by performing running control of the moving body model 32 using the characteristic-reflecting speed pattern Va(t) as a command speed, it is possible to achieve human-like operation, as if the operator were operating the moving body model 32 with the evaluation speed pattern Ve(t) as a target.
[0045] As described above, according to the progress evaluation program and the progress evaluation method executed thereby according to this embodiment, in step S101, the speed of the moving body model 32 is set as the progress operation parameter, a target speed pattern Vf(t) is acquired as the target pattern, an actual operation speed pattern Vr(t) is acquired as the actual operation pattern, and an evaluation speed pattern Ve(t) is acquired as the evaluation pattern. In a feature amount calculation step (step S102), a feature amount f indicating the feature of the operator's operation behavior is calculated based on the difference between the target speed pattern Vf(t) and the actual operation speed pattern Vr(t). In a pattern change step (step S103), the evaluation speed pattern Ve(t) is changed to a feature-reflecting speed pattern Va(t) based on the feature amount f calculated in the feature amount calculation step. By reflecting the feature amount f in the evaluation speed pattern Ve(t), it is possible to change to a feature-reflecting speed pattern Va(t) that reflects the feature of the operator's operation behavior. Therefore, the change in speed over time in the characteristic-reflecting speed pattern Va(t) is close to the change in speed over time when the operator actually operates the moving body model 32 with the evaluation speed pattern Ve(t) as the target. Therefore, in the progress control step (S104), by controlling the moving body model 32 using the characteristic-reflecting speed pattern Va(t) as the command speed, it is possible to make the moving body model 32 run as if the operator were actually running with the evaluation speed pattern Ve(t) as the target. By the pattern changing step and the progress control step, it is possible to construct a driver model that performs human-like operation and to run a driving simulation of the moving body model 32, without repeatedly running a driving simulation.
[0046] In the above embodiment, the moving body is a vehicle, but the present invention is not limited to this, and the technology related to the progress evaluation described above can be applied to an apparatus, program, and method for evaluating the progress of any moving body and its moving body model, such as an aircraft or a ship. For example, if the moving body is an aircraft, a pilot model that performs human-like operations can be constructed without repeatedly running simulations, and the progress of the aircraft and aircraft model can be evaluated based on the pilot model.
[0047] The controller and methods described herein may be implemented by a special-purpose computer configured with a processor and memory programmed to perform one or more functions embodied in a computer program. Alternatively, the controller and methods described herein may be implemented by a special-purpose computer configured with a processor configured with one or more dedicated hardware logic circuits. Alternatively, the controller and methods described herein may be implemented by one or more special-purpose computers configured with a processor and memory programmed to perform one or more functions in combination with a processor configured with one or more hardware logic circuits. Furthermore, the computer program may be stored as instructions executed by a computer on a computer-readable non-transitory storage medium.
[0048] Characteristic configurations extracted from the above-described embodiments are described below. [Configuration 1] A progress evaluation device (1) that progresses a moving body or a moving body model that models the moving body while changing a predetermined progress operation parameter based on a predetermined evaluation pattern, and performs a progress evaluation of the moving body, comprising: a feature calculation unit (11) that calculates a feature amount that indicates a feature of the operation behavior of the operator based on a difference between a target pattern that indicates a change in a target value of the progress operation parameter and an actual operation pattern that indicates a change in an actual measured value of the progress operation parameter when the operator actually operates the moving body in accordance with the target pattern, a pattern change unit (21) that changes the evaluation pattern based on the feature amount calculated by the feature calculation unit, and changes the evaluation pattern to a feature reflecting pattern that reflects the feature of the operation behavior of the operator, and a progress control unit (22) that progresses the moving body or the moving body model based on the feature reflecting pattern when performing the progress evaluation. [Configuration 2] The progress evaluation device according to Configuration 1, wherein the feature calculation unit calculates the feature based on the target pattern and the actual operation pattern, in which the speed of the moving object is set as the progress operation parameter, and the feature includes at least one of a delay time of the actual pattern relative to the target pattern, a following speed, an overshoot at a speed peak, a speed difference, an acceleration / deceleration difference, and a difference in acceleration / deceleration change amount. [Configuration 3] The progress evaluation device according to Configuration 1 or 2, wherein the feature calculation unit, when calculating the feature based on the target pattern and the actual operation pattern input as the progress operation parameter, distinguishes between an acceleration feature in a time domain in which the moving object accelerates and a deceleration feature in a time domain in which the moving object decelerates. [Configuration 4] The progress evaluation device according to any of Configurations 1 to 3, wherein the feature includes at least a one-dimensional or two-dimensional feature and further includes a three-dimensional or higher-dimensional feature. [Configuration 5] The progress evaluation device according to any one of Configurations 1 to 4, wherein the pattern modification unit calculates the feature reflecting pattern by modifying the evaluation pattern using a transfer function set based on the feature amount.[Configuration 6] The progress evaluation device according to any one of Configurations 1 to 5, wherein the pattern change unit calculates the feature reflecting pattern by changing the evaluation pattern by machine learning using the feature amount as learning data. [Configuration 7] A progress evaluation program for progressing a moving object or a moving object model that models the moving object while changing a predetermined progress operation parameter based on a predetermined evaluation pattern, and performing progress evaluation of the moving object, the program causing a computer to execute: a feature amount calculation step of calculating a feature amount that indicates a feature of the operation behavior of the operator based on a difference between a target pattern that indicates a change in a target value of the progress operation parameter and an actual operation pattern that indicates a change in an actual measured value of the progress operation parameter when the operator actually operates the moving object in accordance with the target pattern; a pattern change step of changing the evaluation pattern based on the feature amount calculated in the feature amount calculation step, and changing the evaluation pattern to a feature reflecting pattern that reflects the feature of the operation behavior of the operator; and a progress control step of progressing the moving object or the moving object model based on the feature reflecting pattern when performing the progress evaluation.
[0049] Although the present disclosure has been described with reference to the embodiments, it is understood that the present disclosure is not limited to the embodiments or structures. The present disclosure also encompasses various modifications and equivalent modifications. In addition, various combinations and forms, including only one element, more than one element, or less than one element, are also within the scope and spirit of the present disclosure.
Claims
1. A progress evaluation device (1) that progresses a moving body or a moving body model that models the moving body while changing a predetermined progress operation parameter based on a predetermined evaluation pattern, and performs a progress evaluation of the moving body, comprising: a feature calculation unit (11) that calculates a feature amount that indicates a characteristic of the operator's operating behavior based on a difference between a target pattern that indicates a change in a target value of the progress operation parameter and an actual operation pattern that indicates a change in an actual value of the progress operation parameter when the operator actually operates the moving body in accordance with the target pattern; a pattern change unit (21) that changes the evaluation pattern based on the feature amount calculated by the feature calculation unit, and changes the evaluation pattern to a feature reflection pattern that reflects the characteristic of the operator's operating behavior; and a progress control unit (22) that progresses the moving body or the moving body model based on the feature reflection pattern when performing the progress evaluation.
2. The progress evaluation device of claim 1, wherein the feature calculation unit calculates the feature based on the target pattern and the actual operation pattern, in which the speed of the moving body is set as the progress operation parameter, and the feature includes at least one of the delay time of the actual pattern relative to the target pattern, the following speed, the amount of overshoot at a speed peak, the speed difference, the acceleration / deceleration difference, and the difference in the amount of acceleration / deceleration change.
3. A progress evaluation device as described in claim 1 or 2, wherein when the feature calculation unit calculates the speed of the moving body based on the target pattern and the actual operation pattern input as the progress operation parameters, the feature calculation unit distinguishes between an acceleration feature in a time domain in which the moving body accelerates and a deceleration feature in a time domain in which the moving body decelerates.
4. A progress assessment device according to claim 1 or 2, wherein the feature quantities include at least one-dimensional or two-dimensional feature quantities, and further include three or more-dimensional feature quantities.
5. A progress evaluation device as described in claim 1 or 2, wherein the pattern modification unit calculates the characteristic reflecting pattern by modifying the evaluation pattern using a transfer function set based on the characteristic amount.
6. A progress evaluation device as described in claim 1 or 2, wherein the pattern modification unit calculates the feature-reflecting pattern by modifying the evaluation pattern through machine learning using the features as learning data.
7. A progress evaluation program for progressing a moving body or a moving body model modeled on said moving body while changing a predetermined progress operation parameter based on a predetermined evaluation pattern, and performing a progress evaluation of said moving body or a moving body modeled on said moving body, which causes a computer to execute the following steps: a feature calculation step for calculating a feature indicating a characteristic of the operating behavior of the operator based on a difference between a target pattern indicating a change in a target value of the progress operation parameter and an actual operation pattern indicating a change in an actual value of the progress operation parameter when the operator actually operates the moving body in accordance with the target pattern; a pattern change step for changing the evaluation pattern based on the feature calculated in the feature calculation step, and changing the evaluation pattern to a feature reflecting pattern that reflects the characteristic of the operating behavior of the operator; and a progress control step for progressing the moving body or the moving body model based on the feature reflecting pattern when performing the progress evaluation.
Citation Information
Patent Citations
Vehicle behavior simulation device and method
JP2012220378A
Vehicle data analysis method and system
JP2012242894A
Automatic operation device of vehicle
JP2017061168A
Learning system and learning method for operation deduction learning model for controlling automatic operation robot
JP2021128510A
