Control characteristic selection system and control characteristic selection method

Abstract

This control characteristic selection system comprises a control characteristic selection device that searches for a control characteristic of an actuator that contributes to the travel characteristics of a mobile body, the control characteristic selection device comprising: a database in which a plurality of combinations of control characteristics and sensory evaluation points are stored; a sensory evaluation point estimation unit that accepts input of a control characteristic as a current characteristic and outputs a sensory evaluation point as a current evaluation point; an input unit that accepts a target value, which is a target sensory evaluation point; and a selection unit that selects, from the control characteristics stored in the database, a control characteristic that has a sensory evaluation point similar to the target value and that is similar to the current characteristic, as a selected characteristic.

Description

Control characteristic selection system and control characteristic selection method

[0001] The present invention relates to a control characteristic selection system and a control characteristic selection method.

[0002] When operating an actuator in response to a human input operation, it is essential to make adjustments that take human sensations into consideration, i.e., adjust the control characteristics of the actuator. Patent Document 1 discloses a control parameter adjustment teaching device for an electric power steering control device that performs control based on various control parameter data, the device including: storage means for storing evaluation data during steering related to the electric power steering control device; steering situation data indicating under what circumstances the evaluation data was performed and evaluated, steering situation data associated with the evaluation data; control parameter data of the electric power steering control device that is associated with the steering situation data and the evaluation data and causes the evaluation data to be expressed under the associated steering situation data; data related to the control parameter data, adjusting method data for the control parameter data to resolve the evaluation content of the associated evaluation data; first input means for inputting the evaluation data; second input means for inputting the steering situation data; reading means for reading adjusting method data for the control parameter data associated with the input data from the storage means based on the evaluation data and steering situation data input by the two input means; and output means for outputting the adjusting method data for the read control parameter data.

[0003] Japanese Patent Application Publication No. 2002-173042

[0004] The invention described in Patent Document 1 leaves room for improvement in the selection of control characteristics.

[0005] A control characteristic selection system according to a first aspect of the present invention is a control characteristic selection device that searches for control characteristics of actuators that contribute to the driving characteristics of a moving body, and includes a database in which multiple combinations of the control characteristics and sensory evaluation points are stored, a sensory evaluation point estimation unit that inputs the control characteristics as current characteristics and outputs the sensory evaluation points as current evaluation points, an input unit that accepts a target value that is the sensory evaluation point to be targeted, and a selection unit that selects, from the control characteristics stored in the database, a control characteristic whose sensory evaluation point is similar to the target value and is similar to the current characteristic as a selected characteristic. A control characteristic selection method according to a second aspect of the present invention is a control characteristic selection method executed by a computer capable of searching a database storing multiple combinations of control characteristics of actuators that contribute to the driving characteristics of a mobile body and sensory evaluation scores of the control characteristics, and includes a sensory evaluation point process that inputs the control characteristic as a current characteristic and outputs the sensory evaluation score as the current evaluation point, an input process that accepts a target value that is the sensory evaluation point to be targeted, and a selection process that selects, from the control characteristics stored in the database, a control characteristic whose sensory evaluation score is similar to the target value and is similar to the current characteristic as a selected characteristic, and in the selection process, a control characteristic that satisfies the target value and has a small difference in characteristic value from the current characteristic is selected as the selected characteristic.

[0006] According to the present invention, it is possible to select control characteristics that satisfy the user's requirements while preventing abrupt changes in ride comfort.

[0007] 1. Configuration diagram of a control characteristic selection device. 2. Diagram showing the hardware configuration of a control characteristic selection device. 3. Diagram showing an example of sensory evaluation scores. 4. Diagram showing an example of data stored in a database. 5. Diagram showing the relationship between control characteristics and parameters. 6. Diagram explaining learning of a neural network constituting a control characteristic selection unit. 7. Diagram showing the configuration of a neural network when there are multiple nodes in the output layer. 8. Functional configuration diagram of a control characteristic selection unit. 9. Flowchart showing the processing of a control characteristic selection device. 10. Diagram showing an example display of a display unit. 11. Diagram showing a display when a control characteristic is changed after shipping a vehicle in Modification 1. 12. Diagram showing restrictions on input of target values.

[0008] -First Embodiment- A first embodiment of a control characteristic selection device will be described below with reference to FIGS.

[0009] FIG. 1 is a configuration diagram of a control characteristic selection device 1. In this embodiment, a person who uses the control characteristic selection device 1 is called a "user." In this embodiment, a person involved in the manufacture of a vehicle equipped with the control characteristic selection device 1 is assumed to be the user. The control characteristic selection device 1 includes a sensory evaluation point estimating unit 101, a setting unit 103, a control characteristic selection unit 104, a database 102, an input unit 105, and a display unit 106. A current characteristic 51 is input to the sensory evaluation point estimating unit 101 and the setting unit 103. The control characteristic is the characteristic of an actuator that contributes to the running characteristics of a mobile object, such as the steering, suspension, and brakes. The control characteristic can also be expressed as the magnitude of a second physical quantity corresponding to a first physical quantity. There is only one second physical quantity corresponding to a first physical quantity. The control characteristic can be expressed by a polynomial or the like, and a "parameter," which will be described later, is a simplified version of the control characteristic.

[0010] The sensory evaluation point estimating unit 101 estimates a sensory evaluation point corresponding to the input control characteristic and outputs a current evaluation point 52. The sensory evaluation point estimating unit 101 includes a neural network, as will be described later. This neural network has undergone a learning process, i.e., weights have been determined in advance. A large number of combinations of control characteristics and sensory evaluation points are stored in the database 102. The display unit 106 is a display device, such as a liquid crystal display, that presents information to the user. The display unit 106 displays the current evaluation point 52 output by the sensory evaluation point estimating unit 101 and the selected characteristic 54 output by the setting unit 103. The display unit 106 may further display the current characteristic 51 and the target value 53.

[0011] The input unit 105 is, for example, a keyboard or a mouse. The user inputs a target value 53, which is a sensory evaluation score to be aimed for, using the input unit 105 while looking at the display unit 106. Since absolute evaluation of the sensory evaluation score is not easy, the user inputs the target value 53 using the current evaluation score 52 of the current characteristic 51 as a reference. The current characteristic 51, the evaluation score 502 output by the sensory evaluation point estimation unit 101, and the target value 53 input from the input unit 105 are input to the setting unit 103. The input unit 105 and the display unit 106 may be realized by a single touch panel.

[0012] 2 is a diagram showing the hardware configuration of the control characteristic selection device 1. The control characteristic selection device 1 includes a CPU 41, which is a central processing unit, a ROM 42, which is a read-only storage device, a RAM 43, which is a readable and writable storage device, an input / output device 44, which is a user interface, and a communication device 45. The CPU 41 loads a program stored in the ROM 42 into the RAM 43 and executes it to perform the various calculations described above.

[0013] The control characteristic selection device 1 may be realized by a field programmable gate array (FPGA), which is a rewritable logic circuit, or an application specific integrated circuit (ASIC), which is an integrated circuit for a specific application, instead of the combination of the CPU 41, the ROM 42, and the RAM 43. Furthermore, the control characteristic selection device 1 may be realized by a combination of different configurations, for example, a combination of the CPU 41, the ROM 42, the RAM 43, and the FPGA, instead of the combination of the CPU 41, the ROM 42, and the RAM 43.

[0014] The input / output device 44 is an interface for data input / output with the user. Examples of the input / output device 44 include a display, a mouse, a keyboard, a touch panel, a speaker, and a microphone. The communication device 45 realizes communication with external devices including the database 102. Any communication standard may be used by the communication device 45.

[0015] The control characteristic selection device 1 may receive input from an external device via the communication device 45. For example, a touch panel or buttons may be provided on the instrument panel of a vehicle equipped with the control characteristic selection device 1, and the control characteristic selection device 1 may receive input operations made by a user via the instrument panel. The communication device 45 may also communicate with a mobile terminal carried by the user, and the control characteristic selection device 1 may receive input operations made by the user to the communication device 45. Furthermore, instead of including the input / output device 44, the control characteristic selection device 1 may communicate with a device external to the control characteristic selection device 1 via the communication device 45, and the external device may be used in place of the input / output device 44. In this case, data output by the control characteristic selection device 1 is displayed on the external device.

[0016] 3A and 3B are diagrams showing examples of sensory evaluation scores. FIG. 3A shows a case where there are five evaluation items, and FIG. 3B shows a case where there is one evaluation item. In FIG. 3A, a pentagonal radar chart is used to display the evaluation values ​​of the five evaluation items at once. In FIG. 3B, a bar graph is used. The user can refer to the current evaluation score 52 on the display shown in FIG. 2, and input the target value 53 using the input unit 105 while viewing this display.

[0017] In this example, the sensory evaluation score ranges from 0 to 7 points, with auxiliary lines set every 0.25 points. The numerical range of the sensory evaluation score is an example and may be different. In the example of Figure 3(a), the range of 3.50 to 4.50 points, which is the most frequently answered range, is expanded, but the display method is not limited. Also, while the illustration shows cases with one and five evaluation items, there is no upper limit as long as the number of evaluation items is one or more.

[0018] Fig. 4 is a diagram showing an example of data stored in the database 102. The database 102 stores characteristics and evaluation scores 502 for each evaluation item in association with each other. However, Fig. 4 shows a case where there are multiple evaluation items. When there is one evaluation item as shown in Fig. 3(b), there is one value for the evaluation item corresponding to each characteristic.

[0019] FIG. 5 is a diagram showing the relationship between control characteristics and parameters. Three curves 504-1 to 504-3 shown in FIG. 5 represent different characteristics. These characteristics show the relationship between the steering angle (or pinion angle during steering) and friction torque, with the x-axis representing the steering angle and the y-axis representing the friction torque. All characteristics show a similar tendency for friction torque to increase as the steering angle increases. In this embodiment, the "control characteristics" shown by the curves are converted into "parameters" by sampling at predetermined value increments on the horizontal axis. When converting the control characteristics of steering angle and friction torque into calculation parameters as shown in this figure, the friction torque values ​​are sampled at predetermined steering angle increments, for example, in 1-degree increments, and the friction torque values ​​from 0 degrees to 90 degrees are used as "parameters."

[0020] FIG. 6 is a diagram illustrating the learning process of the neural network 1041 constituting the control characteristic selection unit 104. First, a test driver drives a vehicle to which pre-created characteristics are applied, and determines sensory values. This process is repeated multiple times to create combinations of characteristics and sensory values, i.e., training data for the neural network 1041. The neural network 1041 includes an input layer 305, a hidden layer 306, and an output layer 307. The number of nodes in the input layer 305 is uniquely determined by the parameter determination method. For example, in the example described with reference to FIG. 5, the angle is in 1-degree increments from 0 degrees to 90 degrees, so the input layer 305 has 91 nodes. For convenience of illustration, FIG. 6 shows only one hidden layer 306, but multiple hidden layers 306 may exist. The number of nodes in the hidden layer 306 is arbitrary. The number of nodes in the output layer 307 may be one, as shown in this figure, or multiple, as described below.

[0021] In the learning phase of the neural network 1041, parameters created based on each characteristic are input to the input layer 305, and the weights of the neural network 1041 are determined so that the value of the output layer 307 corresponds to the sensory value determined by the test driver. A known method, such as backpropagation, can be used to determine the weights. The configuration of the hidden layer 306 may also be determined by learning. Because the learning process requires a high computational load, it is desirable to use a device with high computing power and looser power consumption restrictions, such as a server, rather than the control characteristic selection device 1.

[0022] FIG. 7 is a diagram showing the configuration of the neural network 1041 when the output layer 307 has multiple nodes. In this case, the output layer 307 employs one-hot encoding. That is, each node is assigned a sensory evaluation value in advance, with only one set to "1" and all others set to "0." For example, if the output layer 307 has five nodes and the evaluation score 502 ranges from "3.50" to "4.50," the combinations shown at the bottom of FIG. 5 are obtained. For example, if a test driver evaluates the sensory value of a certain test characteristic as "3.50," the weights of the neural network 1041 are trained so that only "Y1" at the top of the output layer 307 is set to "1" and the others are set to "0."

[0023] It is not essential to use an actual vehicle and a test driver to create the training data. A simulator that simulates the behavior of an actual vehicle may be used instead of an actual vehicle, and sensor output may be used instead of the sensory values ​​determined by a test driver. In other words, in addition to the combination of an actual vehicle and a test driver, any of the other three combinations, i.e., a simulator and a test driver, an actual vehicle and a sensor, or a simulator and a virtual sensor, may be used.

[0024] In the case of a combination of a simulator and a test driver, the test driver operates the simulator and determines the sensory value. In the case of a combination of an actual vehicle and a sensor, the output of the sensor mounted on the actual vehicle is classified using predetermined categories to obtain the sensory value. In the case of a combination of a simulator and a sensor simulator, a virtual sensor is installed in the simulator that mimics the vehicle's operation, and the output of this sensor is classified using predetermined categories to obtain the sensory value.

[0025] 8 is a functional configuration diagram of the control characteristic selection unit 104. The control characteristic selection unit 104 includes a presentation number setting unit 701, a mode setting unit 702, a control unit 703, and a reading unit 704. A presentation number command 58 is input to the presentation number setting unit 701. The presentation number command 58 may be input by the user from the input unit 105, or a predetermined value may be input. The presentation number setting unit 701 outputs a presentation number 71 to the control unit 703 based on the presentation number command 58. The presentation number command 58 may be, for example, a natural number, and the presentation number setting unit 701 may output the presentation number command 58 directly to the control unit 703 as the presentation number 71.

[0026] A mode command 59 is input to the mode setting unit 702. The mode command 59 may be input by the user from the input unit 105, or a predetermined value may be input. Based on the mode command 59, the mode setting unit 702 selects one of three modes, namely, a parameter similarity priority mode, a trade-off influence minimum mode, and a specified item priority mode, as the determined mode 72. The mode command 59 is a value indicating one of the three modes.

[0027] The presentation number 71 output by the presentation number setting unit 701 and the decision mode 72 output by the mode setting unit 702 set the operation of the reading unit 704. The presentation number 71 sets the number of characteristics output by the reading unit 704. For example, if the presentation number 71 is "1", the reading unit 704 outputs one selected characteristic 54 that best matches the conditions. If the presentation number 71 is "5", the reading unit 704 outputs the top five selected characteristics 54 that best match the conditions. The difference in operation of the reading unit 704 depending on the decision mode 72 is as follows.

[0028] In the parameter similarity priority mode, the reading unit 704 selects a characteristic that satisfies the target value 53 and has a high similarity to the current characteristic 51. Similarity is the difference in characteristic values, and the smaller the difference in characteristic values, the higher the similarity. For example, if the characteristic 504-1 shown in FIG. 5 is the current characteristic 51, and the characteristics 504-2 and 504-3 are characteristics that satisfy the target value 53, the characteristic 504-2 is selected as the most suitable characteristic. This is because the difference in the friction torque value between the characteristic 504-3 and the characteristic 504-1 at each steering angle is smaller than that between the characteristic 504-3 and the characteristic 504-1.

[0029] The trade-off influence minimum mode and designated evaluation item priority mode, which will be described next, are based on the premise that there are multiple evaluation items for the sensory evaluation score, as shown in FIG. 3A. Among the multiple evaluation items, one or more are designated as representative items. The representative items may be designated by the user each time, or may be preset in the input unit 105.

[0030] In the trade-off influence minimum mode, the reading unit 704 selects a characteristic for the representative item that has a value closer to the target value 53 than the current characteristic 51. If there are multiple characteristics with the same value for the representative item, the characteristic for which the total change in the sensory evaluation score is the smallest is selected for evaluation items other than the representative item. If there are multiple characteristics with the same total change in the sensory evaluation score, the characteristic with the least change from the current characteristic 51 is selected as the selected characteristic 54.

[0031] In the designated evaluation item priority mode, the reading unit 704 selects a characteristic of the representative item that exceeds the target value 53. In this mode, evaluation points 502 of items other than the representative item are not taken into consideration. If there are multiple characteristics of the representative item that exceed the target value 53, the characteristic whose representative item value is closest to the target value 53 is selected. If there are multiple characteristics whose representative item value is closest to the target value 53, the characteristic whose change from the current characteristic 51 is the least is selected as the selected characteristic 54.

[0032] The control unit 703 transfers the search conditions to the reading unit 704 based on the outputs of the presentation number setting unit 701 and the mode setting unit 702. The reading unit 704 searches the database 102 for characteristics that match the search conditions and outputs them to the control unit 703. The control unit 703 transfers the characteristics output by the reading unit 704 to the setting unit 103. The reading unit 704 searches the database 102 in accordance with the mode command 59 set by the mode setting unit 702, and outputs the selected characteristics 54 in the number of presentation number commands 58 set by the presentation number setting unit 701. The operation of the reading unit 704 in each mode is as described above.

[0033] 9 is a flowchart showing the processing of the control characteristic selection device 1. In step S801, the current characteristic 51 is input to the sensory evaluation point estimating unit 101 and the setting unit 103. In the following step S802, the sensory evaluation point estimating unit 101 calculates a current evaluation point 52 and outputs it to the display unit 106 and the setting unit 103, and the process proceeds to step S803. In step S803, the display unit 106 displays the current evaluation point 52. In the following step S804, the input unit 105 accepts input of a target value 53 by the user, and the process proceeds to step S805.

[0034] In step S805, the control characteristic selection unit 104 calculates the selected characteristic 54. In the following step S806, the display unit 106 displays the selected characteristic 54 calculated in step S805. In the following step S807, the setting unit 103 determines whether or not the user has re-inputted a value, i.e., whether or not the user has input a new target value 53 from the input unit 105. If it is determined that the user has re-inputted a value, the process returns to step S805; if it is determined that the user has not re-inputted a value, the process shown in FIG. 9 ends.

[0035] 10A and 10B are diagrams showing examples of the display on the display unit 106. This diagram shows an example in which a mobile terminal is used as the display unit 106. FIG. 10A shows a current characteristic 51 and a current evaluation score 52. The current characteristic 51 may be input by the user, or the current setting may be automatically read. When the user inputs the current characteristic 51, the user may input it on the screen using a touch pen or the like, or the current setting may be read from a CSV file or the like.

[0036] Figure 10(b) shows a target value 53 and a selection characteristic 54. The current evaluation score 52 is shown in a solid line in the upper part of Figure 10(b), and the user can input a target value 53 by evaluating it relative to the current evaluation score 52. This input may be made using a touch panel or by voice input. The display shown in Figure 10 is merely an example, and a different layout may be used, or all of the information shown in Figure 10 may be displayed on one screen.

[0037] The first embodiment described above provides the following advantageous effects. (1) The control characteristic selection device 1 searches for control characteristics of actuators that contribute to the driving characteristics of a mobile object. The control characteristic selection device 1 includes a database 102 that stores a plurality of combinations of control characteristics and sensory evaluation scores, a sensory evaluation point estimating unit 101 that inputs a control characteristic as a current characteristic 51 and outputs a sensory evaluation score as a current evaluation score 52, an input unit 105 that receives a target value 53 that is a target sensory evaluation score, and a control characteristic selection unit 104 that selects, from the control characteristics stored in the database 102, a control characteristic whose sensory evaluation score is similar to the target value 53 and is similar to the current characteristic 51 as a selected characteristic 54. Therefore, a characteristic that is similar not only to the target value 53 specified by the user but also to the current characteristic is selected, thereby satisfying the user's requirements while preventing a sudden change in ride comfort.

[0038] (2) In the parameter similarity priority mode, the control characteristic selection unit 104 selects, as the selected characteristic 54, a control characteristic whose sensory evaluation score is similar to the target value 53 and whose difference from the current characteristic 51 is small.

[0039] (3) The sensory evaluation score includes multiple evaluation items including a representative item. In the trade-off influence minimization mode, the control characteristic selection unit 104 selects, as the selected characteristic 54, a control characteristic whose representative item of the sensory evaluation score is similar to the target value 53 and which minimizes the total change in the sensory evaluation score of the items other than the representative item.

[0040] (4) In the designated evaluation item priority mode, the control characteristic selection unit 104 selects as the selected characteristic 54 the control characteristic whose representative item of the sensory evaluation score exceeds the target value 53 and whose value is closest to the target value 53.

[0041] (5) The control characteristic indicates a second physical quantity corresponding to the first physical quantity, for example, a friction torque corresponding to a steering angle. The sensory evaluation point estimating unit 101 specifies parameters as described with reference to Fig. 5 to evaluate the current characteristic 51. That is, the value of the second physical quantity when the first physical quantity is increased by a predetermined unit value is specified as a parameter, and the parameter is used as an input to the neural network 1041 that has been trained in advance.

[0042] (6) The actuator is one of steering, suspension, and brakes.

[0043] (Variation 1) In the above-described embodiment, it is assumed that a person involved in the vehicle production process is the “user” who operates the control characteristic selection device 1. However, the control characteristic selection device 1 may also be operated by a person who uses the vehicle after the vehicle has been shipped.

[0044] 11A and 11B are diagrams illustrating a display when changing control characteristics after a vehicle is shipped. FIG. 11A illustrates a scene in which the control characteristics selection device 1 is operated using an instrument panel 1001 of a vehicle equipped with the control characteristics selection device 1. FIG. 11B illustrates a scene in which the control characteristics selection device 1 is operated using a mobile terminal 1004 carried by a user. In FIG. 11A, the communication device 45 communicates with the control device of the instrument panel 1001, causing the instrument panel 1001 to function as the input / output device 44. That is, the target value 53 and the selected characteristic 54 are displayed on the instrument panel 1001. In FIG. 11B, the target value 53 and the selected characteristic 54 are displayed on the mobile terminal 1004. The user can input the target value 53 using the instrument panel 1001 or the mobile terminal 1004 and obtain the selected characteristic 54.

[0045] FIG. 12 is a diagram showing limitations on the input of the target value 53. When the target value 53 is set by a general user, rather than by the vehicle manufacturer, after the vehicle has been shipped, it is desirable to narrow the range within which the target value 53 can be set for safety reasons. Changing the characteristics of the suspension, steering, etc. changes the driving feel and ride comfort. Furthermore, applying inappropriate changes may result in vehicle behavior that is unexpected for the driver. Therefore, the settable range of the target value 53 is determined in advance, and the user is allowed to input the target value 53 within the settable range.

[0046] In the example shown in Fig. 12(a), the range that cannot be set is indicated by hatching. In the example shown in Fig. 12(b), the settable values ​​are indicated by white circles. The values ​​that can be set as the target value 53 are the same in Fig. 12(a) and Fig. 12(b).

[0047] (Variation 2) In the above-described embodiment, the control characteristic selection device 1 is configured as a single hardware device as shown in Fig. 2. However, the functions of the control characteristic selection device 1 may be realized by a plurality of hardware devices. In this case, a group of a plurality of devices having functions equivalent to those of the control characteristic selection device 1 can be called a "control characteristic selection system."

[0048] (Variation 3) In the embodiment described above, the control characteristics are converted into parameters by sampling at predetermined horizontal axis value intervals, as explained with reference to FIG. 5, and these parameters are input to the neural network. However, the control characteristics diagram shown in FIG. 5 may be treated as image data, and this image data may be input to the neural network 1041. This image is generated after both the friction torque and the steering angle are set to predetermined minimum and maximum values. When this variation is adopted, it is necessary to re-run the learning of the neural network 1041.

[0049] - Second embodiment - A second embodiment of a control characteristic selection system will be described with reference to Fig. 13. In the following description, the same components as those in the first embodiment are designated by the same reference numerals, and differences will be mainly described. Points that are not particularly described are the same as those in the first embodiment. This embodiment differs from the first embodiment mainly in that data is added to the database.

[0050] 13 is a configuration diagram of a control characteristic selection system S1 according to the second embodiment. The control characteristic selection system S1 further includes a characteristic adding device 1200 in addition to the control characteristic selection device 1 according to the first embodiment. The characteristic adding device 1200 includes a characteristic data generating unit 1201, a database managing unit 1203, and a second sensory evaluation point estimating unit 1202. The operation of the second sensory evaluation point estimating unit 1202 is similar to that of the sensory evaluation point estimating unit 101. Like the sensory evaluation point estimating unit 101, the second sensory evaluation point estimating unit 1202 also includes a trained neural network.

[0051] The characteristic data generating unit 1201 generates a large number of automatically generated control characteristics 1251, which are control characteristics. The automatically generated control characteristics 1251 may be generated randomly, or a general trend, such as a monotonically increasing trend, may be specified in advance. The second sensory evaluation point estimating unit 1202 generates parameters from each of the automatically generated control characteristics 1251, inputs the parameters to a built-in neural network, and outputs an automatically generated evaluation value 1252.

[0052] The database management unit 1203 receives an automatically generated control characteristic 1251 from the characteristic data generation unit 1201 and an automatically generated evaluation value 1252 from the second sensory evaluation point estimator 1202. The database management unit 1203 evaluates the validity of each automatically generated control characteristic 1251. If the database management unit 1203 determines that there is no problem with the validity, it adds the combination of the automatically generated control characteristic 1251 and the automatically generated evaluation value 1252 to the database 102. The evaluation of the validity of the automatically generated control characteristic 1251 by the database management unit 1203 includes a monotonically increasing trend, upper and lower limits of the minimum value, upper and lower limits of the maximum value, and an upper limit of the amount of change.

[0053] The hardware configuration of the characteristic adding device 1200 may be the same as the hardware configuration of the control characteristic selecting device 1 shown in Fig. 2. However, since the characteristic adding device 1200 does not require input from the user, the characteristic adding device 1200 does not need to include the input / output device 44.

[0054] According to the second embodiment described above, the following advantageous effects can be obtained: (7) A database management unit 1203 is provided that adds to the database 102 a combination of an automatically generated control characteristic 1251, which is an automatically generated control characteristic, and an automatically generated evaluation value 1252, which is a sensory evaluation score obtained by evaluating the automatically generated control characteristic 1251. This makes it possible to expand the number of records in the database 102, and to potentially obtain a new control characteristic that is closer to the target value 53 than before.

[0055] (Variation of the Second Embodiment) If the automatically generated evaluation value 1252 for the automatically generated control characteristic 1251 added to the database 102 does not match the sensory evaluation value perceived by the user, the user may perform the following process. That is, the user adds a new combination of the automatically generated control characteristic 1251 and the sensory evaluation value perceived by the user to the learning data of the sensory evaluation point estimating unit 101 and the second sensory evaluation point estimating unit 1202. This improves the estimation accuracy of the sensory evaluation point estimating unit 101 and the second sensory evaluation point estimating unit 1202.

[0056] In the above-described embodiments and modifications, the functional block configurations are merely examples. Some functional configurations shown as separate functional blocks may be integrated, or a configuration shown in a single functional block diagram may be divided into two or more functions. Furthermore, some of the functions of each functional block may be provided by other functional blocks.

[0057] In the above-described embodiments and modifications, the program is stored in the ROM 42, but the program may be stored in the storage unit 46. Furthermore, the control characteristic selection device 1 may be provided with an input / output interface (not shown), and the program may be loaded from another device as needed via the input / output interface and a medium available to the control characteristic selection device 1. Here, the medium refers to, for example, a storage medium detachable from the input / output interface, or a communication medium, i.e., a wired, wireless, or optical network, or a carrier wave or digital signal propagating through the network. Furthermore, some or all of the functions realized by the program may be realized by a hardware circuit or an FPGA.

[0058] The above-described embodiments and modifications may be combined with each other. Although various embodiments and modifications have been described above, the present invention is not limited to these. Other embodiments conceivable within the scope of the technical concept of the present invention are also included within the scope of the present invention.

[0059] 1: Characteristic selection device 44: Input / output device 45: Communication device 51: Current characteristic 52: Current evaluation score 53: Target value 54: Selected characteristic 58: Presentation number command 59: Mode command 101: Sensory evaluation score estimation unit 102: Database 103: Setting unit 104: Control characteristic selection unit 105: Input unit 106: Display unit 502: Evaluation score 703: Control unit 704: Reading unit 1200: Characteristic addition device 1201: Characteristic data generation unit 1203: Database management unit

Claims

1. A control characteristic selection device that searches for control characteristics of an actuator that contributes to the driving characteristics of a moving body, comprising: a database in which multiple combinations of the control characteristics and sensory evaluation points are stored; a sensory evaluation point estimation unit that inputs the control characteristics as current characteristics and outputs the sensory evaluation point as the current evaluation point; an input unit that accepts a target value which is the sensory evaluation point to be targeted; and a selection unit that selects, from the control characteristics stored in the database, a control characteristic whose sensory evaluation point is similar to the target value and is similar to the current characteristic as a selected characteristic.

2. A control characteristic selection system as described in claim 1, wherein the selection unit selects as the selected characteristic a control characteristic whose sensory evaluation score is similar to the target value and whose difference from the current characteristic is small.

3. A control characteristic selection system as described in claim 1, wherein the sensory evaluation score includes a plurality of evaluation items including a representative item, and the selection unit selects as the selected characteristic the control characteristic for which the representative item of the sensory evaluation score is similar to the target value and for which the sum of the changes in the sensory evaluation scores other than the representative item is the smallest.

4. A control characteristic selection system as described in claim 1, wherein the sensory evaluation score includes a plurality of evaluation items including a representative item, and the selection unit selects as the selected characteristic the control characteristic for which the representative item of the sensory evaluation score exceeds the target value and the value of the sensory evaluation score of the representative item is closest to the target value.

5. A control characteristic selection system as described in claim 1, wherein the control characteristic indicates a second physical quantity corresponding to a first physical quantity, and the sensory evaluation point estimating unit identifies, as a parameter, the value of the second physical quantity when the first physical quantity is increased by a predetermined value unit in order to evaluate the current characteristic, and uses the parameter as an input to a pre-trained neural network.

6. A control characteristic selection system according to claim 1, wherein the actuator is one of a steering actuator, a suspension actuator, and a brake actuator.

7. A control characteristic selection system as described in claim 1, further comprising a characteristic adding device which adds to the database a combination of an automatically generated control characteristic, which is the control characteristic that has been automatically generated, and an automatically generated evaluation value, which is the sensory evaluation point obtained by evaluating the automatically generated control characteristic.

8. A control characteristic selection method executed by a computer capable of searching a database in which multiple combinations of control characteristics of actuators that contribute to the running characteristics of a moving body and sensory evaluation points of the control characteristics are stored, the control characteristic selection method including: a sensory evaluation point process in which the control characteristic is input as a current characteristic and the sensory evaluation point is output as the current evaluation point; an input process in which a target value which is the sensory evaluation point to be targeted is received; and a selection process in which a control characteristic whose sensory evaluation point is similar to the target value and similar to the current characteristic is selected as a selected characteristic from the control characteristics stored in the database, the selection process selecting as the selected characteristic a control characteristic that satisfies the target value and has a small difference in characteristic value from the current characteristic.

Images