Vehicle control device

Abstract

The present invention addresses the problem where multiple control modes for an on-vehicle actuator such as comfort, normal, and sports modes are available as functionalities to suit preferences of a driver but can only be switched manually by the driver and have been underutilized. A vehicle control device according to the present invention has a travel situation determination unit 109 and a driving skill determination unit 110. Control modes of on-vehicle actuators 102-104, which have hitherto been manually switched by a driver, are automatically switched on the basis of determination results regarding the driver's driving skill and the travel situation. As a result, the vehicle control device provides ride comfort and steering stability that are highly satisfactory for the driver.

Description

Vehicle control device

[0001] The present invention relates to a vehicle control device.

[0002] 2. Description of the Related Art A suspension control device is known that controls the attitude of a vehicle by changing the stiffness and characteristics of the suspension in accordance with road conditions and driving conditions.

[0003] Patent Document 1 describes a technology for a device that has a sport mode switch, a snow mode switch, and an eco mode switch for switching the vehicle's driving mode, sets the driving mode according to the switching state of these driving mode switching means, and transfers a control command signal including the driving mode to a suspension or an ECU that controls the suspension, with the aim of achieving a desired ride comfort. In an actual vehicle, the driver operates a dial to switch the driving mode when faced with driving in a city mainly at low speeds, driving on a highway or expressway mainly at high speeds, or driving on a mountain pass where quick steering and frequent acceleration and deceleration may be preferred.

[0004] Patent Document 2 discloses a technique for determining the vehicle's driving conditions based on information about the on / off status of an auto-cruise switch and the position of a shift lever, and for changing the suspension characteristics based on the determination results.

[0005] JP 2008-132876 A JP 2013-116641 A

[0006] The device of Patent Document 1 only allows the driver to manually switch the control mode, and does not necessarily achieve both a suitable ride comfort and driving stability, and this function cannot be fully utilized.

[0007] Furthermore, the device in Patent Document 2 shows that the suspension characteristics can be changed according to the driver's operating conditions, even without the driver having any particular intention. However, the only information that is referred to is the driver's operating conditions such as the shift lever, and it is not necessarily possible to select a control mode based on the driving situation.

[0008] The vehicle control device according to the present invention is a vehicle control device capable of setting a plurality of control modes for an actuator mounted on a vehicle, and is characterized by comprising an optimal control mode estimation unit that estimates an optimal control mode from among the plurality of control modes using behavior information of the vehicle, learning parameters obtained by machine learning of the correlation between the behavior information and the plurality of control modes, and information on the driving skill of a driver who drives the vehicle.

[0009] According to the present invention, it is possible to control on-board actuators, such as steering, suspension, and motors, according to the driving situation, thereby improving the ride comfort and handling stability of the vehicle and increasing driver satisfaction. Further features related to the present invention will become apparent from the description of this specification and the accompanying drawings. Furthermore, problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments.

[0010] 1 is a block configuration diagram of a vehicle in the automatically adaptable vehicle system according to the first embodiment of the present invention. FIG. 2 is a diagram showing the relationship between input and output waveforms of an optimal control mode estimation unit 106 in the automatically adaptable vehicle system according to the first embodiment of the present invention. FIG. 3 is a diagram showing an example of a driving situation determination unit 109 according to the first embodiment of the present invention, showing a case where a neural network 300 is applied, and showing the configuration of the neural network 300. FIG. 4 is a diagram showing an example of the driving situation determination unit 109 according to the first embodiment of the present invention, showing a truth table showing the relationship between values ​​of output elements and corresponding control modes. FIG. 5 is a block configuration diagram of the automatically adaptable vehicle system according to the first embodiment of the present invention. FIG. 6 is a diagram showing a flowchart illustrating control processing of a test vehicle in the automatically adaptable vehicle system according to the first embodiment of the present invention. FIG. 7 is a diagram showing a flowchart illustrating control processing of a server in the automatically adaptable vehicle system according to the first embodiment of the present invention. FIG. 8 is a diagram showing a flowchart illustrating control processing of a vehicle in the automatically adaptable vehicle system according to the first embodiment of the present invention. 1 is a diagram showing an example of measuring driving skill by the driving skill determination unit 110 in the automatically adaptable vehicle system of the first embodiment of the present invention. FIG. 2 is a diagram showing a flowchart illustrating processing for determining driving skill by the driving skill determination unit 110 in the automatically adaptable vehicle system of the first embodiment of the present invention. FIG. 3 is a diagram showing an example of conditions for determining driving skill by the driving skill determination unit 110 in the automatically adaptable vehicle system of the first embodiment of the present invention. FIG. 4 is a block configuration diagram relating to an example of the driving skill determination unit 110 in the automatically adaptable vehicle system of the first embodiment of the present invention, focusing on vehicle speed. FIG. 5 is a diagram showing an example of measuring driving skill by the driving skill determination unit 110 in the automatically adaptable vehicle system of the first embodiment of the present invention. FIG. 6 is a diagram relating to an example of the driving skill determination unit 110 in the automatically adaptable vehicle system of the first embodiment of the present invention, showing a case where a neural network is applied. FIG. 7 is a truth table showing the relationship between values ​​of output elements of the driving skill determination unit 110 and corresponding driving skills in the automatically adaptable vehicle system of the first embodiment of the present invention.1 is a block configuration diagram when a control mode is determined in a control mode setting unit 111 of an automatically adapting vehicle system according to a first embodiment of the present invention. FIG. 2 is a block configuration diagram when a control mode is determined by adjusting a signal value in response to behavior information arriving from a CAN in an optimal control mode estimation unit 106 according to a first embodiment of the present invention. FIG. 3 is a diagram showing variations of control mode switching according to a first embodiment of the present invention. FIG. 4 is a truth table showing the relationship between the value of an output element for control mode switching and the corresponding control mode according to a first embodiment of the present invention. FIG. 4 is a diagram showing an example of a case where information is presented to a driver regarding the output of a control mode setting unit 111 according to a first embodiment of the present invention. FIG. 5 is a block configuration diagram of a vehicle in an automatically adapting vehicle system according to a second embodiment of the present invention. FIG. 5 is a diagram relating to an example of processing by an optimal control mode estimation unit 106 in an automatically adapting vehicle system according to a second embodiment of the present invention, where the processing is realized by a neural network. FIG. 6 is a diagram relating to an example of processing by an optimal control mode estimation unit 106 in an automatically adapting vehicle system according to a second embodiment of the present invention, where the processing is realized by a neural network. FIG. 10 is a block configuration diagram of a vehicle in an automatically adapting vehicle system according to a third embodiment of the present invention. FIG. 11 is a table showing the correspondence between input modes such as family mode and switching of control modes in a driving situation determination unit of an automatically adapting vehicle system according to a third embodiment of the present invention. FIG. 12 is a diagram showing an example in which a driver adjusts the specifications of a driving situation determination unit using an information terminal such as a smartphone for the driving situation determination unit of an automatically adapting vehicle system according to the third embodiment of the present invention. FIG. 13 is a diagram showing an example of a specification table and a specification selection screen on an information terminal for the driving situation determination unit of an automatically adapting vehicle system according to the third embodiment of the present invention. FIG. 14 is a block configuration diagram when a control mode is determined in a driving skill determination unit 110 of an automatically adapting vehicle system according to a fourth embodiment of the present invention.

[0011] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The following description and drawings are examples for explaining the present invention, and some omissions and simplifications have been made as appropriate for clarity of explanation. The present invention can be implemented in various other forms. Unless otherwise specified, each component may be singular or plural.

[0012] When there are multiple components with the same or similar functions, they may be described using the same reference numeral with different subscripts. However, when there is no need to distinguish between these multiple components, the subscripts may be omitted.

[0013] Furthermore, in the following description, processing performed by executing a program may be described. However, the program is executed by a processor (e.g., a CPU (Central Processing Unit), a GPU (Graphics Processing Unit)) to perform a predetermined process while appropriately using storage resources (e.g., memory) and / or interface devices (e.g., communication ports), and therefore the processor may be the subject of the processing. Similarly, the subject of the processing performed by executing a program may be a controller, device, system, computer, or node having a processor. The subject of the processing performed by executing a program may be any computing unit, and may include a dedicated circuit (e.g., an FPGA (Field Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit)) that performs specific processing.

[0014] A program may be installed on a device such as a computer from a program source. The program source may be, for example, a program distribution server or a computer-readable storage medium. When the program source is a program distribution server, the program distribution server may include a processor and storage resources for storing the program to be distributed, and the processor of the program distribution server may distribute the program to be distributed to other computers. Also, in the following description, two or more programs may be realized as one program, or one program may be realized as two or more programs.

[0015] [First embodiment] Vehicle driving situations, such as in town, on a highway, or on a mountain pass, are characterized by driver operation information and the resulting vehicle behavior, such as 6DoF (6 Degree of Freedom) information, and all or part of this information is transmitted over a CAN (Controller Area Network). By referring to this information, it was thought that a control mode that was previously manually selected by the driver could be automatically switched.

[0016] Specific examples of physical quantities to be referenced include longitudinal acceleration, lateral acceleration, and vehicle speed, and the optimal control mode estimator realized in this embodiment estimates a suitable control mode corresponding to the driving situation. Information on the control mode estimated by the optimal control mode estimator is transferred to an integrated ECU or an ECU provided for each actuator, and control according to the control mode is performed on various actuators such as steering, suspension, and motors. As a result, suitable actuator control corresponding to the driving situation described above is performed, and good ride comfort and handling stability that satisfy the driver can be achieved.

[0017] However, if the driver's driving skill is low, the driver may not be able to respond to the automatic switching of the control mode, which may cause confusion. Therefore, in this embodiment, the driving skill is taken into consideration when automatically switching the control mode. Briefly, the driving skill is determined by comparing the target vehicle behavior with the actual vehicle behavior, taking into consideration the vehicle's position information via the CAN and road shape and traffic regulations based on map information. Furthermore, the automatic control mode switching function is visualized to improve driver satisfaction, allowing the driver to enjoy the vehicle's responsiveness in response to the driver's operations.

[0018] Next, a first embodiment of the present invention will be described with reference to Figures 1 to 11. Figure 1 is a block diagram of a vehicle in an automatically adaptable vehicle system according to the first embodiment of the present invention. A vehicle 101 has a steering control 102, a suspension control 103, a motor control 104, a CAN 105, an optimal control mode estimation unit 106, a parameter storage unit 107, a switch 108, a map storage unit 112, a display unit 113, and a driver data storage unit 114.

[0019] The optimal control mode estimation unit 106 includes a driving situation determination unit 109, a driving skill determination unit 110, and a control mode setting unit 111. Each of the driving situation determination unit 109, the driving skill determination unit 110, and the control mode setting unit 111 performs calculations and processes based on the parameters stored in the parameter storage unit 107.

[0020] The optimal control mode estimation unit 106 then acquires data relating to the vehicle behavior during driving (hereinafter referred to as behavior information) from the communication data transmitted via the CAN 105, calculates a suitable control mode based on this behavior information and the parameters in the parameter storage unit 107, and transfers it via the switch 108 to each on-board actuator such as the steering control 102, suspension control 103, and motor control 104 in the subsequent stage.

[0021] The driving situation determination unit 109 receives behavior information transmitted via the CAN 105, such as longitudinal acceleration, lateral acceleration, and vehicle speed, and outputs the optimum control mode from among a plurality of preset control modes.

[0022] The driving skill determination unit 110 receives behavior information, such as vehicle speed, steering angle, and position information, transmitted via the CAN 105, road shape information of the driving surrounding area based on the vehicle position information from the map storage unit 112, and the driver ID, and outputs a driving skill level. A set of the driver ID and the driving skill level is stored in the driver data storage unit 114, and when the same driver ID is input on the next or subsequent driving occasions, the corresponding driving skill level is read out and transferred to the control mode setting unit 111.

[0023] The control mode setting unit 111 adjusts the control mode output by the driving situation determination unit 109 based on the output of the driving skill determination unit 110, as will be described in detail later.

[0024] The switch 108 makes it possible to switch between transferring the output of the optimum control mode estimation unit 106 of this embodiment to various actuators in the subsequent stage and manually setting the control mode by a conventional driver, thereby enabling coexistence with conventional products and also allowing the driver's intention of whether or not to select the control mode to be reflected in the automatic control mode switching function by the optimum control mode estimation unit 106 of this embodiment.

[0025] FIG. 2 is a diagram showing the relationship between input and output waveforms of the optimal control mode estimator 106 according to the first embodiment of the present invention.

[0026] 2, reference numeral 201 denotes an example of a waveform of the longitudinal acceleration of the vehicle, 202 denotes an example of a waveform of the lateral acceleration, 203 denotes an example of a waveform of the vehicle speed, 204 denotes an example of a control mode switching waveform, 205 denotes a window, and 206 denotes an example of an output control mode value. In this embodiment, the control mode value 206 has three modes: normal mode, sport mode, and comfort mode.

[0027] The optimal control mode estimation unit 106 refers to behavior information received from the CAN 105, such as longitudinal acceleration 201, lateral acceleration 202, and vehicle speed 203, and selects and outputs a control mode value 206 that is deemed optimal from among a plurality of control modes. Here, the time range over which the optimal control mode estimation unit 106 refers to the behavior information is defined by a window 205. The window 205 is set by going back a certain amount of time from a target time at which the control mode is selected as the starting point. The information referred to by the control mode setting unit 111 is time-series data for a predetermined period (within the window 205) or vehicle behavior events included in time-series data for a predetermined traveling course.

[0028] 3A and 3B are diagrams showing an example of the driving situation determination unit 109 according to the first embodiment of the present invention, in which a neural network 300 is applied. Fig. 3A shows the configuration of the neural network, and Fig. 3B shows a truth table showing the relationship between the values ​​of the output elements and the corresponding control modes.

[0029] 3, the input layer element group 301 is assumed to refer to three systems of behavior information arriving from the CAN 105, with 301a being an element group that inputs longitudinal acceleration, 301b being an element group that inputs lateral acceleration, and 301c being an element group that inputs vehicle speed. 302 is a hidden layer element group, and 303 is an output layer element group.

[0030] As shown, the longitudinal acceleration element group 301a, the lateral acceleration element group 301b, and the vehicle speed element group 301c each have a plurality of elements, and a plurality of data items are set in the window 205. The behavior information transmitted over the CAN 105 is updated, for example, every 10 msec. If the time specified by the window 205 is 10 seconds, the number of data points included during that time will be 1,000 (= 10 sec ÷ 10 msec), and the number of elements in 301a-c will be 1,000 each. Therefore, when referencing three systems of behavior information, the total number of input layer elements will be 3,000 (= 1,000 × 3 systems), which increases the circuit size and the calculation load. To avoid this, the behavior information arriving from the CAN 105 can be thinned out at regular intervals, for example, by capturing it every 1 second, and the number of points included in the window 205 can be reduced to 10 (= 10 sec ÷ 1 sec). The behavior information thus acquired via CAN 105 can be set in the input layer element group 301a-c of the neural network 300, for example, by installing buffers in the front stage of the neural network 300, the number of which is equal to the number of input elements of the input layer element group 301a-c.

[0031] FIG. 3A shows the configuration of a so-called simple perceptron, but since the behavior information handled in the present invention via the CAN 105 is time-series data, a recurrent neural network RNN ​​(Recurrent Neural Network), which is good at handling time-series data, is an effective configuration.

[0032] In Figure 3A, the number of elements in the output layer element group 303 is four, but this setting is an example when there are three control modes. Furthermore, the truth table 304 in Figure 3B shows the relationship between the value of the output element and the corresponding control mode, with three of the four elements corresponding to three control modes. The remaining element is associated with a transitional state of control mode switching or a state that does not belong to any of the three control modes, and in this embodiment, the mode is maintained.

[0033] 3B, any one of the output elements of the output layer element group 303 outputs "1" (high), and the others output "0" (low). When the comfort mode is selected, only Y1 outputs "1" (high), and the others Y2 to Y4 output "0" (low). When the normal mode is selected, only Y2 outputs "1" (high), and the others Y1, Y3, and Y4 output "0" (low). When the sports mode is selected, only Y3 outputs "1" (high), and the others Y1, Y2, and Y4 output "0" (low). When the mode maintenance, which inherits the previous setting, is selected, only Y4 outputs "1" (high), and the others Y1 to Y3 output "0" (low).

[0034] Although longitudinal acceleration, lateral acceleration, and vehicle speed have been given as examples of behavior information to be used, the present invention is not limited to these examples as long as they are physical values ​​that are effective for automatically switching the target control mode. Furthermore, although the present invention gives the names of the multiple control modes as comfort mode, normal mode, and sport mode, the present invention is not limited to these examples as long as they change the control mode of each actuator mounted on the vehicle.

[0035] FIG. 4 is a block diagram showing an implementation configuration for creating parameters for the driving situation determination unit 109 to be stored in the parameter storage unit 107, and is an automatic adaptive vehicle system according to the first embodiment of the present invention.

[0036] The automatically adapting vehicle system of this embodiment includes a vehicle 101, a server 401, and a test vehicle 402, which are connected to each other via a public line 403 through their respective transmission / reception I / Fs 404, 409, and 412. The server 401 includes a control unit 405, a data set storage unit 406, a learning unit 407, and a weight parameter storage unit 408. The test vehicle 402 includes a driving log storage unit 410 and a driver response storage unit 411.

[0037] The vehicle 101 has the configuration shown in Fig. 1, and an optimal control mode estimation unit 106 including a driving situation determination unit 109 automatically switches the control mode based on behavior information transmitted to a CAN 105. Fig. 4 shows that there are a plurality of vehicles 101 incorporating the embodiment of the present invention. There may also be a plurality of test vehicles 402.

[0038] The server 401 acquires, via the public line 403, the driving data from the CAN 105 stored in the driving log storage unit 410 of the test vehicle 402 and the preferred control mode answered by the test driver during driving stored in the driver response storage unit 411. The server 401 acquires the data set of the driving log and the preferred control mode answered by the driver received from the test vehicle 402 via the transmission / reception IF 404, and temporarily stores them in the data set storage unit 406 via the control unit 405.

[0039] The control unit 405 then reads out the data set of behavior information and control mode from the data stored in the data set storage unit 406 and transfers it to the learning unit 407. The learning unit 407 is provided with a neural network learning environment, sets the behavior information on the input side of the neural network, sets the control mode on the output side of the neural network, and performs correlation learning using a method such as back propagation, which is a commonly known learning method. The weight parameters of the neural network acquired in this way are stored in the weight parameter storage unit 408 via the control unit 405. The weight parameters acquired here are distributed by the control unit 405 to the vehicle 101 via the public line 403.

[0040] The vehicle 101 acquires the weight parameters received from the server 401 via the public line 403 and the transmission / reception IF 412, and stores them in the parameter storage unit 107. As described above, the weight parameters are applied to the driving situation determination unit 109 included in the optimal control mode estimation unit 106, and the control mode is updated every time behavior information during driving is input from the CAN 105.

[0041] FIG. 5 is a flowchart according to the first embodiment of the present invention, and FIGS. 5A to 5C are flowcharts for the test vehicle 402, server 401, and vehicle 101 shown in FIG. 4, respectively. In FIG. 5A, 501 indicates acquisition of driving data and driver responses, and 502 indicates server transfer. In FIG. 5B, 503 indicates reception of driving data and driver responses, 504 indicates storage of driving data and driver responses, 505 indicates data set readout, 506 indicates correlation learning, 507 indicates storage of weight parameters, 508 indicates readout of weight parameters, and 509 indicates distribution of weight parameters. In FIG. 5C, 510 indicates reception of weight parameters, 511 indicates storage of weight parameters, 512 indicates readout of weight parameters, 513 indicates trial of new parameters, 514 indicates error determination, 515 indicates application of new parameters, 516 indicates resumption of automatic adaptation, 517 indicates application of old parameters, 518 indicates error reporting, and 519 indicates deletion of new parameters.

[0042] In the test vehicle 402, driving data and driver response acquisition 501 and server transfer 502 are performed. In driving data and driver response acquisition 501, a model driver, a driver to be emulated, or a so-called expert driver drives in simulated urban areas, mountain roads, etc., and the driver is asked to indicate the control mode that the driver considers appropriate for each driving situation and driving operation phase, such as comfort mode when driving in urban areas or on general roads at a constant speed, or sport mode when driving on mountain roads or when accelerating, decelerating, or steering relatively frequently. In addition, CAN data during driving is stored in the corresponding driver response storage unit 411 and driving log storage unit 410. These data sets are managed by linking them with time data, and once a certain amount of data sets have been stored, they are transferred to the server 401 via the transmission / reception IF 409, public line 403, and transmission / reception IF 404 (server transfer 502).

[0043] The server 401 performs processes from receiving driving data and driver responses 503 to distributing weight parameters 509. The receiving driving data and driver responses 503 receives and acquires a data set arriving from the test vehicle 402, and the storing driving data and driver responses 504 stores the acquired data set in the data set storage unit 406. Then, in reading data set 505, the learning unit 407 reads out the behavior information to be used and the control mode answered by the driver from the data set storage unit 406.

[0044] In correlation learning 506, so-called machine learning is performed using the group of data sets read by the learning unit 407. The learning method may be, for example, backpropagation. In weight parameter storage 507, weight parameters of the neural network resulting from the machine learning performed by the learning unit 407 are written to the weight parameter memory unit 408. In weight parameter readout 508, the weight parameters stored in the weight parameter memory unit 408 are read, and in weight parameter distribution 509, the weight parameters are transmitted to the vehicle 101 via the transmission / reception IF 404, the public line 403, and the transmission / reception IF 412.

[0045] In the vehicle 101, steps from weight parameter reception 510 to new parameter deletion 519 are performed. In weight parameter reception 510, weight parameters arriving from the server 401 are received, and in weight parameter storage 511, the weight parameters are written into the parameter storage unit 107.

[0046] In weight parameter reading 512 , the optimum control mode estimation unit 106 reads out the weight parameters from the parameter storage unit 107 and applies them to the driving situation determination unit 109 included in the optimum control mode estimation unit 106 .

[0047] In new parameter trial 513, the weight parameters received from the server 401 are applied inside the vehicle 101 once before they are actually applied, and in error determination 514, it is first checked whether there are any basic errors and whether the driving situation determination unit 109 is outputting appropriately. If there are no problems, the latest parameters transferred from the server 401 are applied in new parameter application 515, and automatic switching of the control mode is realized in automatic adaptation restart 516. If there are problems, the parameters that were used until the weight parameters were transferred from the server 401 are applied in old parameter application 517, and automatic switching of the control mode is realized in automatic adaptation restart 516.

[0048] If a problem is found in error determination 514, this is reported to server 401 in error report 518, and the weight parameters received from server 401 are deleted in new parameter deletion 519 to prevent overloading of the storage capacity in vehicle 101. The flowchart in Figure 5 is described on the assumption that vehicle 101 already has weight parameters, but the initial parameter settings are also applied to vehicle 101 via weight parameter reception 510 and new parameter trial 513, and if it is initial, the initial values ​​may be written to the ECU without going through public line 403.

[0049] 6A and 6B are diagrams showing an example of processing by the driving skill determination unit 110 according to the first embodiment of the present invention, focusing in particular on the steering angle. Fig. 6A is a block diagram, Fig. 6B is a diagram showing an example of measuring the driving skill, Fig. 6C is a flowchart, and Fig. 6D is an example of conditions for determining the driving skill.

[0050] As shown in FIG. 6A, the driving skill determination unit 110 includes a vehicle direction estimation unit 601, a self-position estimation unit 602, a surrounding map acquisition unit 603, a steering angle candidate calculation unit 604, a target steering angle selection unit 605, and a steering angle difference calculation unit 606.

[0051] In the graph shown in Fig. 6B, the vertical axis represents the steering angle of the vehicle, and the horizontal axis represents time. In Fig. 6B, 607 represents the target steering angle, 608 represents the ideal steering angle transition, 609 represents the actual steering angle transition, and 610 represents the steering angle difference.

[0052] 6C is a flowchart illustrating the process of determining the driving skill of the driving skill determination unit 110. Reference numeral 611 denotes driver ID acquisition, 612 denotes ID registration determination, 613 denotes driving skill information acquisition, 614 denotes beginner setting, 615 denotes initialization, 616 denotes road / driving state detection, 617 denotes intersection presence / absence determination, 618 denotes target steering angle candidate calculation, 619 denotes steering angle information acquisition, 620 denotes right / left turn operation completion determination, 621 denotes steering angle difference D calculation, 622 denotes error integration, 623 denotes count-up, 624 denotes driving process completion determination, 625 denotes average error calculation, 626 denotes driving skill determination, and 627 denotes driver data update (addition). The steering angle information acquired in steering angle information acquisition 619 is a vehicle behavior event included in the time-series data of a predetermined driving process.

[0053] The vehicle direction estimation unit 601 and the vehicle's own position estimation unit 602 estimate the vehicle's direction and position by referring to GPS information and the like transmitted over the CAN 105. The surrounding map acquisition unit 603 refers to the map storage unit 112 to identify the road shape around the traveling vehicle 101 and the actual traveling lane. The steering angle candidate calculation unit 604 uses the presence of an intersection or side road ahead of the traveling lane as a trigger to derive an angle for each road on which the vehicle may turn right or left, and lists candidate target steering angles. The target steering angle selection unit 605 receives actual steering information transmitted over the CAN 105 and selects one target steering angle from the listed candidate steering angles. The steering angle difference calculation unit 606 compares an ideal steering angle transition 608 from the pre-steering state to the target steering angle 607 with an actual steering angle transition 609. The steering angle difference calculation unit 606 calculates some quantitative value, for example, a steering angle difference 610 between an ideal steering angle transition 608 and an actual steering angle transition 609 .

[0054] FIG. 6C shows a manner in which the steering angle difference 610 is derived in a single driving situation, but the driving skill does not have to be determined based on only one driving situation.

[0055] 6D is a diagram showing an example of conditions for determining the driving skill of the driving skill determination unit 110. An example is shown in the specification table of FIG. 6D. The steering angle difference 610 is accumulated for each right or left turn during a driving trip on a certain day, and if the accumulated value is equal to or less than a certain value x, the driver is determined to be an "expert driver," and the driving skill is treated as an "expert driver" in the next trip from the next day onwards. If the accumulated value is greater than the certain value x and equal to or less than the certain value X, the driver is determined to be an "average driver," and the driving skill is treated as an "average driver" in the next trip from the next day onwards. If the accumulated value is greater than the certain value X, the driver is determined to be a "novice driver," and the driving skill is treated as a "novice driver" in the next trip from the next day onwards.

[0056] In this embodiment, an example of calculating driving skill under driving conditions of turning right or left is shown, but it is also possible to assume a case where a road transitions from a straight road to a curved road and compare the target steering angle with the actual steering angle in relation to the curvature of the curve.

[0057] FIG. 6C is a flowchart relating to the driving skill determination unit 110, and particularly shows the process of management and driving skill determination based on the driver ID.

[0058] Driver ID acquisition 611 is a process for acquiring driver information about the driver who is driving, and this information may be managed by the ignition key, input by the driver, or obtained from driver recognition results from various sensors. ID registration determination 612 checks whether a driver ID for managing the results of the driving skill determination performed in the present invention is already present in the driver data storage unit 114. If an ID is present, driving skill information acquisition 613 refers to driving skill information, which is a previous determination result also managed in the driver data storage unit 114. If an ID is not present, it is determined that the driver's driving skill is unknown based on a certain standard, and the driving skill is initially set to beginner in beginner setting 614. Hereinafter, the description will be given assuming that the driving skill determination is performed regardless of the result of ID registration determination 612.

[0059] First, the parameters N and Diff used in the calculation of the driving skill determination are set to "0" in initialization 615. The driving skill determination is performed for each driving process, for example, from turning the ignition on to turning it off, and initialization 615 is performed after the driver holding the steering wheel is identified by the driver ID.

[0060] Road / driving state detection 616 acquires GPS position information from CAN 105 and map information including road shapes from map storage unit 112, and checks which road / lane on the map the vehicle 101 is traveling on and in which direction. Then, intersection presence / absence determination 617 checks whether there is a fork ahead where the vehicle 101 may turn right or left.

[0061] If there is a branch where there is a possibility of turning right or left, i.e., an intersection, steering angles (θ1, θ2, θ3) are derived for the number of branches, for example, for three branches excluding the road currently being traveled, in target steering angle candidate calculation 618. After that, when the road on which the vehicle 101 will proceed based on the actual steering by the driver can be determined in steering angle information acquisition 619, θx, one of the multiple steering angles listed in target steering angle candidate calculation 618, is adopted, and it is determined in right or left turn operation completion determination 620 whether the vehicle 101 has reached the target steering angle.

[0062] If the result of the right / left turn operation completion determination 620 is NO, the process returns to the steering angle information acquisition 619 to continue the determination, and if the result of the right / left turn operation completion determination 620 is YES, the process calculates D as the difference between the target steering angle and the actual steering angle in the steering angle difference D calculation 621. Then, the accumulation of the steering angle difference D for each steering event is calculated as Diff in the error integration 622, and the number of steering events in one traveling process is also counted in the count up 623.

[0063] In the travel process end determination 624, the end of one travel process is determined, for example, by turning off the ignition, and if the travel is to continue, the process returns to the road and travel condition detection 616, and if the travel is to end, the process proceeds to the average error calculation 625. In the average error calculation 625, the average value of the steering angle difference D is calculated, and specifically, Diff / N is calculated.

[0064] Driving skill determination 626, the details of which will be described later, determines the driving skill of the driver based on the calculation result of average error calculation 625. Finally, driver data update (addition) 627 writes the determination result of driving skill determination 626 in the driver data storage unit 114 in association with the driver ID. If the driver ID has already been registered, it is an update, and if the driver ID is not registered, it is an addition of data.

[0065] 6D shows an example of conditions for determining driving skill, where driving skill is defined as three levels: expert, average, and beginner. Two levels, x and X, are defined as boundary conditions, and a determination is made by comparing the average error Diff / N calculated in the average error calculation 625 (x<X). For example, if the average error Diff / N is equal to or less than x, the level is set to 1 (expert); if the average error Diff / N is greater than X, the level is set to 3 (beginner); and if the average error Diff / N is greater than x but equal to or less than X, the level is set to 2 (average).

[0066] While the driving skill is determined using the determination conditions shown in FIG. 6D here, other methods may be used as long as they can measure the driving skill in a similar manner. For example, the number of oversteers or understeers that occurred during one driving process may be counted and the driving skill may be determined based on whether or not a certain number of oversteers or understeers occurred during one driving process. The aforementioned certain number of oversteers or understeers may be determined in advance, or may be set as a dynamically changing value obtained by performing statistical processing with reference to separately acquired driving log data. Furthermore, although the number of driving skill levels has been described as three, in the automatic control mode switching that is the object of the present invention, the number of levels is set according to the switching mode, and is not limited to three.

[0067] 7A and 7B are diagrams showing an example of the processing of the driving skill determination unit 110 according to the first embodiment of the present invention, focusing particularly on the vehicle speed. Fig. 7A is a block diagram, and Fig. 7B is a diagram showing an example of measuring the driving skill.

[0068] As shown in FIG. 7A, the driving skill determination unit 110 includes a vehicle direction estimation unit 601, a self-position estimation unit 602, a surrounding area map acquisition unit 603, a speed limit information extraction unit 701, a congestion information acquisition unit 702, and a vehicle speed difference calculation unit 703.

[0069] In the graph shown in Fig. 7B, the vertical axis represents vehicle speed, and the horizontal axis represents time. In Fig. 7B, reference numeral 704 denotes a target vehicle speed, 705 denotes an ideal vehicle speed transition, 706 denotes an actual vehicle speed transition A, 707 denotes a vehicle speed difference A, 708 denotes an actual vehicle speed transition B, and 709 denotes a vehicle speed difference B.

[0070] The vehicle direction estimation unit 601 and the self-position estimation unit 602 refer to GPS information and the like transmitted over the CAN 105, and the surrounding map acquisition unit 603 refers to the map storage unit 112. The speed limit information extraction unit 701 extracts the road shape around the traveling vehicle 101, the lane on which the vehicle is actually traveling, and the speed limit information for that lane.

[0071] Furthermore, in the case of Japan, the congestion information acquisition unit 702 checks whether or not there is congestion by referring to VICS (registered trademark: Vehicle Information and Communication System). The vehicle speed difference calculation unit 703 interprets the speed limit transferred from the speed limit information extraction unit 701 as a target vehicle speed 704, and compares an ideal vehicle speed transition 705 up to the target vehicle speed 704 with an actual vehicle speed transition A 706 and an actual vehicle speed transition B 708 from the vehicle speed information transmitted on the CAN 105.

[0072] The vehicle speed difference calculation unit 703 calculates the vehicle speed difference between the speed limit information and the vehicle speed information. The vehicle speed difference calculation unit 703 calculates some quantitative value, such as the vehicle speed difference A707 or the vehicle speed difference B709. The driving skill determination unit 110 determines the driving skill based on the vehicle speed difference. While FIG. 7B shows a manner in which the vehicle speed difference A707 or the vehicle speed difference B709 is derived in a single driving phase, the driving skill does not need to be determined based on only one driving phase. As with the steering angle example in FIG. 6B , the vehicle speed difference A707 or the vehicle speed difference B709 is accumulated for each acceleration during a single driving process, and the value of the average error Diff / N derived from the accumulated value Diff is used to determine whether the driver is an expert driver, an average driver, or a novice driver.

[0073] Here, the driving skill is determined by calculating the average error Diff / N, but other methods may be used as long as they can measure the driving skill in a similar manner. For example, the number of sudden accelerations and decelerations that occurred during one driving process, that is, the number of times that the absolute acceleration exceeded a certain value, may be counted, and the driving skill may be determined based on whether or not the certain number of times has been exceeded. Furthermore, the above-mentioned absolute acceleration and certain number of times may be determined in advance, or may be set as dynamically changing values ​​obtained by performing statistical processing with reference to separately acquired driving log data.

[0074] 8A and 8B are diagrams showing an example of the processing of the driving skill determination unit 110 according to the first embodiment of the present invention, particularly when the processing is realized by a neural network. While Fig. 8A shows the configuration of a so-called simple perceptron, since the behavior information handled in the present invention via the CAN 105 is time-series data, a recurrent neural network (RNN), which is good at handling time-series data, is an effective configuration.

[0075] 8B is a truth table showing the relationship between the value of an output element and the corresponding driving skill. In FIG. 8, 801 is an input layer element group, 802 is a hidden layer element group, 803 is an output layer element group, and 804 is a truth table. This is one example of a method for implementing the driving skill determination unit 110. The input to the neural network is assumed to be time-series data of the driver's operation, and physical quantities such as the accelerator 801a, brake 802b, and steering angle 803c, which is the steering operation, are input. It is believed that these physical quantities can represent the transient state of the driver's operation and can determine the quality of the driving skill.

[0076] When driving skill determination is realized using a neural network, there is no need to express the determination rules as processing processes as shown in Figures 6A and 7A, and it is sufficient to prepare a plurality of drivers categorized into expert drivers, average drivers, novice drivers, etc., and learn the correlation between the actual driving data of each driver and the category of driving skill. Since it is similar to the driving situation determination unit 109 in that it handles time-series data, learning can be performed by the server 401.

[0077] Truth table 804 is a table showing the relationship between output Y and the skill judgment result when judging driving skill using a neural network. As with truth table 304 (FIG. 3B) of driving situation judgment unit 109, an output element is assigned to each driving skill category, and learning is performed so that only the output element for the corresponding driving skill category becomes 1 and the other output elements become 0.

[0078] Although the accelerator, brake, and steering angle have been given as examples of operation information to be referenced, the present invention is not limited to these, and any physical value that is effective for determining the target driving skill may be used.

[0079] Fig. 9A is a diagram showing an example of processing by the control mode setting unit 111 according to the first embodiment of the present invention. Reference numeral 901 in Fig. 9A is a diagram illustrating processing content in the control mode setting unit 111. Fig. 9A is a block configuration diagram for determining an optimal control mode by correcting the output of the driving situation determination unit 109 based particularly on the driving skill that is the determination result of the driving skill determination unit 110.

[0080] 6A, 7A, or 8, or a combination thereof, and the control mode setting unit 111 receives the output a of the driving situation determination unit 109 and the output b of the driving skill determination unit 110. Then, in accordance with the processing content 901, the control mode setting unit 111 corrects the output a and outputs a'. If the output b of the driving skill determination unit 110 is 1 (expert), it is determined that the control mode should be switched, and the same value as the output a of the driving situation determination unit is output.

[0081] Furthermore, if the output b is 2 (general), the control mode is one level lower than the output a of the driving situation determination unit 109, for example, if the output a is the sport mode, the normal mode is selected, and if the output a is the normal mode, the comfort mode is selected. Furthermore, if the output b is 3 (beginner), the control mode is not switched on the vehicle 101 side, and the normal mode is fixed. Note that this is just an example, and the number of determination levels for driving skill is not limited to 3, and the type of control mode for each level is not limited to the processing content 901, as long as the switching mode of the control mode is similarly changed depending on the driving skill.

[0082] Next, a method of adjusting a signal value in a stage preceding the driving situation determination unit 109 shown in Fig. 9B will be described. Fig. 9B is a block diagram showing a case where a control mode is determined by adjusting a signal value in response to behavior information received from the CAN.

[0083] The signal adjustment unit 902 corrects the vehicle behavior information according to the determination result of the driving skill determination unit 110. The driving situation determination unit 109 outputs one of a plurality of control modes using the behavior information corrected by the signal adjustment unit 902 and the learning parameters. The sport mode is characterized by its ability to quickly respond to conditions with large acceleration / deceleration changes and large steering amounts, and as a result is a suitable mode when the behavior information, such as longitudinal acceleration, lateral acceleration, or vehicle speed, is large. Therefore, the signal adjustment unit 902 adjusts the longitudinal acceleration, lateral acceleration, and vehicle speed received from the CAN 105 according to processing content 903.

[0084] When the output b of the driving skill determination unit 110 is 1 (expert), the longitudinal acceleration, lateral acceleration, and vehicle speed received from the CAN 105 are transferred to the downstream driving situation determination unit 109. When the output b is 2 (general), the longitudinal acceleration, lateral acceleration, and vehicle speed received from the CAN 105 are multiplied by 0.95, and the resulting values ​​are transferred to the downstream driving situation determination unit 109.

[0085] When the output b is 3 (beginner), the longitudinal acceleration, lateral acceleration, and vehicle speed received from CAN 105 are multiplied by 0.90, and the resulting value is transferred to the downstream driving situation determination unit 109. Note that this is just one example, and as long as the control mode switching manner is similarly varied depending on the driving skill, the number of determination levels for driving skill is not limited to 3, and the manner of the control mode for each level is not limited to the processing content 903.

[0086] In either case, when the output b of the driving skill determination unit 110 is 3 (beginner) or 2 (average), adjustments are made to prevent the vehicle from easily transitioning to sports mode, thereby providing a safer and more stable control mode.

[0087] 10A and 10B are diagrams showing an example of processing by the control mode setting unit 111 according to the first embodiment of the present invention, in which FIG. 10A is a diagram showing variations in control mode switching, and FIG. 10B is a truth table showing the relationship between the value of the output element for control mode switching and the corresponding control mode.

[0088] Reference numeral 1001 in FIG. 10A is a diagram illustrating the relationship between the control modes and the control gains when there are three types of control modes (a sparse mode set), reference numeral 1002 is a diagram illustrating the relationship between the control modes and the control gains when there are ten types of control modes (a dense mode set), and reference numeral 1003 in FIG. 10B is a truth table for the case where there are ten types of control modes. When the driver's driving skill is lower than a standard, the control mode setting unit 111 provides multiple control modes with smaller differences in characteristics between the control mode settings compared to when the driver's driving skill is higher than the standard. The control mode setting unit 111 includes a sparse mode set consisting of multiple control modes with large differences in parameters between the control modes, and a dense mode set consisting of multiple control modes with small differences in parameters between the control modes. When the driving skill determination unit 110 determines that the driver has low skill, the dense mode set is selected. When the driving skill determination unit 110 determines that the driver has high skill, the sparse mode set is selected, or the driver can select either the dense mode set or the sparse mode set.

[0089] For example, if the driver's driving skill is not relatively high, if the actuator characteristics change significantly while the driver is driving, the driver may be surprised, especially if the change is unexpected, and this may cause problems in driving behavior. Considering this, a control mode 1002 (dense mode set) with 10 levels, in which the difference in characteristics between settings is small, is provided, instead of three levels like the control mode 1001 (sparse mode set). In contrast to the method described above in which it is difficult to enter the sport mode based on the determination result of the driving skill determination unit 110, this method is characterized by the fact that the transition to the sport mode is made as a result, but the transitional state up to that point is finely detailed.

[0090] 10A and 10B, an example has been described in which the output of control mode setting unit 111 is a multi-stage value, but the output of control mode setting unit 111 is updated in the same manner every time the behavior information received from CAN 105 is updated. Therefore, it is also possible to calculate a moving average value for the output of control mode setting unit 111 and adjust the control gain with an analog value.

[0091] 11A and 11B are diagrams showing an example of the case where information is presented to the driver on the display unit 113 in relation to the output of the control mode setting unit 111 according to the first embodiment of the present invention. In Fig. 11A, 1101 is the instrument panel of the vehicle 101, 1102 is an information presentation unit, and 1103 is the information presentation content displayed on 1102. In Fig. 11B, 1104 is an information terminal, 1105 is a behavior information waveform, and 1106 is the information presentation content.

[0092] The control mode is presented in an information presentation section 1102 in an instrument panel 1101 that can be seen by the driver while driving, as shown in Fig. 11A, and the presented information 1103 is the driver ID and the set control mode. The information may be presented not only by the instrument panel 1101, but also by an information terminal 1104 that can communicate with the vehicle 101, as shown in Fig. 11B. Examples of the presented information on the information terminal 1104 include a behavior information waveform 1105, and, as in the example of the instrument panel 1101, the presented information 1106 is the driver ID and the set control mode.

[0093] [Second Embodiment] The second embodiment of the present invention is characterized by adding the setting of the driver's driving preferences to the first embodiment, which takes into account the driving situation and driving skill. Basically, the control mode is automatically switched depending on the driving situation, but the second embodiment is characterized by the driver being able to change the switching manner, specifically the switching conditions, and this can be adjusted by setting a driver number. In this embodiment, it is assumed that there are four types of drivers, and by specifying a driver number, it is possible to experience the corresponding control mode switching manner.

[0094] Next, a second embodiment of the present invention will be described with reference to Figures 12A to 12C. Figure 12A is a block diagram of a vehicle in an automatically adapting vehicle system according to the second embodiment of the present invention.

[0095] 12A shows a vehicle 1201, a driving preference setting unit 1202, a parameter storage unit 1203, and a driving situation determination unit 1204. Fig. 12B shows an example of the processing of the optimal control mode estimation unit 106 in the automatically adaptive vehicle system according to the second embodiment of the present invention, and is a diagram showing the case where the processing is realized by a neural network.

[0096] 12B shows an input layer element group 1205, which includes an element group 1205a for inputting longitudinal acceleration, an element group 1205b for inputting lateral acceleration, an element group 1205c for inputting vehicle speed, an element group 1205d for inputting driver number, an element group 1206 for hidden layer elements, and an element group 1207 for output layer elements. FIG. 12C is a truth table showing the relationship between driver number and driving preference model.

[0097] The difference from the first embodiment of the present invention is that a group of input elements 1205d for driver number has been added to the input of the neural network, and the driving logs and driver responses of the test vehicle 402 are acquired from a plurality of model drivers, and when acquiring a data set, information on the drivers who responded at the same time is acquired.

[0098] If four model drivers are prepared, the server 401 will also manage a data set for each driver who has driven the test vehicle 402, and by adding two inputs related to the driver No. described above to the neural network during learning, different control mode switching modes can be reproduced depending on the driver No. In the driving preference setting unit 1202, the driver can set driving preferences by inputting the driver No. of a model driver who has the driving style he or she prefers.

[0099] The truth table 1208 is for the case where there are four model drivers, and since it can be expressed in two bits, two input elements of the neural network can be added. The driving situation determination unit 1204 outputs one of a plurality of control modes using the driver's driving preferences, vehicle behavior information, and learning parameters. Details other than those described here are the same as those in the first embodiment of the present invention, so detailed descriptions will be omitted.

[0100] [Third Embodiment] The third embodiment of the present invention is characterized by varying the control mode depending on the passenger conditions during driving. More specifically, when there are children or many passengers, and there is a possibility that some of them may suffer from motion sickness, it is considered to prevent aggressive driving. To achieve this, the vehicle is equipped with all or any of a weight sensor that determines the presence or absence of passengers, an in-vehicle camera that can acquire information including passenger characteristics, and a family mode input means that can be set by the driver. When the control mode is automatically switched based on information acquired from the weight sensor or the in-vehicle camera, it is adjusted to meet pre-specified specifications. Furthermore, if a family mode ON / OFF function is provided, it can be adjusted at the driver's discretion. Note that even when the control mode is automatically switched, the driver may be able to adjust the specifications using an application running on an information processing device such as a smartphone.

[0101] Next, a third embodiment of the present invention will be described with reference to Fig. 13 and Fig. 14. Fig. 13A is a block diagram of an automatically adaptable vehicle system including an optimal control mode estimation unit 106 according to the third embodiment of the present invention, and Fig. 13B is a table showing the correspondence between input modes such as a family mode and switching of control modes in the driving situation determination unit of the third embodiment of the present invention.

[0102] In FIG. 13A, 1301 denotes a vehicle, 1302 denotes a driving situation determination unit, 1303 denotes a weight sensor, 1304 denotes an in-vehicle camera, 1305 denotes a control mode setting unit, and 1306 in FIG. 13B denotes processing contents.

[0103] The difference from the first embodiment of the present invention is that the vehicle control device of the vehicle 1301 has a driving situation determination unit 1302, and refers to the additional weight sensor 1303, the in-vehicle camera 1304, and input related to the family mode as one of the passenger information input by the driver himself, and reflects this in the manner of switching the control mode.

[0104] The driving situation determination unit 1302 acquires passenger information of the vehicle and determines the driving situation of the vehicle. The driving situation determination unit 1302 receives a family mode signal, a passenger determination result based on the measurement results of the weight sensor 1303 and information from the in-vehicle camera 1304, and transfers passenger information, such as the number of passengers and whether the passengers include children or elderly people, to the control mode setting unit 1305. The weight sensor 1303 may be a sensor used to determine whether a seat belt is fastened, and the in-vehicle camera 1304 is installed in the front of the living space of the vehicle 1301 in a position where the faces of the passengers can be confirmed and where the age group, etc., can be determined using existing image recognition technology. The control mode setting unit 1305 sets an optimal control mode using the driving situation of the vehicle determined by the driving situation determination unit 1302, the control mode output by the driving situation determination unit 109, and the driving skill information.

[0105] 13B shows an example of processing content 1306 of the control mode setting unit 1305 based on these inputs. First, when the family mode input by the driver is 0 (not selected), the same processing as in the first embodiment of the present invention is performed. In other words, the output of the driving situation determination unit 1302 is not reflected in the processing of the control mode setting unit 1305. Furthermore, when the measurement result of the weight sensor 1303 indicates that there is only one occupant, i.e., the driver, the same processing as in the first embodiment of the present invention is performed regardless of the value of the family mode.

[0106] When the family mode is 1 (selected), the processing of the control mode setting unit 1305 is changed based on the information obtained from the weight sensor 1303 and the in-vehicle camera 1304. One example is processing content 1306, which, for example, when no children are on board regardless of the number of occupants, is the same processing as in the first embodiment of the present invention, but when a child is on board, a control mode one level lower than the output a of the driving situation determination unit 109 is selected, for example, when the output a is the sport mode, the normal mode is selected, and when the output a is the normal mode, the comfort mode is selected.

[0107] These are just examples, and the processing contents of the control mode setting unit 1305 and the number of occupants and the age range of the passengers that can be obtained by the in-vehicle camera are not limited to the processing contents 1306.

[0108] On the other hand, FIGS. 14A and 14B show an example of a case where the process content 1306 can be adjusted by the driver.

[0109] 14A is a diagram showing an example in which a driver adjusts the specifications of the driving situation determination unit 110 according to the third embodiment of the present invention using an information terminal 1104 such as a smartphone. FIG. 14B is a diagram showing an example of a specification table and a specification adjustment screen on an information terminal for the driving situation determination unit of the automatically adaptable vehicle system according to the third embodiment of the present invention. Reference numerals 1401 and 1402 are the specification table, 1403 is an example of the specification adjustment screen, and 1404 is a selection frame.

[0110] The information terminal 1104 may be the same as the information terminal that serves as means for presenting the control modes to the driver as described in the first embodiment of the present invention, and is capable of displaying a specification table 1401 such as the processing content 1306 in the same manner as the application that presents the control modes, or in a separate linked application. The specification table 1402 is an enlarged version of the specification table 1401, but the output result of the weight sensor 1303, the detection result of the in-vehicle camera 1304, and the output a'' of the control mode setting unit 1305 can each be changed by the driver holding the information terminal 1104.

[0111] When a cell / area to be changed is selected, a specification adjustment screen 1403 and a selection frame 1404 are displayed, and the driver can change the specifications of the output a'' of the control mode setting unit 1305 by moving the selection frame 1404. Of course, the weight and passenger conditions for which the output a'' of the control mode setting unit 1305 is to be adjusted can also be changed in the same way.

[0112] This makes it possible to switch the control mode in a way that reflects the driver's intentions, in contrast to the specifications regarding switching of the control mode at the time of shipment.

[0113] The details other than those described here are the same as those in the first embodiment of the present invention, and therefore will not be described in detail.

[0114] [Fourth Embodiment] A fourth embodiment of the present invention is characterized in that the driver can adjust the manner of automatic control mode switching. Expert drivers who are highly skilled may wish to adjust the control mode more precisely to their driving behavior. For example, this may be the case when the existing switching manner for acceleration and deceleration is sufficient, but the driver wishes to have a sportier steering feel. To achieve this, the fourth embodiment of the present invention allows the driver to adjust the accelerator, brake, and steering angle signals input to the driving situation determination unit 109.

[0115] A fourth embodiment of the present invention will be described with reference to Fig. 15. Fig. 15 is a block diagram showing a case where a control mode is determined in the driving skill determination unit 110 of the automatically adaptable vehicle system according to the fourth embodiment of the present invention. In Fig. 15, reference numeral 1501 denotes a signal adjustment unit, 1502 denotes a specification table, 1503 denotes an example of a specification adjustment screen, and 1504 denotes a selection frame.

[0116] As shown in specification table 1502, signals can be adjusted for operation information received from CAN 105, in this case, the physical quantities of the accelerator, brake, and steering angle. These signal adjustments can be performed on the instrument panel 1101 or information terminal 1104 described in the first embodiment of the present invention, and specification table 1502, for example, is displayed on the screen of the running application. When the driver specifies the item to be adjusted, specification adjustment screen 1503 is displayed, allowing the setting to be selected in selection frame 1504. Note that if all adjustment values ​​are ×1.0, the same control mode as in the first embodiment of the present invention is selected, and fine adjustments can be made by varying the adjustment values.

[0117] The adjustment targets listed in the specification table 1502 are just an example, and although the options on the specification adjustment screen 1503 are in increments of 0.1, it goes without saying that they are not limited to this. Rather than setting the individual values ​​of the accelerator, brake, and steering angle, it is also possible to select the relative relationship between the three, such as setting only the steering control mode closer to a sports mode, or setting the acceleration / deceleration control mode closer to a sports mode.

[0118] This makes it possible to switch the control mode in a way that reflects the driver's intentions, in contrast to the specifications regarding switching of the control mode at the time of shipment.

[0119] The present invention is not limited to the above-described embodiment, and other embodiments that are conceivable within the scope of the technical concept of the present invention are also included within the scope of the present invention as long as they do not impair the characteristics of the present invention. Furthermore, a configuration that combines the above-described embodiment with multiple modified examples may also be adopted.

[0120] 101...vehicle, 102...steering control, 103...suspension control, 104...motor control, 105...CAN, 106...optimum control mode estimation unit, 107...parameter storage unit, 108...switch, 109...driving situation determination unit, 110...driving skill determination unit, 111...control mode setting unit, 112...map storage unit, 113...display unit, 114...driver data storage unit, 201...longitudinal acceleration waveform, 202...lateral acceleration waveform, 203...vehicle speed waveform, 204...control mode waveform, 20 5...Window, 206...Control mode value, 301...Input layer element group, 302...Hidden layer element group, 303...Output layer element group, 304...Truth value, 401...Server, 402...Test vehicle, 403..., 501...Acquire driving data & driver's answer, 502...Transfer to server, 503...Receive driving data & driver's answer, 504...Store driving data & driver's answer, 505...Read data set, 506...Correlation learning, 507...Store weight parameter, 508...Read weight parameter, 509...Weight parameter Distribution, 510...receive weight parameters, 511...store weight parameters, 512...read weight parameters, 513...try new parameters, 514...determine error, 515...apply new parameters, 516...resume automatic adaptation, 517...apply old parameters, 518...report error, 519...erase new parameters, 601...vehicle direction estimation unit, 602...self-position estimation unit, 603...surrounding area map acquisition unit, 604...steering angle candidate calculation unit, 605...target steering angle selection unit, 606...steering angle difference calculation unit, 607...target Steering angle, 608... Ideal steering angle transition, 609... Actual steering angle transition, 610... Steering angle difference, 611... Driver ID acquisition, 612... ID registration determination, 613... Driving skill information acquisition, 614... Beginner setting, 615... Initialization (N, Diff), 616... Road / driving state detection, 617... Intersection presence / absence determination, 618... Target steering angle candidate calculation, 619... Steering angle information acquisition, 620... Right / left turn operation completion determination, 621... Steering angle difference D calculation, 622... Error accumulation (Diff + = D), 623... Count up (N),624...Travel process end determination, 625...Average error calculation, 626...Driving skill determination, 627...Driver data update (addition), 701...Speed ​​limit information extraction unit, 702...Congestion information acquisition unit, 703...Vehicle speed difference calculation unit, 704...Target vehicle speed, 705...Ideal vehicle speed transition, 706...Actual vehicle speed transition A, 707...Vehicle speed difference A, 708...Actual vehicle speed transition B, 709...Vehicle speed difference B, 801...Input layer element group, 80 2...Hidden layer element group, 803...Output layer element group, 804...Truth table, 901...Processing content, 902...Signal adjustment unit, 903...Processing content, 1001...Control mode (sparse mode set), 1002...Control mode (dense mode set), 1003...Truth table, 1101...Instrument panel, 1102...Information presentation unit, 1103...Information presentation content, 1104...Information terminal, 1105...Behavior Information waveform, 1106...information presentation content, 1201...vehicle, 1202...driving preference setting unit, 1203...parameter storage unit, 1204...driving situation determination unit, 1205...input layer element group, 1206...hidden layer element group, 1207...output layer element group, 1208...truth table, 1301...vehicle, 1302...driving situation determination unit, 1205...input layer element group, 1206...hidden layer element group, 1207... Output layer element group, 1208... truth table, 1301... vehicle, 1302... driving situation determination unit, 1303... weight sensor, 1304... in-vehicle camera, 1305... control mode setting unit, 1306... processing content, 1401... specification table, 1402... specification table, 1403... specification adjustment screen, 1404... selection frame, 1501... signal adjustment unit, 1502... specification table, 1503... specification adjustment screen, 1504... selection frame,

Claims

1. A vehicle control device capable of setting multiple control modes for an actuator mounted on a vehicle, characterized in that it comprises an optimal control mode estimation unit that estimates an optimal control mode from among the multiple control modes using behavior information of the vehicle, learning parameters obtained by machine learning of the correlation between the behavior information and the multiple control modes, and information on the driving skill of the driver who drives the vehicle.

2. The vehicle control device described in claim 1, characterized in that the optimal control mode estimation unit comprises: a driving situation determination unit that outputs one of the plurality of control modes using the vehicle behavior information and the learning parameters; a driving skill determination unit that acquires information on the driving skill based on operation information of the vehicle; and a control mode setting unit that sets the optimal control mode using the control mode output by the driving situation determination unit and the information on the driving skill.

3. The vehicle control device described in claim 2, characterized in that the driving skill determination unit outputs one of the multiple driving skill levels using operation information of the vehicle by the driver and learning parameters that have been machine-learned to determine the correlation between the operation information and multiple driving skill levels.

4. The vehicle control device described in claim 2, characterized in that the driving skill determination unit comprises: a steering angle candidate calculation unit that calculates multiple target steering angle candidates based on the vehicle direction, its own position, and map information; a target steering angle selection unit that selects one of the multiple target steering angle candidates as the target steering angle based on the vehicle's steering angle information; and a steering angle difference calculation unit that compares an ideal steering angle transition from the vehicle's pre-steering state to the target steering angle with an actual steering angle transition, and determines the driving skill based on the steering angle difference between the ideal steering angle transition and the actual steering angle transition.

5. The vehicle control device described in claim 2, characterized in that the driving skill judgment unit comprises: a speed limit information extraction unit that extracts speed limit information based on the vehicle direction, its own position, and map information; and a vehicle speed difference calculation unit that calculates the vehicle speed difference between the speed limit information and the vehicle's speed information, and judges the driving skill based on the vehicle speed difference.

6. The vehicle control device according to claim 2, characterized in that the control mode setting unit corrects the output of the driving situation determination unit based on the driving skill, and sets the optimal control mode based on the corrected output.

7. A vehicle control device as described in claim 2, further comprising a signal adjustment unit that corrects the vehicle behavior information according to the judgment result of the driving skill judgment unit, and wherein the driving situation judgment unit outputs one of the plurality of control modes using the behavior information corrected by the signal adjustment unit and the learning parameters.

8. A vehicle control device as described in claim 2, further comprising a driving preference setting unit that sets the driving preferences of the driver, and wherein the driving situation determination unit outputs one of the plurality of control modes using the driving preferences of the driver, the vehicle behavior information, and the learning parameters.

9. A vehicle control device as described in claim 2, characterized in that it comprises a driving situation determination unit that acquires passenger information of the vehicle and determines the driving situation of the vehicle, and the control mode setting unit sets an optimal control mode using the driving situation of the vehicle determined by the driving situation determination unit, the control mode output by the driving situation determination unit, and information on the driving skill.

10. The vehicle control device according to claim 2, characterized in that the vehicle behavior information is time series data of an in-vehicle network, and the information referenced by the control mode setting unit is vehicle behavior events included in time series data for a predetermined period of time or time series data for a predetermined driving course.

11. The vehicle control device described in claim 2, characterized in that the control mode setting unit provides multiple control modes with smaller characteristic differences between the control mode settings when the driver's driving skill is lower than a standard, compared to when the driver's driving skill is higher than the standard.

12. The vehicle control device described in claim 2, characterized in that the control mode setting unit comprises a sparse mode set consisting of a plurality of control modes in which the differences in parameters between the respective control modes are large, and a dense mode set consisting of a plurality of control modes in which the differences in parameters between the respective control modes are small, and selects the dense mode set when the driving skill determination unit determines that the driver has low skill, and selects the sparse mode set when the driving skill determination unit determines that the driver has high skill, or allows the driver to select either the dense mode set or the sparse mode set.

13. The vehicle control device according to claim 2, wherein the learning parameters are acquired from a server.

Images