WO2013132961A1

WO2013132961A1 - Driving assistance device

Info

Publication number: WO2013132961A1
Application number: PCT/JP2013/053109
Authority: WO
Inventors: 水野将弘
Original assignee: アイシン精機株式会社
Priority date: 2012-03-05
Filing date: 2013-02-08
Publication date: 2013-09-12
Also published as: JP2015099406A

Abstract

Provided is a learning-type driving assistance device that is capable of learning done so as to conform to a travel tendency of a driver. Provided are: a host vehicle travel behavior acquisition unit that acquires travel behavior information related to the travel behavior of a host vehicle that travels; a mental/physical information acquisition unit that acquires mental/physical information related to the mental/physical condition of the driver; a concentration level assessment unit that on the basis of the mental/physical information, assesses the driving-related concentration level of the driver; a travel behavior learning unit that on the basis of the travel behavior information acquired by the host vehicle travel behavior acquisition unit, learns the travel behavior of the driver; and a learning prohibition unit that prohibits learning if the concentration level assessed by the concentration level assessment unit is no more than an assessment condition.

Description

Driving assistance device

The present invention relates to a learning type driving support apparatus that learns a driving tendency of a driver.

Driving by considering the driving tendency of the driver in the steering control device that returns the vehicle to the center of the lane based on the amount of lateral deviation between the center point in the width direction of the lane where the vehicle is traveling and the vehicle position A learning-type steering control device that avoids forced steering control contrary to the intention of a person is known from Patent Document 1.
In this learning type steering control device, from the driver travel tendency information obtained by the travel tendency learning means for learning which position in the lane width direction the driver tends to travel based on the vehicle position information. The driver travel tendency correction amount is calculated. The basic vehicle travel line is corrected using the driver travel tendency correction amount. The steering control of the vehicle is performed so that the vehicle travels on the corrected vehicle travel line as a learning result by actual travel, thereby reducing the uncomfortable feeling given to the driver. In this steering control, when the vehicle is traveling on a curve having a predetermined curvature or more, the steering control of the vehicle is performed so that the vehicle travels on the corrected vehicle travel line, and the steering control is adapted to the driving tendency of the driver. Has been done.

Japanese Patent Laying-Open No. 2005-059643 (paragraph number [0002-0006], FIG. 5)

However, the driver does not always travel according to his / her driving tendency, and when driving while not concentrating on driving, the driving trajectory during driving is the driving tendency of the driver It can be different. When learning about traveling in this state, during normal driving, there arises a problem that steering control contrary to the will of the driver is performed. Further, in actual driving, particularly in curve driving, the driving tendency of the driver changes depending on the vehicle speed, date, weather, road conditions, and the like. For this reason, if learning is performed in association with only the curvature of the curve, there is a possibility that steering control that is also contrary to the will of the driver may be performed.
In view of the above circumstances, there is a demand for a learning-type driving support device that can learn to match the driving tendency of the driver.

The features of the driving support device according to the present invention are: a host vehicle travel behavior acquisition unit that acquires travel behavior information related to the travel behavior of the host vehicle traveling; a mind / body information acquisition unit that acquires mind / body information related to the mind / body state of the driver; A concentration degree evaluation unit that evaluates the degree of concentration related to driving of the driver based on mind and body information, and a driving behavior that learns the driving behavior of the driver based on the driving behavior information acquired by the own vehicle driving behavior acquisition unit A learning unit and a learning prohibiting unit that prohibits the learning when the concentration evaluated by the concentration evaluation unit is equal to or less than a determination condition.

According to this configuration, when the driving support device learns the driving behavior of the vehicle by the driver, it is determined whether the driver is in a normal or normal state of mind and body driving. The degree of concentration related to the driving of the driver used for this determination is evaluated based on the mind-body information relating to the mind-body state of the driver. Furthermore, only when it is determined that the vehicle is in a normal or normal driving state, the traveling behavior learning is executed. Conversely, if normal or normal driving is not performed, traveling behavior learning is not executed. Therefore, the driving behavior learning for matching the driving tendency of the driver is performed only when the driver is in a mental and physical state in which he / she performs his / her own driving. Therefore, it is possible to realize driving behavior learning that is optimal for the driving tendency of the driver.

One of the driving behaviors suitable for providing driving assistance that matches the driving tendency of the driver is to maintain the position in the width direction of the driving lane, that is, lane keeping.
In one preferred embodiment of the present invention, the traveling behavior is a position in the traveling lane width direction of the own vehicle traveling in the traveling lane, and the own vehicle traveling behavior acquisition unit is the own vehicle in the width direction of the traveling lane. It is comprised as the own vehicle horizontal position acquisition part which acquires a horizontal position. According to this configuration, the lateral driving position learning for adjusting to the driving tendency of the driver is performed only when the driver is in a mental and physical state where the driver performs his own driving. For this reason, learning of the vehicle side position optimal for the driving tendency of the driver is realized.

In one preferred embodiment of the present invention, the mind-body information is information based on a face image of a driver photographed by a vehicle-mounted camera, and the driver identification unit further provided is photographed by the vehicle-mounted camera. The driver is identified based on the photographed face image of the driver. According to this configuration, the driver's face photographed image necessary for creating the mind and body information and the face photographed image for identifying the driver used in the driver identification unit are acquired from a common in-vehicle camera. Therefore, it is possible to share the configuration that reduces the cost.

One preferable example of the mind-body information is gaze information related to the driver's gaze or face orientation information related to the driver's face orientation. If the line of sight while driving is unnaturally moving or gazing at a place deviated from the traveling direction, it is not concentrated on driving, and it can be considered that normal or normal driving cannot be performed. Further, instead of the line-of-sight information or the face orientation information, biological information such as heart rate information, respiratory information, body motion information, sputum information, brain waves, body temperature, etc., and information on temporal changes thereof may be used. Furthermore, it is also possible to use information on the travel behavior, the driver's driving behavior (steering operation, braking operation, accelerator operation) and their temporal variation.

In a preferred embodiment of the present invention, a deviation allowable range setting unit that sets a deviation allowable range with respect to the reference of the driving behavior is provided, and the driving behavior learning unit is configured to allow the deviation allowable based on the driving behavior information. Configured to modify the range.
According to this configuration, setting or correction of the target travel line that guides the host vehicle through automatic steering control is performed when it is determined that the driver is in a normal or normal state of mind and body driving. Therefore, it is possible to set a target travel line adapted to the driver's optimal travel behavior.
In addition, when it is configured to support driving so that the driving behavior of the host vehicle is maintained within the allowable deviation range in the width direction of the driving lane, the driver must be normal or normal in learning the allowable deviation range. It is determined whether or not the vehicle is in a state of mind and body in which to drive. Thereby, the driving behavior learning for adjusting the deviation allowable range to the driving tendency of the driver is performed only when the driver is in a mental and physical state where he / she performs his own driving. For this reason, the driving behavior learning optimal for the driving tendency of the driver is realized.
As described above, in order to maintain the vehicle traveling so that the traveling behavior falls within the deviation allowable range, the position that prompts the driver to return from the separation when the traveling behavior has deviated from the deviation allowable range. It is important that a retention promoting part is provided. The following two are proposed in this invention as a suitable form of this position maintenance promotion part.
One of them is that the position retention promoting unit is constructed as a notification unit that notifies the driver of departure from the deviation allowable range, and the other one is that the vehicle returns to the deviation allowable range. Thus, it is constructed as a return control unit that executes vehicle return control. For example, the notification unit can be incorporated in a notification ECU that notifies the driver of various driving information. The return control unit can be incorporated in a vehicle control ECU that performs steering control, drive / brake control, and the like.
Further, as a functional unit that can be used in place of the return control unit or selectively with the return control unit, a lane keeping control unit that always gives steering control information to the vehicle control ECU in order to maintain the set target travel line. It can also be incorporated in a vehicle control ECU. At this time, a target travel line setting unit that sets a target travel line of the host vehicle that can be corrected by the travel behavior learning unit is provided, and vehicle control is performed so as to maintain the target travel line.

Since there are an infinite number of roads to be supported by the driving support device, it is necessary to quickly find and set the allowable deviation range for each road to be supported.
For this purpose, in a preferred embodiment of the present invention, road information relating to the road obtained from the car navigation system and / or road information relating to the road obtained from the processing result of the image taken by the in-vehicle camera are used. An allowable range derivation table for deriving the allowable deviation range using information as an input parameter is provided. In particular, by using a deviation tolerance range derivation table (a deviation tolerance range database) using road information from a car navigation system with a high degree of spread as an input parameter (search condition), a deviation tolerance range can be quickly derived. On the other hand, it is possible to generate road information about a traveling road acquired from a processing result of a captured image of an in-vehicle camera without using a car navigation system, and this road information may be used as an input parameter.
The driving route of each driver often changes depending on the driving environment, for example, depending on the type of road (highway, general road), weather, season, date and time. Therefore, it is convenient to use any one of the own vehicle traveling speed, the road curvature of the traveling lane, the weather, the date and time, or a combination thereof as the input parameter.

When a car is used by a family or company, multiple people are drivers. When the driving support device of the present invention is applied to such a vehicle used by a plurality of drivers, a driver identification unit that identifies the driver as the input parameter in order to set the deviation allowable range for each driver. It is preferable that the generated driver identification information is used.

It is a schematic diagram illustrating the basic principle of the learning process in the learning type driving support device of the present invention. It is a schematic diagram illustrating the basic principle of the driving assistance process in the learning type driving assistance apparatus of this invention. It is a schematic diagram illustrating the basic principle of the driving assistance process in the learning type driving assistance apparatus of this invention. It is a functional block diagram which shows an example of the control system of the vehicle carrying the learning type | mold driving assistance device of this invention. It is a schematic diagram illustrating the flow of data in a learning type driving support device. It is a flowchart figure which shows the flow of control in a learning type driving assistance device. It is a schematic diagram illustrating the basic principle of the learning process in the driving assistance device that employs the inter-vehicle distance as the running behavior. It is a schematic diagram illustrating the basic principle of a driving support process in a driving support device that employs an inter-vehicle distance as a running behavior. It is a schematic diagram illustrating the basic principle of a driving support process in a driving support device that employs an inter-vehicle distance as a running behavior.

Before describing a specific embodiment of the driving support apparatus according to the present invention, the basic principle will be described.
The driving behavior in which driving assistance is performed by the driving assistance device is based on maintaining the position in the width direction of the traveling lane, that is, lane keeping. In order to explain the present invention in an easy-to-understand manner, the driving support device detects a lateral deviation amount between a preset target position (reference position) in the width direction of the lane and the actual own vehicle position, and the lateral deviation is detected. Necessary information is given to the driver based on the quantity, and necessary vehicle control is performed. Then, it is assumed that learning is performed so that the target position used at that time matches the driving tendency of the driver.
FIG. 1 illustrates the basic principle of the learning process. The roads that are subject to driving assistance by the driving assistance device may be all roads, high-risk roads, or curved roads that need to be consciously cornered. May be.
In the following explanation, a curved road is taken up as a support target road. In addition, the amount of lateral displacement can be detected by calculating the position of a road image (an image including a sign such as a center line or lane line) obtained by the own camera, calculating the position using high-precision GPS, or calculating the position using road-to-vehicle communication. Can be adopted. In the present invention, the method of detecting the lateral deviation amount is not limited.

First, during road driving, road information 20 is acquired from a car navigation system, VICS (registered trademark) (Vehicle Information and Communication System), or the like (# 01). When a road to be driven is a target for driving support, a road ID, a road width, a curve radius of curvature, an inclination angle, a curved road method (clockwise, counterclockwise), and the like are read from the acquired road information 20. Next, using at least one of these as an input parameter, the deviation allowable range 21 in the width direction of the traveling lane is derived from the allowable range deriving table (# 02). In addition, you may take in environmental conditions, such as a own vehicle travel speed, a weather, a date, and a season, as an additional input parameter.
The derived deviation allowable range 21 is set to a road (curved road) that is a target of driving assistance (# 03). In order to set the deviation allowable range 21, a reference travel line is required. Therefore, as shown in FIG. 1, the deviation allowable range 21 is equally distributed between the in side and the out side with respect to the reference travel line which is the center line of the road width. The allowable deviation range 21 may be the same value from the beginning to the end of the curved road, but may be set for each predetermined distance or for each position where the allowable deviation range 21 changes.
FIG. 1 shows a mode in which a deviation allowable range 21 is set at positions P1 to P5. P0 indicates the vehicle position when the allowable deviation range 21 is set.

Next, mind-body information 22 relating to the mind-body state of the driver is acquired. Here, eyes are extracted from the face image 32 of the driver photographed by the camera 10 disposed in the vicinity of the windshield, and information on the line of sight, the direction of the line of sight, changes in the line of sight over time, etc. are used as the mind-body information 22. Yes.
The driver's concentration 23 is evaluated based on parameters included in the mind-body information 22 and additional parameters such as the posture and behavior of the driver that are adopted as necessary. Compared with the determination condition (threshold value) set in advance, the degree of concentration 23 is a state where the driver is concentrated (concentration), or a state where the driver is not concentrated (de-concentration) when the determination condition is below Is determined (# 04). The state of being concentrated indicates a state of mind and body in which driving is performed normally or normally for the driver. The state of not concentrating indicates a state of mind and body in which driving is not performed normally or normally.

A feature of the present invention is that the running position learning is performed to correct the above-described deviation allowable range 21 based on the vehicle lateral position 27 during actual running, and the running position learning is performed normally or normally. It is only executed when driving. Therefore, when it is determined that the driver is in a non-concentrated state, traveling position learning is prohibited.

If it is determined that the driver is in a concentrated state, learning information for driving position learning is acquired at predetermined time intervals or predetermined driving distances (# 05). The learning information includes a map coordinate position acquired from the car navigation system or a vehicle position 26 that is a road position in the direction of travel, and the vehicle in the width direction of the travel lane that can be acquired by image processing the captured image from the camera 10. The vehicle side position 27, which is the position, additional information 28 such as the vehicle traveling speed and environmental conditions that can be acquired from the sensor system of the vehicle. Furthermore, when the allowable range deriving table is constructed for each driver, driver specifying information 29 for specifying the driver is also acquired.

When the learning information is acquired, learning data 33 in which the amount of lateral displacement of the vehicle for each traveling position (P1 to Pn) is recorded is generated together with the identified driver ID and the like (# 06). The lateral displacement amount of the host vehicle is an amount that is out of the allowable displacement range 21, where + indicates a deviation in the out direction and-indicates a deviation in the in direction. When the driving on the road to be driven is finished, the compatible part of the deviation allowable range table is updated based on the generated learning data 33 (# 07).

Next, the basic principle of the driving assistance that suppresses the departure from the deviation allowable range 21 will be described with reference to the schematic diagrams of FIGS. 2A and 2B. The driving support control can be performed simultaneously with the learning control.
When it is confirmed from the road information 20 acquired from the car navigation system or the like that the vehicle enters the driving support target road, the driver deviation allowable range 21 from the allowable range derivation table using the road information 20 and environmental conditions as input parameters. Is derived. Furthermore, the derived deviation allowable range 21 is set to a road (curved road) (see FIG. 2A).

Next, as the vehicle enters the driving assistance target road, the vehicle lateral position 27 in the width direction of the traveling lane on this road is detected. The own vehicle lateral position 27 at each traveling position (P1 to Pn) is compared with the deviation allowable range 21 set in advance, and it is checked whether or not the own vehicle has left the allowable deviation range 21. For example, in FIG. 2B, the departure in the out direction from the deviation allowable range 21 is detected at the traveling position P3. When the departure from the deviation allowable range 21 is detected, in order to prompt the driver to return to the deviation allowable range 21, the departure from the deviation allowable range 21 is visually or audibly or both. Informed. Furthermore, when the departure information is sent to the vehicle control system, the operation torque of the steering operation system is changed so that the steering operation toward the departure side becomes heavy (steering restriction control to the separation side), or oversteering The control constant of the stabilizer mechanism or the like is changed so as to become (vehicle traveling control in the reverse direction of separation).

In the case where the corrected or newly created deviation allowable range 21 is used through learning when the driver is traveling on the same or similar road in advance, the set deviation allowable range 21 is out in accordance with the driving tendency of the driver. It has been corrected to the side or the in side. For this reason, driving assistance in accordance with the driving tendency of the driver is performed. For example, for a driver who tends to bulge out, no warning or vehicle control is executed even if the vehicle bulges out, as long as it is within the corrected deviation allowable range 21.

Hereinafter, specific embodiments of the driving support apparatus according to the present invention will be described. The driving support device is mounted on a car and can be represented by functional blocks as shown in FIG. Electronic control units (hereinafter simply referred to as ECUs) that are control systems for automobiles are a camera ECU 2, an image processing ECU 3, a driving assistance ECU 4, an in-vehicle sensor ECU 5, a vehicle control ECU 6, a notification ECU 7, and a car navigation ECU 8.

The camera 10 handled in the present embodiment includes a driver photographing camera 10a that photographs a driver, and a lateral position detection camera 10b that photographs a sign that can serve as a reference in the width direction of a traveling lane including a lane line. It is. The camera ECU 2 classifies the captured images sent from the cameras 10 according to purpose at a predetermined timing, and transfers them to the image processing ECU 3. The image processing ECU 3 performs face detection and face feature information generation for driver identification using the face photographed image 32 obtained by the driver photographing camera 10a. In addition, information on the driver's line of sight is generated as the driver's mind-body information 22 from the detected pupil movement in the face. Furthermore, the face photographed image 32 from the same driver photographing camera 10a is also used for identifying the driver. The image processing ECU 3 also generates a lateral position detection image 35 such as an edge image of a lane line using the road photographed image.
The camera 10 and the camera ECU 2 including the driver photographing camera 10a may be separate units, but the camera 10 and the camera ECU 2 may be integrated. In that case, it is also possible to integrate with one camera ECU2 for every camera. Further, the image processing ECU 3 may be integrated with the camera 10 and the camera ECU 2.

The in-vehicle sensor ECU 5 processes a signal from a sensor that detects control amounts of various drive devices mounted on the vehicle, and transfers the signal to the ECU that requires the signal. The vehicle control ECU 6 controls the drive by sending a control signal to various drive devices based on the sensor signal and the driving control signal sent from the vehicle-mounted sensor ECU 5. For example, it is possible to realize oversteering or understeering by giving a control signal to a driving device of the steering system so as to make the steering in one direction heavier or adjusting the control constant of the stabilizer mechanism.
The notification ECU 7 outputs display data to a monitor constituted by a liquid crystal panel and the like, and visually gives various information to the driver, and also gives sound and buzzer sound through a speaker.

The car navigation ECU 8 has a function of calculating the own vehicle traveling position 26 and applying the own vehicle traveling position 26 to the road on the map extracted from the map database. Since the car navigation ECU 8 has a function of specifying the traveling road and a function of calculating the own vehicle traveling position 26 in the traveling direction on the road, these functions are used in the present embodiment.
As is well known, the car navigation ECU 8 uses position calculation by GPS and position calculation by dead reckoning navigation in order to calculate the vehicle traveling position 26. However, in some cases, it is also possible to improve the accuracy of the vehicle traveling position 26 by road-to-vehicle communication or the like. In dead reckoning, a distance sensor, which is a sensor that detects the vehicle speed and travel distance of the vehicle, a vehicle speed pulse sensor that outputs a pulse signal every time a drive shaft or wheel of the vehicle rotates a certain amount, and the acceleration of the host vehicle C are detected. An azimuth sensor composed of a yaw / G sensor that performs the detection and a circuit that integrates the detected acceleration is used.
The car navigation ECU 8 can output road information 20 that is information about a road to be traveled. The road information 20 includes, for example, a radius of curvature of a curved road and a road width.

The driving support ECU 4 includes an allowable deviation range database 9 that functions as an allowable range derivation table for deriving the allowable deviation range 21 based on input parameters as search conditions. The allowable deviation range 21 stored in the allowable deviation range database 9 includes not only various parameters included in the road information 20 sent from the car navigation ECU 8, but also environmental conditions such as traveling speed, traveling lane weather, date and time, and driver ID. Is linked. The deviation allowable range 21 is extracted using these as input parameters.

The driving support ECU 4 includes, as functional units related to the present invention, a deviation allowable range setting unit 41, a travel behavior acquisition unit (which functions as a vehicle lateral position acquisition unit in this embodiment) 42, a mind-body information acquisition unit 43, a concentration A degree evaluation unit 44, a travel behavior learning unit (which functions as a travel position learning unit in this embodiment) 45, a learning prohibition unit 46, and a driver identification unit 47 are included. The flow of control data generated by each of these functional units and transferred to each functional unit is schematically shown in FIG.

The driver identifying unit 47 identifies the driver from the face photographed image 32 of the driver who is driving while sitting in the driver's seat, which is sent from the image processing ECU 3, and stores the driver ID as the current driver in a predetermined memory. Keep a record. For this purpose, a facial photograph or facial feature information of a person who may drive this car is created in advance. Of course, instead of such face authentication, a method of specifying a driver by inputting an identification code or inputting a fingerprint may be adopted.

The deviation allowable range setting unit 41 specifies the parameters included in the road information 20 sent from the car navigation ECU 8 and the environmental conditions sent from the in-vehicle sensor ECU 5 and the like, and further specifies the driver sent from the driver identification unit 47. The deviation allowable range database 9 is accessed using the driver identification information 34 such as the driver ID as a search condition. And the deviation tolerance | permissible_range 21 for the identified driver | operator is extracted in the road (curved road) from which it will drive | work from now on. The deviation allowable range setting unit 41 assigns the extracted deviation allowable range 21 to the road to be supported, thereby setting the deviation allowable range 21 for the road to be traveled from now on.

The own vehicle lateral position acquisition unit 42 acquires the own vehicle lateral position 27 in the width direction of the travel lane from the lateral position detection image 35 such as an edge image of the lane line sent from the image processing ECU 3.
The mind-body information acquisition unit 43 inputs line-of-sight information sent from the image processing ECU 3 as mind-body information 22 relating to the mind-body state of the driver. The mind and body information acquisition unit 43 may generate line-of-sight information based on the face photographed image 32 sent from the image processing ECU 3.

The concentration degree evaluation unit 44 is in a concentrated state where the driver can perform normal or normal driving based on the line-of-sight information, in particular, a change in the line-of-sight direction or the line-of-sight direction over time, or normal or normal. The degree of concentration 23 for determining whether or not the vehicle is not in a concentrated state is calculated. If the concentration degree evaluation unit 44 determines that the driver is concentrated, a learning command is given to the traveling position learning unit (traveling behavior learning unit) 45. If the concentration evaluation unit 44 determines that the driver is not concentrated, the learning prohibition unit 46 prohibits the travel learning in the travel position learning unit 45.

The travel position learning unit 45 includes a host vehicle travel position 26 (own vehicle position in the travel direction) acquired from the car navigation ECU 8, and a host vehicle lateral position 27 (in the road width direction) given from the host vehicle lateral position acquisition unit 42. The lateral deviation of the own vehicle is calculated based on the own vehicle position) and the allowable deviation range 21 set by the allowable deviation range setting unit 41, and the calculated value is generated as learning data 33. Furthermore, the corresponding data in the deviation allowable range database 9 is corrected using the generated learning data 33.

The flow of driving support processing in the driving support apparatus configured as described above will be described with reference to the flowchart of FIG. FIG. 5 shows an example of a driving support routine. This driving support routine is instructed to execute control for driving support so as to be within an allowable range set on the road on which the vehicle is traveling. As long as it is, it is executed repeatedly.

As pre-processing of the driving support routine, driver identification processing for identifying a driver is performed by the driver identification unit 47, and a driver ID for identifying the driver is stored (# 11). If the driver identification is performed before the driving support routine is activated and the driver ID is stored at a predetermined address, the driver identification process is omitted.
During the execution of the driving support routine, the road information 20 is acquired from the car navigation ECU 8 or the driving support ECU 4 at predetermined intervals (# 12). Based on the acquired road information 20, it is checked whether or not the road to be traveled from now on is a target road for driving assistance (# 13). If the road is not a support target road (# 13 No branch), the process returns to step # 12 and waits for the next road information 20. If it is a road to be supported (# 13 Yes branch), the deviation allowable range database 9 is accessed using the road information 20, environmental conditions, driver ID, etc. as search conditions, and the corresponding deviation allowable range 21 is extracted. Read (# 14).
The deviation allowable range setting unit 41 sets the deviation allowable range 21 read for the support target road (# 15). Here, the setting of the allowable deviation range 21 is mapping so that the allowable deviation range 21 can be called according to the vehicle position associated with the travel on the support target road. Next, the driver's concentration level 23 is calculated by the linkage between the mind and body information acquisition unit 43 and the concentration level evaluation unit 44 (# 16).

If the calculated concentration level 23 is normal or normal operation can be expected (# 21 Yes branch), “ON” is set to the learning flag (# 23). If the calculated degree of concentration 23 is normal or normal operation cannot be expected (# 21 No branch), the learning flag is set to “OFF” (# 22).

Next, the host vehicle travel position 26 is acquired from the car navigation ECU 8 (# 31), and the host vehicle lateral position acquisition unit 42 calculates the host vehicle lateral position 27 (# 32). Next, the deviation allowable range 21 set based on the own vehicle traveling position 26 and the own vehicle lateral position 27 are compared, and the deviation amount in the lane width direction of the own vehicle is calculated (# 33). Here, the contents of the learning flag are checked (# 34).
If the learning flag is “ON” (# 34 Yes branch), the travel on this support target road is learned, so the learning data 33 is created by combining the own vehicle travel position 26 and the deviation amount on the support target road. (# 35). The learning data 33 is used for updating the permissible deviation allowable range 21 of the deviation allowable range database 9 after completion of the driving support routine or as parallel processing.

If the learning flag is “OFF” (# 34 No branch), since the current driver state is not suitable for learning, traveling learning on this support target road is prohibited.

In any case, when the amount of deviation is calculated, it is compared with a preset determination value (allowable range information 36) in order to determine whether the amount of deviation is negligible (# 36). If the deviation amount is greater than or equal to the determination value (# 36 Yes branch), the driver is warned through the speaker or monitor that the deviation is out of the deviation tolerance range 21 (# 41). Furthermore, in order to suppress this separation according to the amount of deviation, vehicle control is performed via the vehicle control ECU 6 to adjust the control constants of the steering system and the stabilizer system (# 42). That is, in this case, the vehicle control ECU 6 functions as a return control unit that executes vehicle return control so that the host vehicle returns to the allowable deviation range 21.
If the deviation amount is less than the determination value (# 36 No branch), of course, control that suppresses the departure from the deviation allowable range 21 is not performed. The process from step # 31 to step # 42 is continued until it finishes driving | running | working the said assistance object road (# 43 No branch). When traveling on the support target road is completed (Yes in # 43), the process returns to step # 12 to wait for the next support target road.

Next, with reference to FIG. 6, FIG. 7A and FIG. 7B, the driving support device in the embodiment in which the inter-vehicle distance is adopted as the traveling behavior will be described. FIG. 6 is a schematic diagram illustrating the basic principle of the learning process, and FIGS. 7A and 7B are schematic diagrams illustrating the basic principle of the driving support process.
Also in this embodiment, the roads that are subject to driving assistance may be all roads, or may be limited to roads with high risk, highways, and slopes. Acquisition of the inter-vehicle distance information 50 is based on inter-vehicle distance calculation based on signals from in-vehicle sensors such as millimeter wave radars and laser radars, inter-vehicle distance calculation by image processing using a front image taken by the in-vehicle camera 10, and inter-vehicle communication. Can be used to calculate distance between vehicles. In the present embodiment, the method for acquiring the inter-vehicle distance is not limited. The traveling behavior acquisition unit 42 functions as an inter-vehicle distance acquisition unit, and the traveling behavior learning unit 45 functions as an inter-vehicle distance learning unit.

First, during road driving, road information 20 and environmental information 38 are obtained from a car navigation system, a driving support ECU 4 and VICS (registered trademark), and vehicle speed information 39 is obtained from an in-vehicle sensor ECU 5 (# 101). When the next road to be driven is subject to driving assistance, road characteristics such as road ID and road width are read from the acquired road information 20, weather and date / time are read from the environment information 38, and vehicle speed is read from the vehicle speed information 39. Using at least one of these as input parameters, an allowable range deviation 21 of the inter-vehicle distance is derived using an allowable range derivation table (# 102). The derived deviation allowable range 21 is set (# 103). The allowable deviation range 21 here is set as + ΔL and −ΔL as a positive / negative deviation allowable range 21 with respect to the standard inter-vehicle distance: L.

Next, mind and body information 22 regarding the mind and body state of the driver is acquired. The driver's concentration 23 is evaluated based on parameters included in the mind-body information 22 and additional parameters such as the posture and behavior of the driver that are adopted as necessary. The concentration degree 23 is in a state where the driver is concentrated (concentration) based on a preset determination condition (threshold value), or in a state where the concentration is below the determination condition (non-concentration) It is determined whether there is (# 104). The state of being concentrated indicates a state of mind and body in which driving is performed normally or normally for the driver. The state of not concentrating indicates a state of mind and body in which driving is not performed normally or normally.

If it is determined that the driver is in a concentrated state, driving behavior learning is executed, and learning information is acquired at predetermined time intervals or predetermined driving distances (# 105). In addition to the inter-vehicle distance information 50, information acquired in learning is additional information 28 such as a map coordinate position or a vehicle travel position 26 that is a road position, a vehicle travel speed, and environmental conditions. Also here, when the deviation allowable range deriving table is constructed for each driver, the driver specifying information 29 for specifying the driver is also acquired.

When the learning information is acquired, learning data 33 in which the amount of deviation of the inter-vehicle distance for each predetermined road area from the standard inter-vehicle distance is recorded is generated together with the additional information 28 such as the identified driver ID and environmental conditions. (# 106). When the driving on the road to be driven is finished, the compatible part of the deviation allowable range table is updated based on the generated learning data 33 (# 107).

Next, the basic principle of driving assistance that suppresses the departure from the deviation allowable range 21 will be described with reference to the schematic diagrams of FIGS. 7A and 7B. Driving support control can be performed simultaneously with learning control.
First, when it is confirmed from the road information 20 obtained from the car navigation system or the like that the vehicle enters the driving support target road, the driving is performed from the allowable range derivation table using the road information 20, the vehicle speed information 39, and the environment information 38 as input parameters. The allowable deviation range 21 regarding the inter-vehicle distance for the person is derived. Furthermore, the allowable deviation range 21 is set based on the derived allowable deviation range 21 (see FIG. 7A).

Next, the inter-vehicle distance is detected as the vehicle enters the support target road portion. The actual inter-vehicle distance is compared with the deviation allowable range 21 set in advance, and it is checked whether or not the host vehicle has left the deviation allowable range 21. For example, in FIG. 7B, the actual inter-vehicle distance is short, and the departure from the deviation allowable range 21 is detected. When the departure from the deviation allowable range 21 is detected, in order to prompt the driver to return to the deviation allowable range 21, the departure from the deviation allowable range 21 is visually or audibly or both. Informed. Further, when the separation information is sent to the vehicle control system via the vehicle control ECU 6, braking control for increasing the inter-vehicle distance and a reaction force for depressing the accelerator pedal so as to increase the inter-vehicle distance are increased. Vehicle travel control such as accelerator control is performed.

[Another embodiment]
(1) In the above-described embodiment, the allowable range derivation table for deriving the allowable deviation range 21 based on the search conditions such as the road information 20, the environmental condition, and the driver ID is constructed as the allowable deviation range database 9. . However, it is also possible to use a simpler deviation allowable range derivation table with fewer search conditions.
As such a simple deviation allowable range derivation table, there is a table that derives the deviation allowable range 21 using the curvature radius of the curved road and the vehicle speed as parameters. That is, the curvature radius of the curved road is read from the road information 20, only the vehicle speed is acquired from the in-vehicle sensor ECU 5, and the deviation allowable range 21 is obtained from the curvature radius of the curved road and the vehicle speed. At this time, if the radius of curvature is divided in units of 500 m and the speed is divided in units of 30 km / h, the deviation allowable range derivation table has a small capacity, and its management becomes easy.

(2) In the above-described embodiment, in order to explain the configuration of the driving support ECU 4 in an easy-to-understand manner, as shown in FIG. 3, the functions are blocked according to functions. However, the present invention is limited to the functional blocks partitioned in this way. I don't mean. For example, it may be constructed by combining any of the function units, or each function unit may be further divided. Further, the driving support ECU 4 itself may incorporate all or some of the functions in another ECU.

(3) In the above-described embodiment, the vehicle lateral position and the inter-vehicle distance are individually taken up as the traveling behavior, but both may be adopted at the same time. Further, the distance to the temporary stop position, the distance to the signal, the distance to the intersection, and other driving characteristics may be adopted as the driving behavior.

(4) In the driving support device of the above-described embodiment, information from the car navigation ECU 8 is used, but the presence of the car navigation ECU 8 is not an essential condition of the present invention. Information obtained from the car navigation ECU 8 is based on information from the in-vehicle sensor ECU 5 that processes signals from various sensors mounted on the vehicle, image processing information for an image captured by the camera 10, and communication with other external facilities and devices. It can be replaced by the acquired information.
(5) When the function of the vehicle maintenance support system that performs vehicle control to maintain the target travel line set in the vehicle control ECU 6 is constructed, the target travel line setting unit that sets the target travel line of the own vehicle Alternatively, the vehicle control ECU 6 or the driving support ECU 4 or another configuration may be used. The target travel line set by the target travel line setting unit can be corrected by the travel behavior learning unit 45 that learns the target travel line.

The present invention can be used for a learning-type driving support device that suppresses departure from a target.

10: Camera 2: Camera ECU
3: Image processing ECU
4: Driving support ECU
41: Deviation allowable range setting unit 42: Travel behavior acquisition unit (own vehicle lateral position acquisition unit: inter-vehicle distance acquisition unit)
43: Mind and body information acquisition unit 44: Concentration evaluation unit 45: Travel behavior learning unit (travel position learning unit: inter-vehicle distance learning unit)
46: Learning prohibition unit 47: Driver identification unit 5: In-vehicle sensor ECU
6: Vehicle control ECU (target travel line setting unit)
7: Notification ECU
9: Allowable deviation database 8: Car navigation ECU

Claims

A host vehicle travel behavior acquisition unit for acquiring travel behavior information related to the travel behavior of the host vehicle traveling;
A psychosomatic information acquisition unit for acquiring psychosomatic information relating to a driver's psychosomatic state;
A concentration degree evaluation unit that evaluates a concentration degree related to driving of the driver based on the mind-body information;
A driving behavior learning unit that learns the driving behavior of the driver based on the driving behavior information acquired by the host vehicle driving behavior acquisition unit;
A learning prohibition unit that prohibits the learning when the concentration degree evaluated by the concentration evaluation unit is equal to or less than a determination condition;
A driving support apparatus comprising:
The travel behavior is a position in the travel lane width direction of the host vehicle traveling in the travel lane, and the host vehicle travel behavior acquisition unit is a host vehicle lateral position acquisition unit that acquires the host vehicle lateral position in the width direction of the travel lane. The driving support device according to claim 1, which is configured.
The mind-body information is information based on a driver's face image captured by the in-vehicle camera, and further includes a driver identification unit that identifies the driver based on the driver's face image captured by the in-vehicle camera. The driving support apparatus according to claim 1 or 2.
The driving support device according to claim 3, wherein the mind-body information is gaze information about the driver's gaze or face orientation information about the driver's face orientation.
The deviation allowable range setting part which sets the deviation allowable range with respect to the reference | standard of the said driving behavior is provided, The said driving behavior learning part corrects the said deviation allowable range based on the said driving behavior information. The driving support device according to one item.
When the travel behavior departs from the deviation allowable range, a position holding promotion unit is provided to prompt the driver to return from the deviation, and the position holding promotion unit causes the driver to leave the deviation allowable range. The driving support device according to claim 5, wherein the driving support device is a notifying unit that notifies the vehicle and / or a return control unit that executes vehicle return control so that the vehicle returns to the allowable deviation range.
The target travel line setting part which sets the target travel line of the own vehicle which can be corrected by the said travel behavior learning part is provided, and vehicle control is performed so that the said target travel line may be maintained. The driving support device according to one item.
Allowable range derivation for deriving the allowable displacement range using road information about the road acquired from the car navigation system and / or road information about the road acquired from the processing result of the image taken by the in-vehicle camera as input parameters. The driving support device according to claim 5 or 6, further comprising a table.
The driving support device according to claim 8, wherein any one of a host vehicle traveling speed, a road curvature of a traveling lane, weather, date and time, or a combination thereof is used as the input parameter.
The driving support device according to claim 8 or 9, wherein driver specifying information generated by a driver identifying unit that identifies a driver is used as the input parameter in order to set the deviation allowable range for each driver.