Travel support device

Abstract

Provided is a driving assistance device capable of detecting signs of a driver falling into a state of reduced functionality without causing the driver to feel annoyed, and alerting the driver at an earlier stage. This driving assistance device is provided with: a normal tracking operation learning unit 016 that learns whether the fluctuation tendency of the tracking operation of the driver of a host vehicle with respect to a preceding vehicle is within a normal range, from a feature quantity of at least one of a change in the state of the host vehicle as calculated on the basis of prescribed data detected in time series, and a change in the conditions between the host vehicle and the preceding vehicle; a tracking operation abnormality determination unit 017 that determines that the current tracking operation is abnormal if the fluctuation of the current tracking operation deviates from the normal range by more than a certain amount; and an alarm control unit 018 that indicates an alarm if it is determined that the current tracking operation is abnormal.

Description

Technical Field

[0001] The present invention relates to a driving support device that prevents or reduces a collision with an obstacle that may occur due to driver inattention when the driver of a host vehicle is in a state of mild functional decline.

Background Art

[0002] In this technical field, an invention related to a driver state detection device that detects an abnormality of a driver who has fallen into a state of mild functional decline by a front vehicle detection sensor and an acceleration / deceleration sensor that detects acceleration / deceleration is known (see Patent Document 1 below).

[0003] For example, in Patent Document 1, a driver state detection device that detects an abnormality of a driver, a front vehicle detection sensor that detects a vehicle traveling in front of the host vehicle or on the side and in front of the host vehicle, an acceleration / deceleration sensor that detects the acceleration / deceleration of the host vehicle, an acceleration / deceleration calculation unit that calculates an appropriate acceleration / deceleration for driving the host vehicle to follow the preceding vehicle detected by the front vehicle detection sensor based on an acceleration / deceleration model, the acceleration / deceleration calculated by this acceleration / deceleration calculation unit, and the actual acceleration / deceleration of the host vehicle detected by the acceleration / deceleration sensor, and an abnormality determination unit that compares them to determine the presence or absence of an abnormality of the driver. When another vehicle that may enter between the preceding vehicle being followed and the host vehicle is detected by the front vehicle detection sensor, the abnormality determination unit determines that the driver has an abnormality when the degree of coincidence between the acceleration / deceleration calculated by the acceleration / deceleration calculation unit and the actual acceleration / deceleration of the host vehicle is higher than a predetermined threshold value.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in Patent Document 1, a functional degradation state is detected when the difference between the acceleration / deceleration of the driver and the appropriate acceleration / deceleration for following the preceding vehicle becomes equal to or greater than a predetermined threshold value. Therefore, the functional degradation state can be detected only after a delay occurs in the driver's operation with respect to the preceding vehicle. Therefore, even if an alarm is issued after detecting the functional degradation state when the deceleration of the preceding vehicle is strong, there is a possibility that the time until the driver's function returns to the normal state cannot be secured. Also, although it is possible to detect the functional degradation state earlier by changing the above threshold value to a smaller value, if it is made too early, false detection of the functional degradation state may occur, and the driver may feel annoyed.

[0006] Therefore, in view of the above circumstances, an object of the present invention is to provide a driving support device that can detect a sign of a driver falling into a functional degradation state so that the driver does not feel annoyed and can alarm the driver earlier.

Means for Solving the Problems

[0007] To achieve the above object, a driving support device of the present invention includes a preceding vehicle recognition unit that detects a preceding vehicle traveling ahead of the host vehicle, and at least one feature amount of a change in the state of the host vehicle calculated based on predetermined data detected in time series and a change in the state between the host vehicle and the preceding vehicle. A following operation normal learning unit that learns whether the tendency of fluctuation of the following operation of the driver of the host vehicle with respect to the preceding vehicle is included in a normal range, and a following operation abnormality determination unit that determines that the current following operation is abnormal when there is a difference of a certain amount or more from the normal range in the current fluctuation of the following operation, and an alarm control unit that instructs an alarm when it is determined that the current following operation is abnormal. By providing a driving support device characterized by this, a sign of a driver falling into a functional degradation state is detected, and the driver is alarmed earlier.

Effects of the Invention

[0008] According to the present invention, by detecting a sign of a driver falling into a functional degradation state, the driver can be alarmed earlier, and a collision with an obstacle can be prevented or reduced.

[0009] Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Modes for Carrying Out the Invention

[0011] Hereinafter, modes for carrying out the present invention will be described for a driving support device based on the drawings.

[0012] <Example 1: Example where the learning result is a parameter> FIG. 1 is a configuration diagram of a driving support device according to an embodiment of the present invention.

[0013] [Configuration Explanation] The driving support device 010 of the present embodiment is, for example, an electronic control unit (ECU) mounted on a vehicle (own vehicle) 001 such as a gasoline vehicle, a diesel vehicle, a natural gas vehicle, a hybrid vehicle, an electric vehicle, a fuel cell vehicle, or a hydrogen engine vehicle. Although not shown, the driving support device 010 is composed of, for example, an input / output unit, a central processing unit (CPU), a memory (including both non - volatile memory and volatile memory), and one or more microcontrollers including a timer.

[0014] (Own vehicle 001) The own vehicle 001 is equipped with, for example, an own vehicle sensor 002, an external sensor 003, a car navigation system (CNS) 008, a voice output device 004, an image display device 005, an acceleration device 006, and a deceleration device 007. Although not shown, the own vehicle 001 includes a drive system, a steering system, a braking system, and a control system for driving, turning, decelerating, and stopping the own vehicle 001.

[0015] (Own vehicle sensor 002) The own vehicle sensor 002 includes various sensors for detecting the state of the own vehicle, such as a wheel speed sensor, an acceleration sensor, a shift position sensor, a gyro sensor, a steering angle sensor, a turn signal, etc., and detects the state of the own vehicle including the speed, acceleration, shift position of the transmission, yaw rate, steering angle, turn signal operation state, and abnormalities of the own vehicle 001, and outputs it to the driving support device 010. The abnormalities of the own vehicle 001 detected by the own vehicle sensor 002 include, for example, abnormalities such as tire air pressure, fuel remaining amount, engine, ABS (Anti-Lock Brake System), airbag, brake, hydraulic pressure, battery, water temperature, etc.

[0016] (External sensor 003) The external sensor 003 includes, for example, a millimeter-wave radar that uses reflected waves of radio waves such as millimeter waves, a monocular camera, a stereo camera, a LiDAR (Light Detection and Ranging, Laser Imaging Detection and Ranging) that measures scattered light against pulsed laser irradiation to obtain the distance to an object, an ultrasonic sensor that uses reflected waves of ultrasonic waves, a vehicle-to-roadside communication device, a vehicle-to-vehicle communication device, an illuminance sensor, a raindrop sensor, a humidity sensor, and the like. The external sensor 003 detects, for example, objects around the host vehicle 001 including roads, lane lines, signs, traffic signals, vehicles, pedestrians, obstacles, and the surrounding environment including illuminance, rainfall, humidity, and visibility of obstacles, and outputs the detected information to the driving support device 010. Taking FIG. 3 as an example, the millimeter-wave radar, monocular camera, stereo camera, LiDAR, etc. of the external sensor 003 detect the vehicles 402 and 403 included in the detection range 401 with respect to the host vehicle 400 and output the detected information to the driving support device 010. Note that the detection range 401 is an example.

[0017] (CNS008) CNS008 includes, for example, a map information storage device, a route calculation device, a vehicle-to-vehicle communication device, a vehicle-to-roadside communication device, and a global navigation satellite system (GNSS) receiver. In addition, CNS008 is provided with, for example, an input device for the driver of the host vehicle 001 to input a destination. CNS008 outputs, for example, point information on the map based on the position information of the host vehicle 001, route information from the current location of the host vehicle 001 to the destination, intersection position information that appears from the current value to the destination, curve information such as radius of curvature, gradient information, stop line information, lane width information, traffic signal information, etc. to the driving support device 010.

[0018] (Voice output device 004) The voice output device 004 is, for example, a speaker provided in the passenger compartment of the host vehicle 001, and outputs an alarm sound or voice guidance based on a control signal input from the driving support device 010.

[0019] (Image display device 005) The image display device 005 is, for example, a liquid crystal display device, an organic EL display device, or a head-up display, and displays an image based on a control signal input from the driving support device 010. Note that the image display device 005 may include an input device such as a touch panel or operation buttons. The driver of the host vehicle 001 can input information such as a destination to the CNS 008 via the input device of the image display device 005, for example. Also, via the input device of the image display device 005, the driver may be able to input a determination result as to whether an alarm sound or alarm display output to the audio output device 004 or the image display device 005 is correct. Additionally, it may be possible to select which of a plurality of parameters used in the processing in the driving support device 010 to use.

[0020] (Acceleration device 006) The acceleration device 006 is, for example, an engine or a motor, and accelerates the host vehicle 001 based on a control signal input from the driving support device 010. Also, based on a request for acceleration suppression input from the driving support device 010, it has a function of not accelerating the host vehicle 001 even when the driver steps on the accelerator.

[0021] (Deceleration device 007) The deceleration device 007 is, for example, a brake, and decelerates the host vehicle 001 based on a control signal input from the driving support device 010.

[0022] (Driving support device 010) The driving support device 010 of the present embodiment is mounted on the host vehicle 001 and functions as a driving support device that detects a sign of a state of decline in the function of the driver of the host vehicle 001 and supports the restoration of the driver's function to a normal state. The driving support device 010 is a driving support device including a surrounding environment recognition unit 013, a preceding vehicle recognition unit 014, a following operation abnormality determination suppression unit 015, a following operation normal learning unit 016, a following operation abnormality determination unit 017, an alarm control unit 018, and a speed control unit 019.

[0023] (Surrounding environment recognition unit 013) The surrounding environment recognition unit 013 detects object information and road information around the host vehicle. For example, in terms of object information, it refers to objects such as vehicles (including four-wheel vehicles, two-wheel vehicles, and bicycles), pedestrians, and obstacles traveling in the adjacent lanes of the host vehicle. Also, for example, in terms of road information, it refers to information such as intersection position information, curve information including the radius of curvature, stop line information, lane width information, traffic signal position, and traffic signal state.

[0024] The surrounding environment recognition unit 013 outputs recognition results such as the position, speed, acceleration, and deceleration of surrounding objects and the position of the surrounding road to the following operation abnormality determination suppression unit 015.

[0025] (Leading vehicle recognition unit 014) The leading vehicle recognition unit 014 detects a leading vehicle traveling in front of the host vehicle. Taking FIG. 3 as an example, a vehicle 402 traveling in front of the host vehicle 400 is an example of a leading vehicle. The front refers to the area in the traveling direction within the lane of the host vehicle. Note that the leading vehicle includes four-wheel vehicles, two-wheel vehicles, and bicycles, and refers to an object within a predetermined distance within the detection range 401 in front of the host vehicle.

[0026] The leading vehicle recognition unit 014 outputs recognition results such as the position, speed, and acceleration of the leading vehicle to the following operation abnormality determination suppression unit 015 and the following operation normal learning unit 016.

[0027] (Following operation abnormality determination suppression unit 015) The following-action anomaly determination suppression unit 015 determines whether to suppress the anomaly determination of the following-action anomaly determination unit 017. By this determination, a situation where the anomaly determination by the following-action anomaly determination unit 017 may be incorrect can be excluded from the targets of processing. In other words, it is determined (predicted) whether the current fluctuation of the following action is in a transient state where it is impossible to determine whether it is normal or abnormal. Note that the fluctuation of the following action refers to the relative movement between the host vehicle and the preceding vehicle when the driver of the host vehicle manually follows the preceding vehicle. The relative movement refers to the change in the state of the host vehicle (calculated based on predetermined data) from the host vehicle sensor 002 detected in time series or the change in the state between the host vehicle and the preceding vehicle from the external sensor 003. Specifically, the change in the state of the host vehicle includes changes in acceleration and deceleration and changes in the host vehicle speed. Also, the change in the state between the host vehicle and the preceding vehicle includes changes in relative speed, changes in the inter-vehicle distance, and changes in relative acceleration.

[0028] Generally, when a driver manually follows a preceding vehicle, the driver follows so as to bring the inter-vehicle distance and relative speed to values close to the target values. It is known that fluctuations occur in the following action at that time. For example, the fluctuation of the following action can be regarded as the change in relative speed shown in FIG. 6. In FIG. 6, that the fluctuation of the following action is normal means that the change 501 in relative speed fluctuates up and down within a range of ±5 km / h centered on 0 km / h (5 km / h is an example). Also, that the fluctuation of the following action is abnormal means that there is a case where the change 502 in relative speed exceeds the range of ±5 km / h, and also the number of times the relative speed fluctuates up and down is less compared to the case where the following action is normal.

[0029] A situation where the anomaly determination by the following-action anomaly determination unit 017 may be incorrect refers to, for example, a situation where the preceding vehicle suddenly accelerates or decelerates. This is because the change in the inter-vehicle distance and relative speed differs depending on the acceleration and deceleration of the preceding vehicle, and it can be expected that the driver will perform a following action different from normal. Therefore, since it is difficult to determine whether the following action is normal or abnormal, the anomaly determination is suppressed. Note that the determination of sudden acceleration or deceleration can be made when the deceleration of the preceding vehicle from the preceding vehicle recognition unit 014 is equal to or greater than a predetermined threshold value, or the acceleration is equal to or greater than a predetermined threshold value.

[0030] Also, when it can be determined from the information of the monocular camera or stereo camera of the external sensor 003 whether the blinker of the leading vehicle is operating, it is expected that the leading vehicle will leave its own lane (driving lane) in the future. Therefore, the driver may intentionally accelerate or decelerate the host vehicle, changing the inter-vehicle distance and relative speed, and the driver's movement cannot be uniquely predicted. Accordingly, since it is difficult to determine whether the following operation is normal or abnormal, the abnormality determination is suppressed.

[0031] Also, when there is a sharp curve or steep slope ahead of the leading vehicle, the leading vehicle is likely to decelerate. However, depending on whether the driver recognizes the sharp curve or steep slope ahead, the driver's acceleration / deceleration behavior changes, and the driver's movement cannot be uniquely predicted. Accordingly, since it is difficult to determine whether the following operation is normal or abnormal, the abnormality determination is suppressed. Note that the sharp curve can be determined by whether the radius of curvature of the curve from the monocular camera or stereo camera of the external sensor 003 and CNS008 falls within a predetermined threshold range. The steep slope can be determined by whether the road gradient from CNS008 falls within a predetermined threshold range.

[0032] Also, when it can be determined from the temporary stop line information from CNS008 that there is a temporary stop line ahead of the leading vehicle, the leading vehicle is likely to decelerate. However, depending on whether the driver recognizes the existence of the temporary stop line ahead, the driver's deceleration behavior changes, and the driver's movement cannot be uniquely predicted. Accordingly, since it is difficult to determine whether the following operation is normal or abnormal, the abnormality determination is suppressed.

[0033] Also, when it can be determined from the information of the monocular camera or stereo camera of the external sensor 003 and the lane width information from CNS008 that the lane width ahead of the leading vehicle is decreasing, the leading vehicle is likely to decelerate. However, whether the leading vehicle actually decelerates depends on the awareness of the driver of the leading vehicle. Accordingly, since it is difficult to determine whether the following operation is normal or abnormal, the abnormality determination is suppressed. Note that the decrease in the lane width can be determined by the amount of change in the lane width.

[0034] Also, when it can be determined from the monocular camera or stereo camera of the external sensor 003 that there is a red signal in front of the leading vehicle, it is highly likely that the leading vehicle will decelerate. However, since the driver's deceleration behavior changes depending on whether the driver can recognize the color of the traffic signal ahead, the driver's movement cannot be uniquely predicted. Therefore, since it is difficult to determine whether the following operation is normal or abnormal, the abnormality determination is suppressed.

[0035] Also, as shown in FIG. 4, when it can be determined from the information of the monocular camera or stereo camera of the external sensor 003 that the pedestrian 405 is walking near the own lane, the driver may decelerate regardless of the movement of the leading vehicle 406. Therefore, it is possible that the priority of the driver's following operation changes to avoid colliding with the pedestrian 405. Therefore, since it is difficult to determine whether the following operation is normal or abnormal, the abnormality determination is suppressed.

[0036] Also, as shown in FIG. 3, when it can be determined from the information of the monocular camera or stereo camera of the external sensor 003 that the adjacent vehicle 403 cuts in front of the own vehicle, the driver may decelerate to allow the adjacent vehicle 403 to cut in regardless of the movement of the leading vehicle 402, or the driver may accelerate to prevent the adjacent vehicle 403 from cutting in. Therefore, the driver's movement cannot be uniquely predicted, and since it is difficult to determine whether the following operation is normal or abnormal, the abnormality determination is suppressed. Note that the cut-in determination can be made by determining whether the relative lateral speed of the adjacent vehicle 403 is equal to or greater than a predetermined threshold and approaching the own vehicle. Also, the cut-in determination may be made by determining whether the relative lateral position of the adjacent vehicle 403 is within a predetermined threshold range after a predetermined time.

[0037] Also, when it can be determined from the acceleration / deceleration information of the acceleration sensor of the own vehicle sensor 002 that rapid acceleration or deceleration of the own vehicle has occurred, it is highly likely that an event has occurred that requires the driver to prioritize over the following operation to the leading vehicle. Therefore, since it is difficult to determine whether the following operation is normal or abnormal, the abnormality determination is suppressed. Note that rapid acceleration or deceleration can be determined by the acceleration / deceleration being within a predetermined threshold range.

[0038] Also, when it can be determined from the blinker of the host vehicle sensor 002 that the driver has operated the blinker, it can be predicted that the driver will stop following the preceding vehicle and move to another lane. Therefore, there will be no object in front of the host vehicle. Thus, abnormal determination is suppressed.

[0039] Also, in a situation where there is a road 407 that merges into the traveling lane of the host vehicle as shown in FIG. 5, when there is a merging vehicle 404 on the road 407, regardless of the movement of the preceding vehicle 408, the driver may think that the merging vehicle 404 may merge into the traveling lane of the host vehicle, and there is a possibility that the host vehicle will decelerate. Therefore, the inter-vehicle distance and relative speed with the preceding vehicle 408 may temporarily change, and there is a risk of misjudging that the following operation is abnormal. Thus, in order to prevent misjudgment of abnormal determination of the following operation, abnormal determination is suppressed. Note that it can be determined that it is a merging situation when the relative lateral position of the merging vehicle 404 is within a predetermined threshold value. Further, whether there is a road merging into the front of the host vehicle from the CNS008 may be added to the determination condition of the merging situation.

[0040] Also, when the reliability of the recognition of the preceding vehicle is low, since the accuracy of the inter-vehicle distance and relative speed with the preceding vehicle from the external sensor 003 is likely to be low, abnormal determination is suppressed. Note that the reliability of the recognition of the preceding vehicle can be determined by the fact that the state where the detection of the preceding vehicle has continued from the start of detection is within a predetermined time. Alternatively, it may be determined from the information regarding the reliability of the recognition from the external sensor 003. Alternatively, it may be determined by the fact that there is a sudden change in which the current inter-vehicle distance and relative speed values exceed a predetermined threshold range with respect to the past inter-vehicle distance and relative speed values.

[0041] That is, the following operation abnormal determination suppression unit 015 predicts a transient state in which it cannot be determined whether the current fluctuation of the following operation is normal or abnormal, and suppresses the abnormal determination of the following operation abnormal determination unit 017. The transient state is a state in which sudden acceleration or deceleration of the preceding vehicle has occurred, a state in which the blinker of the preceding vehicle has operated, a state in which there is a sharp curve in front of the preceding vehicle, a state in which there is a sharp slope in front of the preceding vehicle A state where a stop line exists in front of the leading vehicle, A state where there is a red traffic light in front of the leading vehicle, A state where the lane width in front of the leading vehicle is decreasing, A state where a pedestrian is walking near the own lane, A state where it can be predicted that an adjacent vehicle will cut in in front of the own vehicle, A state where sudden acceleration or deceleration of the own vehicle has occurred, A state where the driver of the own vehicle has operated the turn signal, A state where an object that can be predicted to move forward exists in the traveling lane of the own vehicle, Including any one or more of the states where the reliability of the recognition of the leading vehicle is decreasing.

[0042] (Following operation normal learning unit 016) The following operation normal learning unit 016 learns the normal range of the following operation when the driver of the own vehicle is following the leading vehicle. And, based on the learned normal range of the following operation, if the current following operation is within the normal range, it is determined that the current following operation is normal. That is, the following operation normal learning unit 016 learns whether the fluctuation tendency of the driver's following operation with respect to the leading vehicle is included in the normal range. Note that the normal range refers to the following operation in a state where the driver of the own vehicle is following the leading vehicle while paying attention to driving.

[0043] In this embodiment, it is assumed that the normal range of the following operation stores a predetermined threshold value learned in advance on board in the non-volatile memory of the driving support device 010 as a parameter, and reads out and uses the parameter when the driving support device 010 is started. Then, the current following operation of the own vehicle is calculated from the change in the state of the own vehicle and the change in the state between the own vehicle and the leading vehicle, and compared with the learned normal range parameter to determine whether the following operation is normal.

[0044] To calculate the current following operation, for example, a method of obtaining the frequency, amplitude, and bandwidth representing the tendency of the change in the relative speed is used. The calculation method is as follows.

[0045] First, since there is noise in the relative speed from the external sensor 003, as preprocessing, noise is removed by passing only the low-frequency components through a filter process such as a moving average. Note that the filter method may be switched to another method (band-pass filter, high-pass filter, etc.) according to the specifications of the input information such as the external sensor 003. Also, when the vehicle speed is slow compared to when it is fast, it may be easier to start or change the driving route, and it is considered that the probability of sudden acceleration or deceleration of the host vehicle is high. Therefore, when the assumed acceleration or deceleration according to the host vehicle speed is equal to or higher than a predetermined threshold, it may be regarded as a temporary deviation operation in the following operation, and a filter process that can exclude the change in the relative speed at that time may be performed. Note that not only the host vehicle, but also a filter process that excludes the change in the relative speed in a temporary deviation operation in the following operation according to the speed and acceleration / deceleration of the preceding vehicle may be performed.

[0046] Next, when obtaining the frequency, amplitude, and bandwidth representing the tendency of the change in relative speed, it is necessary to specify the range of the target data. In this embodiment, time-series data obtained by extracting the relative speed values from a predetermined time ago (past) to the present is acquired. Note that since there are more disturbance factors that interfere with the following operation of the host vehicle on general roads than on highways, it is considered that the time for which the following operation continues varies depending on the speed range of the host vehicle. The disturbance factors refer to factors such as a vehicle cutting in on the host vehicle lane or a pedestrian crossing, which temporarily prevent the host vehicle from continuing the following operation. For example, when the host vehicle speed is slow, it is often the case that the vehicle is driving on a general road, and there are many disturbance factors that interfere with the following operation of the host vehicle, so the following operation continues for a shorter time than on a highway. Taking advantage of such characteristics, the predetermined time may be changed according to the host vehicle speed. Alternatively, the predetermined time may be changed according to the road on which the vehicle is driving using the information from CNS008 on whether the vehicle is driving on a general road or a highway. Also, the predetermined time is preferably set to a relatively long time of several tens of seconds to several minutes so that the following operation can be captured.

[0047] Then, based on the obtained time-series data of the relative speed, perform a fast Fourier transform (FFT) or the like to perform frequency conversion and calculate the amplitude spectrum. For example, assume an amplitude spectrum 701 as shown in FIG. 7. The amplitude spectrum 701 has a frequency F1 with a peak at amplitude P1 and has a bandwidth 703. If the amplitude P1 is within a predetermined threshold, it is assumed that the amplitude is within the normal range. Also, if the frequency F1 is within the range of the predetermined threshold, it is assumed that the frequency is within the normal range. Further, if the bandwidth 703 is within the predetermined threshold, it is assumed that the bandwidth is within the normal range. Note that when all of these conditions are satisfied, it is assumed that the following operation by the driver is within the normal range. However, depending on the time-series data targeted, it is not always the case that a peak value occurs only at the frequency F1 as shown in FIG. 7. Therefore, when a plurality of amplitudes of the same level and a plurality of different frequencies are represented in the frequency spectrum, the normal determination of the following operation may be suppressed.

[0048] Finally, if the following operation being within the normal range continues for a predetermined time, it is determined that the following operation is normal.

[0049] Here, the relative speed is used as an example of the change in the relative relationship between the preceding vehicle and the host vehicle. However, the same fast Fourier transform operation may be performed using the inter-vehicle distance or THW (Time Headway) obtained by dividing the inter-vehicle distance by the host vehicle speed.

[0050] That is, the following operation normal learning unit 016 calculates any one or more of the frequency, amplitude, or bandwidth by performing frequency conversion on the change in the relative relationship (relative speed, inter-vehicle distance, THW, etc.) between the preceding vehicle and the host vehicle, and learns as normal (included in the normal range) when any one or more of the calculated frequency, amplitude, or bandwidth are within the range of the predetermined threshold.

[0051] Also, when a driver generally follows a leading vehicle, it is known that the THW is approximately 2 seconds on average. Therefore, apart from the method of frequency conversion, a range of 2 seconds ± a predetermined threshold may be set as the normal range, and if the average value of the THW falls within the range of the predetermined threshold, it may be determined that the following operation is normal. The method of determining the predetermined threshold here may be to cut out the portion where the following operation is within the normal range from the time-series data on board and use the standard deviation obtained by calculating and analyzing the THW. Note that the average THW is approximately 2 seconds in a situation where the road is not congested, but it can be expected that the average THW will be shorter than 2 seconds when the road is congested. This is based on the driver's intention to reduce the inter-vehicle distance so that the adjacent vehicle does not cut into the own lane when the adjacent vehicle cuts in front of the own vehicle during congestion. Therefore, when the road is congested, the parameter of the average THW may be set to a value shorter than 2 seconds, such as approximately 1.5 seconds. Also, since these average values of the THW vary depending on individual differences, 2 seconds and 1.5 seconds are just examples.

[0052] Note that the object of calculating the average value of the change in the relative relationship between the leading vehicle and the own vehicle may be the relative speed or the inter-vehicle distance.

[0053] That is, the following operation normal learning unit 016 may learn that it is normal (included in the normal range) when the average value of the change in the relative relationship (relative speed, inter-vehicle distance, THW, etc.) between the leading vehicle and the own vehicle is within the range of a predetermined threshold.

[0054] In addition, the parameter of the predetermined threshold value representing the normal range of the following operation may be obtained from the cloud via the Internet for the parameters learned not only by the driver of the host vehicle but also by drivers across the country, and stored in the non-volatile memory in the driving support device 010 of the host vehicle. That is, the following operation normal learning unit 016 may obtain the learned parameters (representing the normal range) from the cloud via the Internet (for each driver). The learned parameters obtained from the cloud are not single, and it is advisable to select those that match the driver's conditions from among a plurality of parameters. For example, it may be determined based on information scoring the driving behavior of the driver, such as the driver's age, driving experience, number of accidents, and presence or absence of sudden acceleration or deceleration in normal times. Also, the driver may be able to select the learned parameters that suit him / herself from the touch panel or operation buttons of the image display device 005.

[0055] (Following operation abnormality determination unit 017) Based on the normal range and normal determination of the following operation normal learning unit 016, the following operation abnormality determination unit 017 determines whether the current following operation is abnormal when the driver of the host vehicle is following the preceding vehicle. In the following operation abnormality determination unit 017, when the state changes from the state where the normal determination of the following operation normal learning unit 016 has been made to a state deviating from the normal range, the current following operation of the driver is determined to be abnormal. The state deviating from the normal range refers to a state where the fluctuation of the current following operation has a difference of a certain amount or more with respect to the normal range. That is, the following operation abnormality determination unit 017 determines that the current following operation is abnormal when the fluctuation of the current following operation has a difference of a certain amount or more with respect to the normal range.

[0056] Whether it is in a state deviating from the normal range is determined as follows.

[0057] If the amplitude P2 of the amplitude spectrum 702 in FIG. 7 is greater than a predetermined threshold value, it is assumed that the amplitude has deviated from the normal range. Also, if the frequency F2 is outside the range of the predetermined threshold value, it is assumed that the frequency has deviated from the normal range. Further, if the bandwidth is greater than the predetermined threshold value, it is assumed that the bandwidth has deviated from the normal range. If any of these conditions is satisfied, it is assumed that the following operation by the driver has deviated from the normal range. Also, if the fact that the following operation has deviated from the normal range continues for a predetermined time, it is determined that the following operation is abnormal. In other words, when the fluctuation of the current following operation has a difference of a certain amount or more from the normal range for a predetermined time, it is determined that the following operation is abnormal. Note that these normal ranges may have a margin with respect to the predetermined threshold value in order to perform hysteresis processing so as not to frequently repeat normal and abnormal states.

[0058] Also, when following a preceding vehicle at the same own vehicle speed for a long time, the following operation is likely to become abnormal due to driver fatigue or the like. Therefore, a condition for determining the abnormality of the following operation may be added when a state where the speed change from the own vehicle speed at the time when the following operation is determined to be normal is within a predetermined range continues for a predetermined time.

[0059] (Warning control unit 018) When the abnormality of the current following operation is determined by the following operation abnormality determination unit 017, the warning control unit 018 requests (instructs) the voice output device 004 to sound a warning sound. The warning sound may be a beep sound or a voice guidance notifying the driver of a functional decline. Also, the warning control unit 018 requests (instructs) the image display device 005 to display a warning.

[0060] The warning control unit 018 may be implemented to request the sounding of a warning sound or the display of a warning when the determination of the abnormality of the following operation continues for a predetermined time after the abnormality of the following operation is determined in order to suppress a false alarm.

[0061] Although an alarm sound is emitted and an alarm is displayed, if the driver is actually not in a state of reduced functionality, the driver operates the image display device 005 to correct the error of the following operation abnormality determination unit 017. For example, an operation button for selecting the correctness of the alarm can be arranged on the display of the alarm, and a method can be considered in which the driver operates and selects whether the alarm is correct or incorrect.

[0062] If it is selected that the alarm is incorrect, since it is a false alarm, the normal learning range of the following operation in the following operation normal learning unit 016 is changed to unlearned. Further, after changing to unlearned, the driver may operate the image display device 005 again so that a parameter different from that at the time of false alarm can be selected from a plurality of parameters representing the normal range in the following operation normal learning unit 016. Another parameter is, for example, to be able to specify the tendency of the fluctuation of the driver's own following operation from large, medium, and small fluctuations of the following operation. When the driver resets the parameter, the normal range is corrected and learned in the following operation normal learning unit 016, and the following operation abnormality determination unit 017 becomes capable of performing abnormality determination again. That is, the following operation normal learning unit 016 corrects the learned normal range based on the response from the driver.

[0063] When the warning control unit 018 starts a warning, if a predetermined time has elapsed since the condition for starting the warning no longer holds, the warning is canceled. Alternatively, if the driver's driving intention can be confirmed since the condition for starting the warning no longer holds, the warning is canceled. When a warning has been started once (in other words, when the warning control unit 018 is instructing a warning), since the driver may still be in a state of reduced function afterwards, a measure is required to reduce the cancellation of false warnings. For example, after the abnormality of the following operation is confirmed and the warning is started, when a predetermined time has elapsed since the following operation is no longer abnormal in the following operation abnormality determination unit 017 (in other words, when a predetermined time has elapsed since the abnormality in the following operation abnormality determination unit 017 has been resolved), the warning is canceled. Also, when the preceding vehicle no longer exists during a warning request, the warning is canceled when a predetermined time has elapsed since the preceding vehicle no longer exists. Also, when a predetermined time has elapsed since the start of suppression in the following operation abnormality determination suppression unit 015 after the warning has been started, the warning is canceled. Also, when a predetermined time has elapsed since the learning in the following operation normal learning unit 016 has been reset after the warning has been started, the warning is canceled. Also, after the warning has been started, when a predetermined time has elapsed since the driver has operated the accelerator or brake or blinker or steering while the following operation is determined to be normal in the following operation abnormality determination unit 017 (in other words, in a state where the abnormality in the following operation abnormality determination unit 017 has been resolved), the warning is canceled. Note that the cancellation condition may be applied to the speed control in the subsequent speed control unit 019.

[0064] That is, in a state where the warning control unit 018 is instructing a warning, when the abnormality in the following operation abnormality determination unit 017 has been resolved, or when the start of suppression in the following operation abnormality determination suppression unit 015 has been started, or when the learning in the following operation normal learning unit 016 has been reset, or when a predetermined time has elapsed since the start of operation of the accelerator or brake or blinker or steering by the driver in a state where the abnormality in the following operation abnormality determination unit 017 has been resolved, the warning control unit 018 cancels the warning.

[0065] (Speed control unit 019) After the alarm control unit 018 outputs an alarm, if the driver does not change the driving behavior, the speed control unit 019 requests acceleration suppression for the acceleration device 006. When acceleration suppression is requested, the host vehicle 001 does not accelerate even if the driver presses the accelerator pedal. Thereby, the possibility of colliding with the preceding vehicle when the driver erroneously presses the accelerator pedal in a state of reduced function can be reduced.

[0066] Alternatively, the speed control unit 019 requests deceleration for the deceleration device 007. When deceleration is requested, the host vehicle 001 automatically starts decelerating. Thereby, the possibility of colliding with the preceding vehicle when the driver's braking operation is delayed in a state of reduced function can be reduced. Note that the requested deceleration may be calculated based on an equal acceleration linear motion model so that THW does not become less than a predetermined value.

[0067] Whether the driver changes the driving behavior or not is determined by whether the output of the alarm has elapsed for a predetermined time. Alternatively, it is determined by whether the collision risk degree with the preceding vehicle has become equal to or higher than a predetermined threshold value. The collision risk degree may use TTC (Time to Collision) or THW as an index.

[0068] [Flowchart description] FIG. 2 is a flowchart showing an example of a driving support routine executed in the driving support device 010. This flowchart is repeatedly executed at a predetermined cycle by a CPU provided in an ECU that realizes the driving support device 010.

[0069] When starting the driving support routine, at step 300, the front vehicle recognition unit 014 determines whether a front vehicle exists. If it is determined that no front vehicle exists, the abnormality determination of the following operation is not performed, and the process proceeds to step 316. In addition, when an abnormality has occurred in the host vehicle 001, it may also proceed to step 316. If it is determined that a front vehicle exists, at step 302, the following operation abnormality determination suppression unit 015 determines whether the condition for suppressing the abnormality determination of the following operation is satisfied. If it is determined that the abnormality determination of the following operation is suppressed, the abnormality determination of the following operation is not performed, and the process proceeds to step 316.

[0070] If it is determined that the abnormality determination of the following operation is not suppressed, at step 305, the following operation normal learning unit 016 determines whether the normal range has already been learned. If the normal range has not been learned, at step 312, the normal range of the following operation is learned.

[0071] Then, in order to be able to use the learned normal range even when the power of the driving support device 010 is turned off, at step 313, the learning result of the normal range is recorded. Note that the learning result is stored in the non-volatile memory of the driving support device 010. The stored learning result is read from the non-volatile memory at the next startup of the host vehicle when the power of the driving support device 010 is turned on.

[0072] Alternatively, the learning result may be stored in a cloud server via the Internet. The stored learning result is read from the cloud server at the next startup of the host vehicle when the power of the driving support device 010 is turned on, and stored in the non-volatile memory within the driving support device 010.

[0073] That is, the following operation normal learning unit 016 stores the normal range learned for each driver (in the non-volatile memory or in the cloud server), and reads the stored normal range of the corresponding driver (from the non-volatile memory or from the cloud server) at the next startup of the host vehicle when the power of the driving support device 010 is turned on.

[0074] After the recording in step 313, proceed to step 316.

[0075] If the normal range has been learned, in step 306, the follow-up operation normal learning unit 016 evaluates the current follow-up operation. The follow-up operation calculated by a method such as the fast Fourier transform (FFT) described above for the current follow-up operation is compared with the learned normal range, and it is determined whether the current follow-up operation is normal (i.e., included in the normal range).

[0076] Next, in step 307, the follow-up operation abnormality determination unit 017 determines whether the previous follow-up operation was normal. If the previous follow-up operation was not normal, proceed to step 316. If the previous follow-up operation was normal, in step 308, the current follow-up operation is compared with the learned normal range, and it is determined whether there is a difference of a certain amount or more between the fluctuation of the current follow-up operation and the learned normal range. If there is no difference of a certain amount or more, proceed to step 316. If there is a difference of a certain amount or more, it is determined that the current follow-up operation is abnormal, and proceed to step 309.

[0077] In the alarm control unit 018, in step 309, an alarm is instructed to operate. In the alarm control unit 018, in step 316, it is determined whether the alarm release condition is satisfied. If the alarm release condition is not satisfied, do nothing and end the routine once. If the alarm release condition is satisfied, in step 317, the alarm is released. Then, end the routine once.

[0078] When the routine ends, the CPU executes from step 300 in the next cycle.

[0079] [Description of Effects] As described above, by detecting an abnormality in the driver's follow-up operation and giving an alarm or suppressing acceleration or deceleration, the possibility of a collision can be reduced.

[0080] <Example 2: Example of Learning> The configuration diagram of the driving support device according to an embodiment of the present invention is the same as that in the first embodiment, and will only be described below when different processes are performed for each configuration in FIG. 1.

[0081] [Configuration Explanation] In addition to the content of the first embodiment, the host vehicle sensor 002 includes, for example, a driver monitor which is a camera that captures the passenger compartment. In the driver monitor, for example, it detects that the driver is in a state of reduced function from the driver's line of sight, face orientation, expression, etc., and outputs it to the driving support device 010. The state of reduced function refers to a state where the driver is driving absent-mindedly or is sleepy.

[0082] In addition to the content of the first embodiment, the driving support device 010 has, for example, a non-volatile memory for storing teacher data. Teacher data refers to learning data that serves as the correct answer for supervised learning in machine learning.

[0083] When the power is turned on, the driving support device 010 always collects time-series data including a portion that falls within the normal range of the following operation acquired from the host vehicle sensor 002 and the external sensor 003 during normal driving, and stores a plurality of time-series data as teacher data in the memory. By setting a narrowing-down condition in advance during collection, it is possible to reduce the amount of memory used.

[0084] The narrowing-down condition is, for example, a condition where the risk of collision between the host vehicle and the preceding vehicle is equal to or lower than a certain threshold value (a non-dangerous situation). Note that the risk of collision is the same as the calculation method mentioned in the description of the speed control unit 019 in the first embodiment. For example, when the TTC is greater than a predetermined time, the following operation is likely to be within the normal range.

[0085] Furthermore, it may also include a condition that the driver is not performing all of sudden acceleration / deceleration, sudden steering, turn signal operation, and shift operation.

[0086] Also, the following method may be included as a condition.

[0087] First, after the driver is driving and the above narrowing conditions are met and a predetermined time has elapsed (after time-series data of about several seconds to several minutes has been accumulated), the voice output device 004 or the image display device 005 outputs a voice or display to confirm to the driver whether the current driving is a following operation within the normal range. Then, the driver hears or visually confirms the voice or display, and if it is a correct following operation within the normal range, the driver presses the correct answer operation button on the image display device 005 to convey to the driving support device 010 that it is the correct answer. When the correct answer operation button is pressed, assuming that permission has been obtained from the driver to use the time-series data as teacher data, the driving support device 010 stores the time-series data in the memory as teacher data. That is, it may include the condition that it has been confirmed to the driver that the current driving of the host vehicle is a normal following operation of the driver with respect to the preceding vehicle.

[0088] Also, when the above narrowing conditions are met and the driver is not in a state of reduced function as determined by the driver monitor of the host vehicle sensor 002 (in other words, the driver's reduced function is above a certain value), the driving support device 010 stores the time-series data in the memory as teacher data.

[0089] Also, the above narrowing conditions are obtained from the cloud, and the driving support device 010 may store the time-series data that satisfies the narrowing conditions in the memory as teacher data.

[0090] Note that the teacher data may be uploaded to the cloud.

[0091] Also, the narrowing conditions may include the following method as a condition.

[0092] First, when the driver starts driving, time-series data is collected and uploaded to the cloud. Next, after the driver finishes driving, the driver accesses the cloud from an external terminal such as a smartphone or a personal computer, and selects teacher data by specifying, on the external terminal, the time-series data in which the driver is not in a state of functional decline. The selected teacher data is acquired by the driving support device 010 from the cloud at the timing when the power of the next driving support device 010 is turned on, and the driving support device 010 stores the acquired teacher data in the memory.

[0093] Also, before the driver starts driving, a method may be adopted in which the driver selects, on the external terminal, the time zone, location, and upload conditions for uploading to the cloud in advance. The upload conditions assume conditions such as that the collision risk degree between the host vehicle and the preceding vehicle is below a certain threshold value (a non-dangerous situation), or that the driver is not performing all of sudden acceleration / deceleration, sudden steering, turn signal operation, and shift operation.

[0094] (Following operation normal learning unit 016) In the following operation normal learning unit 016, using a method such as deep learning with the time-series data acquired from the host vehicle sensor 002 and the external world sensor 003 as inputs and taking the plurality of teacher data stored in the memory as correct answers, the normal range of the following operation is learned.

[0095] That is, the following operation normal learning unit 016 uses, as teacher data (learning data that becomes the correct answer), at least one of the feature amounts when the collision risk degree with the preceding vehicle is below a certain value (a non-dangerous situation), the feature amount when the driver of the host vehicle is not performing all of sudden acceleration / deceleration, sudden steering, turn signal operation, and shift operation, or the feature amount when the driver of the host vehicle can confirm the correctness that the current driving of the host vehicle is a normal following operation of the driver of the host vehicle with respect to the preceding vehicle, and learns the normal range of the following operation.

[0096] Also, in the following behavior normal learning unit 016, from a plurality of teacher data stored in the memory, the frequency, amplitude, and bandwidth of the relative speed are calculated by a method using a fast Fourier transform (FFT) or the like to calculate the fluctuation of the following behavior in the following behavior normal learning unit 016 described in the first embodiment. These numerical values are read as teacher data, and using a method such as deep learning, the normal range of the frequency, amplitude, and bandwidth of the relative speed, in other words, a predetermined range of any one or more of the frequency, amplitude, and bandwidth of the relative speed is learned.

[0097] Also, in the following behavior normal learning unit 016, the average values of THW, relative speed, inter-vehicle distance, etc. are calculated from a plurality of teacher data stored in the memory, these numerical values are read as teacher data, and the normal range of the average values of THW, relative speed, inter-vehicle distance, etc., in other words, the predetermined threshold values of the average values of THW, relative speed, inter-vehicle distance, etc. may be learned.

[0098] That is, the following behavior normal learning unit 016 uses at least one of the feature quantities when the collision risk degree with the preceding vehicle is below a certain value (a situation that is not dangerous), the feature quantity when none of the sudden acceleration / deceleration, sudden steering, turn signal operation, and shift operation of the driver of the host vehicle is being performed, or the feature quantity when it can be confirmed to the driver of the host vehicle that the current driving of the host vehicle is a normal following behavior of the driver of the host vehicle with respect to the preceding vehicle as teacher data (correct learning data), and learns a predetermined threshold value of any one or more of the frequency, amplitude, or bandwidth, or a predetermined threshold value of the average value of the change in the relative relationship (relative speed, inter-vehicle distance, THW, etc.) between the preceding vehicle and the host vehicle.

[0099] The following behavior normal learning unit 016 narrows down the teacher data under the condition that the driver's functional decline is above a certain value (not in a functional decline state) by the driver monitor, and learns the normal range of the following behavior. Alternatively, it obtains the narrowing-down condition of the teacher data from the cloud, narrows down the teacher data based on the narrowing-down condition, and learns the normal range of the following behavior. Alternatively, it narrows down the teacher data based on any one of the time, location, and conditions specified by the driver, and learns the normal range of the following behavior.

[0100] Further, the following-following operation normal learning unit 016 learns the normal range of the following-following operation from the teacher data acquired from the cloud.

[0101] [Description of Effects] As described above, by efficiently preparing the teacher data and using it for learning the normal range of the driver's following-following operation, it is possible to automate or semi-automate the response to the driver's habits and characteristics, and reduce false alarms in the state of the driver's functional decline.

[0102] [Embodiment 3: Abnormality determination of following-following operation by detecting wobbling based on lane line information] The configuration diagram of the driving support device according to an embodiment of the present invention is the same as that in Embodiment 1, and only the cases where different processes are performed in each configuration in FIG. 1 are described below.

[0103] [Configuration Explanation] (Following-following operation normal learning unit 016) The following-following operation normal learning unit 016 has a lane-keeping operation normal learning unit, and learns the normal range of the fluctuation of the lane-keeping operation when the driver of the host vehicle is maintaining within the lane.

[0104] The fluctuation of the lane-keeping operation refers to the relative movement between the host vehicle and the lane line when the driver of the host vehicle manually maintains within the lane. The relative movement in this embodiment refers to the change in the state of the host vehicle (calculated based on predetermined data) from the host vehicle sensor 002 detected in time series or the change in the state between the host vehicle and the lane line from the external sensor 003. Specifically, the change in the state of the host vehicle includes changes in the steering angle and yaw rate. Further, the change in the state between the host vehicle and the lane line includes a change in the relative lateral position with respect to the lane line.

[0105] Then, based on the learned normal range of the lane-keeping operation, the following-following operation normal learning unit 016 determines that it is normal if the current lane-keeping operation is within the normal range. The normal range refers to the lane-keeping operation in a state where the driver of the host vehicle is maintaining within the lane while paying attention to driving.

[0106] In this embodiment, it is assumed that the normal range of the in-lane maintenance operation is stored in the non-volatile memory of the driving support device 010 with a predetermined threshold value learned in advance on a computer as a parameter, and the parameter is read out and used when the driving support device 010 is started. Then, the current in-lane maintenance operation of the host vehicle is calculated from the change in the state of the host vehicle and the change in the state between the host vehicle and the lane marking, and compared with the learned normal range parameter to determine whether the in-lane maintenance operation is normal.

[0107] In order to calculate the current in-lane maintenance operation, for example, a method of obtaining the frequency, amplitude, and bandwidth representing the tendency of the change in the relative lateral position with respect to the lane marking is adopted. Since the calculation method uses a technique such as fast Fourier transform (FFT) in the same way as the following operation method in the first embodiment, the details are omitted. Also, similar to the following operation method in the first embodiment, it may be determined using the average value of the change in the relative lateral position with respect to the lane marking.

[0108] That is, the in-lane maintenance operation normal learning unit 016 (the in-lane maintenance operation normal learning unit of the following operation normal learning unit) calculates one or more of the frequency, amplitude, or bandwidth by frequency-converting the change in the relative relationship between the host vehicle and the lane marking (such as the change in the relative lateral position with respect to the lane marking), and when one or more of the calculated frequency, amplitude, or bandwidth are within the range of the predetermined threshold value, or when the average value of the change in the relative relationship between the host vehicle and the lane marking (such as the change in the relative lateral position with respect to the lane marking) is within the range of the predetermined threshold value, it learns as normal (included in the normal range).

[0109] When the power of the driving support device 010 is turned on, it always collects time-series data including a part that becomes the normal range of the in-lane maintenance operation obtained by the host vehicle sensor 002 and the external sensor 003 during the normal driving of the driver, and stores a plurality of time-series data in the memory as teacher data. By setting a narrowing condition in advance at the time of collection, it is possible to reduce the amount of memory used.

[0110] The narrowing-down condition may be, for example, a condition in which the possibility of deviation between the host vehicle and the lane line is equal to or less than a certain threshold value (a non-dangerous situation). Note that, for example, when the time until the host vehicle's TTLC (Time To Line Crossing) deviates from the lane line is longer than a predetermined time, the following operation is likely to be within the normal range. Note that TTLC can be calculated by dividing the lateral speed of the host vehicle by the relative lateral position of the lane line.

[0111] Regarding other narrowing-down conditions, since they are the same as the following operation in the second embodiment, the details are omitted.

[0112] The in-lane maintenance operation normal learning unit uses, as correct answers, a plurality of teacher data stored in the memory in the same manner as the following operation normal learning unit 016, and inputs time-series data acquired from the host vehicle sensor 002 and the external sensor 003, and learns the normal range of the following operation using a method such as deep learning.

[0113] That is, the in-lane maintenance operation normal learning unit of the following operation normal learning unit 016 uses, as teacher data (learning data that becomes the correct answer), a feature amount in which the possibility of deviation from the change in the relative relationship between the host vehicle and the lane line (such as the change in the relative lateral position with respect to the lane line) is equal to or less than a certain value (a non-dangerous situation), and learns the normal range of the following operation.

[0114] [Description of Effects] As described above, regardless of the presence or absence of the preceding vehicle, the state of the driver's functional decline can be detected, so that the applicable range can be expanded.

[0115] <Summary of Embodiments 1 to 3> As described above, the driving support device 010 according to the present embodiment includes a preceding vehicle recognition unit 014 that detects a preceding vehicle traveling in front of the host vehicle, and at least one of a change in the state of the host vehicle calculated based on predetermined data detected in time series and a change in the state between the host vehicle and the preceding vehicle. A following operation normal learning unit 016 that learns whether the tendency of fluctuation of the driver's following operation with respect to the preceding vehicle is within a normal range, and a following operation abnormality determination unit 017 that determines that the current following operation is abnormal when there is a difference of a certain amount or more from the normal range in the fluctuation of the current following operation, and a following operation abnormality determination suppression unit 015 that predicts a transient state in which the fluctuation of the current following operation cannot be determined as normal or abnormal and suppresses the determination of the following operation abnormality determination unit 017, and an alarm control unit 018 that instructs an alarm when the current following operation is determined to be abnormal.

[0116] That is, the driving support device 010 according to the present embodiment detects the tendency of fluctuation (fluctuation different from normal) of following driving with respect to the preceding vehicle, detects the state of deterioration of the driver's function, and issues an alarm.

[0117] According to the present embodiment, by detecting a sign of a driver falling into a state of functional decline, it is possible to alarm the driver earlier and prevent or reduce a collision with an obstacle.

[0118] Note that the present invention is not limited to the above-described embodiments, and includes various modifications. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described.

[0119] In addition, each of the above-described configurations, functions, processing units, processing means, etc. may be realized in hardware by designing a part or all of them, for example, by means of an integrated circuit. Further, each of the above-described configurations, functions, etc. may be realized in software by a processor interpreting and executing a program for realizing each function. Information such as a program, a table, a file, etc. for realizing each function can be placed in a memory, a storage device such as a hard disk or an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or a DVD.

[0120] Also, control lines and information lines show those considered necessary for explanation, and not necessarily all control lines and information lines are shown on the product. In fact, it may be considered that almost all configurations are interconnected.

Explanation of Reference Numerals

[0121] 001 Vehicle (own vehicle) 002 Own vehicle sensor 003 External sensor 004 Voice output device 005 Image display device 006 Acceleration device 007 Deceleration device 008 Car navigation system (CNS) 010 Driving support device 013 Surrounding environment recognition unit 014 Recognition unit for preceding vehicle 015 Follow-up operation abnormality determination suppression unit 016 Follow-up operation normal learning unit 017 Follow-up operation abnormality determination unit 018 Alarm control unit 019 Speed control unit

Claims

1. A preceding vehicle recognition unit that detects a preceding vehicle traveling in front of the host vehicle, a following operation normal learning unit that learns whether the tendency of fluctuations in the following operation of the driver of the host vehicle with respect to the preceding vehicle is included in a normal range from at least one feature amount of a change in the state of the host vehicle calculated based on predetermined data detected in time series and a change in the state between the host vehicle and the preceding vehicle, a following operation abnormality determination unit that determines that the current following operation is abnormal when there is a difference of a certain amount or more from the normal range in the current fluctuations of the following operation, and an alarm control unit that instructs an alarm when it is determined that the current following operation is abnormal, wherein the traveling support device is characterized by comprising the same.

2. In the traveling support device according to Claim 1, the following operation normal learning unit calculates any one or more of frequency, amplitude, or bandwidth by frequency-converting a change in the relative relationship between the preceding vehicle and the host vehicle, and learns as normal when any one or more of the calculated frequency, amplitude, or bandwidth are within a range of a predetermined threshold value. The traveling support device is characterized by this.

3. In the traveling support device according to Claim 1, the following operation normal learning unit learns as normal when an average value of a change in the relative relationship between the preceding vehicle and the host vehicle is within a range of a predetermined threshold value. The traveling support device is characterized by this.

4. In the traveling support device according to Claim 1, the following operation normal learning unit uses, as teacher data, at least one feature amount from the feature amounts, a feature amount in which the risk of collision with the preceding vehicle becomes a certain value or less, a feature amount when all of sudden acceleration / deceleration, sudden steering, turn signal operation, and shift operation of the driver of the host vehicle are not performed, or a feature amount when it can be confirmed by the driver of the host vehicle that the current traveling of the host vehicle is a normal following operation of the driver of the host vehicle with respect to the preceding vehicle, and learns the normal range. The traveling support device is characterized by this.

5. In the traveling support device according to Claim 2, the following operation normal learning unit uses, as teacher data, at least one feature amount from the feature amounts, a feature amount in which the risk of collision with the preceding vehicle becomes a certain value or less, a feature amount when all of sudden acceleration / deceleration, sudden steering, turn signal operation, and shift operation of the driver of the host vehicle are not performed, or When it is possible to confirm for the driver of the host vehicle that the current driving of the host vehicle is a normal following operation of the driver of the host vehicle with respect to the preceding vehicle, at least one feature amount of the feature amounts is used as teacher data, and the running support device is characterized in that it learns any one or more of the predetermined threshold values of the frequency, the amplitude, or the bandwidth.

6. In the running support device according to claim 3, the following operation normal learning unit, from the feature amount, a feature amount in which the collision risk degree with the preceding vehicle becomes equal to or less than a certain value, a feature amount when all of sudden acceleration / deceleration, sudden steering, turn signal operation, and shift operation of the driver of the host vehicle are not performed, or When it is possible to confirm for the driver of the host vehicle that the current driving of the host vehicle is a normal following operation of the driver of the host vehicle with respect to the preceding vehicle, at least one feature amount of the feature amounts is used as teacher data, and the running support device is characterized in that it learns the predetermined threshold value of the average value of the change in the relative relationship between the preceding vehicle and the host vehicle.

7. In the running support device according to claim 1, the following operation normal learning unit stores the normal range learned for each driver and reads out the stored normal range of the corresponding driver when the host vehicle is started next time. The running support device is characterized by this.

8. In the running support device according to claim 4, the following operation normal learning unit, under the condition that the function deterioration of the driver is equal to or more than a certain value by the driver monitor, under the condition of narrowing down the teacher data acquired from the cloud, or Based on any one of the time, place, and conditions specified by the driver, the teacher data is narrowed down and the normal range is learned. The running support device is characterized by this.

9. In the running support device according to claim 1, the following operation normal learning unit corrects the normal range based on a response from the driver. The running support device is characterized by this.

10. In the running support device according to claim 4, the following operation normal learning unit learns the normal range from the teacher data acquired from the cloud. The running support device is characterized by this.

11. In the running support device according to claim 7, the following operation normal learning unit acquires the learned normal range from the cloud. The running support device is characterized by this.

12. In the running support device according to claim 1, The follow-up operation normal learning unit calculates any one or more of frequency, amplitude, or bandwidth by frequency-converting the change in the relative relationship between the host vehicle and the lane line in addition to the feature amount, and when any one or more of the calculated frequency, amplitude, or bandwidth is within a predetermined threshold range, or when the average value of the change in the relative relationship between the host vehicle and the lane line is within a predetermined threshold range, it learns as normal. A driving support device characterized by this.

13. In the driving support device according to claim 1, The follow-up operation normal learning unit uses, as teacher data, a feature amount whose deviation possibility from the change in the relative relationship between the host vehicle and the lane line is equal to or less than a certain value in addition to the feature amount, and learns the normal range. A driving support device characterized by this.

14. In the driving support device according to claim 1, A driving support device further comprising a follow-up operation abnormality determination suppression unit that predicts a transient state in which it is impossible to determine whether the current fluctuation of the follow-up operation is normal or abnormal, and suppresses the determination of the follow-up operation abnormality determination unit.

15. In the driving support device according to claim 14, The transient state is A state in which sudden acceleration or deceleration of the preceding vehicle has occurred, A state in which the turn signal of the preceding vehicle has been activated, A state in which there is a sharp curve in front of the preceding vehicle, A state in which there is a sharp slope in front of the preceding vehicle, A state in which a stop line exists in front of the preceding vehicle, A state in which there is a red signal in front of the preceding vehicle, A state in which the lane width in front of the preceding vehicle is decreasing, A state in which a pedestrian is walking near the host vehicle lane, A state in which it can be predicted that an adjacent vehicle will cut in front of the host vehicle, A state in which sudden acceleration or deceleration of the host vehicle has occurred, A state in which the driver of the host vehicle has operated the turn signal, A state in which there is an object that can be predicted to progress in the driving lane of the host vehicle, A driving support device characterized by including any one or more of the states in which the reliability of recognition of the preceding vehicle is decreasing.

Images