Vehicle travel control system

Abstract

The present invention comprises: a first reliability setting unit that sets a first reliability, which is the reliability of the result of recognition by a surroundings recognition unit mounted on a host vehicle; a second reliability setting unit that sets a second reliability, which is the reliability of data of roads in the vicinity of the host vehicle acquired from map data; a correlation degree setting unit that sets the degree of correlation between a lane shape obtained on the basis of the recognition result and a lane shape obtained on the basis of the road data; an adjustment unit that adjusts the difference between the lane shape based on the recognition result and the lane shape based on the road data, on the basis of the first reliability, the second reliability, and the degree of correlation; and a lane keeping control unit that performs lane keeping control of the host vehicle using the result of the adjustment by the adjustment unit. This makes it possible to continue lane keeping even when traveling in a section for which no prior information regarding lane keeping is available.

Description

Vehicle Driving Control System

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

[0002] As a technology to assist drivers in driving, for example, a technology known as lane keep (LK) is known, which generates a target route for the vehicle to travel and maintains the lane by following the generated target route.

[0003] Known technology relating to such driver assistance is, for example, that described in Patent Document 1. Patent Document 1 discloses a vehicle driving control device provided in a vehicle control system that includes map information storage means for storing map information, positioning means for receiving signals from satellites to acquire positioning information of the vehicle's position, and driving environment information acquisition means for acquiring driving environment information including lane marking information ahead of the vehicle, for the purpose of continuing appropriate driving control even when lane marking information cannot be acquired due to snow accumulation, and includes control information calculation means for calculating control information for driving control of the vehicle based on the map information, the positioning information, and the lane marking information, driving control means for controlling the vehicle based on the control information, and a plurality of estimated positions related to the current vehicle position, each calculated by a plurality of calculation methods using information on the vehicle's position on the map information most recently calculated based on the lane marking information. a vehicle driving control device that includes an estimated position calculation means that calculates an estimated position reliability that changes cumulatively based on the comparison results of position information including the estimated position; an estimated position reliability calculation means that calculates the control information using the estimated position when the lane marking information cannot be acquired due to snow accumulation; and an abnormality control information calculation means that calculates the control information using the estimated position when the vehicle's position on the map information is newly calculated based on features acquired as the driving environment information, even when the lane marking information cannot be acquired, the estimated position calculation means calculates the estimated position using information related to the newly calculated vehicle position, and the driving control means continues the driving control until the estimated position reliability becomes equal to or less than a threshold, even when the lane marking information cannot be detected.

[0004] Japanese Patent Application Laid-Open No. 2018-173304

[0005] In the above-mentioned conventional technology, the driver's steering input for offsetting the vehicle's driving path from the road center is stored in advance, and when information such as lane markings cannot be obtained, driving control is continued by offsetting the target route from the road center based on the pre-stored steering input. However, driving control cannot be continued in driving locations where the offset amount, which is advance information related to lane keeping, is not stored in advance. Furthermore, since the offset amount is predetermined for each driving location, it cannot be adjusted according to the situation, and there is room for improvement in reliability.

[0006] The present invention has been made in consideration of the above, and aims to provide a vehicle driving control system that can continue to keep the vehicle in its lane even when driving in a section where there is no prior information regarding lane keeping.

[0007] The present application includes multiple means for solving the above-described problems, and one example thereof includes a first reliability setting unit that sets a first reliability which is the reliability of the recognition result of an external environment recognition unit installed in the host vehicle; a second reliability setting unit that sets a second reliability which is the reliability of road data in the vicinity of the host vehicle obtained from map data; a correlation setting unit that sets a correlation between a lane shape obtained based on the recognition result and a lane shape obtained based on the road data; an adjustment unit that adjusts the difference between the lane shape based on the recognition result and the lane shape based on the road data based on the first reliability, the second reliability, and the correlation; and a lane keeping control unit that performs lane keeping control of the host vehicle using the adjustment result by the adjustment unit.

[0008] According to the present invention, it is possible to provide a vehicle driving control system that can continue to keep the vehicle in its lane even when traveling through a section where there is no prior information regarding lane keeping.

[0009] FIG. 1 is a functional block diagram schematically showing the configuration of a vehicle driving control system according to a first embodiment. FIG. 2 is a flowchart showing the processing flow of the vehicle driving control system according to the first embodiment. FIG. 3 is a diagram showing an example (func1) of a road data reliability calculation table. FIG. 4 is a flowchart showing the processing flow of the vehicle driving control system according to a second embodiment. FIG. 5 is a diagram showing an example (func2) of a road data reliability calculation table. FIG. 6 is a diagram showing an example (func3) of a road data reliability calculation table. FIG. 7 is a diagram showing an example (func4) of a road data reliability calculation table. FIG. 8 is a diagram showing an example (func5) of a road data reliability calculation table. FIG. 9 is a flowchart showing the processing flow of a vehicle driving control system according to a third embodiment. FIG. 10 is a diagram showing an example (func6) of a road data reliability calculation table. FIG. 11 is a functional block diagram schematically showing the configuration of a vehicle driving control system according to a fourth embodiment. FIG. 12 is a flowchart showing the processing flow of the vehicle driving control system according to the fourth embodiment.

[0010] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the following description, a case will be described in which each functional unit constituting the present invention is implemented in an ECU (Electronic Control Unit) of an ADAS (Advanced Driver-Assistance Systems), but the present invention is not limited to this. For example, the present invention can be applied even when some functions are realized by other devices.

[0011] First Embodiment A first embodiment of the present invention will be described with reference to FIGS. 1 to 3. FIG.

[0012] This embodiment illustrates a case in which the vehicle uses lane keeping control to pass through a road (e.g., a road at an intersection) without or with unclear white lines indicating the driving lane, in a driving state in which there is a discrepancy between the actual lane shape around the vehicle detected by an external recognition sensor such as a stereo camera and the lane shape around the vehicle obtained based on the vehicle's driving location determined from a Global Navigation Satellite System (GNSS) and map information. In the following description, it is assumed that the road information on the map is not affected by GNSS multipath or matching errors, and that the GNSS is updated at an appropriate interval (e.g., every 50 ms). The series of processes in this embodiment are also assumed to be performed at an appropriate interval (e.g., once every 50 ms) within an ADAS controller installed in the vehicle.

[0013] FIG. 1 is a functional block diagram showing a schematic configuration of a vehicle driving control system according to this embodiment.

[0014] 1, vehicle driving control system 100 is mounted on a vehicle driven by a driver (hereinafter referred to as the host vehicle) and performs control related to driving assistance for the driver of the host vehicle, and includes a lane recognition unit 110, a driving lane shape correction unit 120, a lane keeping control unit 130, and a vehicle control device (VMC: Vehicle Motion Controller) 140. Note that the respective functional units of the driving lane shape correction unit 120 and the lane keeping control unit 130 are implemented in, for example, an ADAS controller (ECU).

[0015] The driving lane shape correction unit 120 outputs the lane shape calculated based on information input from the external environment recognition sensor 101, the yaw rate sensor 102, the vehicle speed sensor 103, the GNSS (Global Navigation Satellite System), the road information acquisition unit 105, the city information acquisition unit 106, and the lane recognition unit 110 to the lane keeping control unit 130.

[0016] The lane recognition unit 110 acquires the shape of the current lane around the vehicle (first lane shape) based on information such as images acquired by the external environment recognition sensor 101 (for example, a stereo camera), and outputs the acquired shape to the driving lane shape correction unit 120. The first lane shape is expressed, for example, by the coefficients of an approximate curve representing the lane recognized by the lane recognition unit 110.

[0017] The driving lane shape correction unit 120 is roughly composed of a road data lane shape calculation unit 121, a lane recognition reliability calculation unit 122, a road data reliability calculation unit 123, a correlation degree determination unit 124, an offset calculation unit 125, a lane shape correction unit 126, and a lane shape switching unit 127.

[0018] The road data lane shape calculation unit 121 calculates the lane shape around the vehicle (second lane shape) taking into consideration information prepared in advance, using the position information (positioning result) of the vehicle acquired by the GNSS 104, the speed of the vehicle (vehicle speed information) acquired by the vehicle speed sensor 103, the yaw rate of the vehicle acquired by the yaw rate sensor 102, the shape of the lanes around the vehicle (first lane shape) acquired by the lane recognition unit 110, and road data acquired by the road information acquisition unit 105. Note that the second lane shape is expressed, for example, by the coefficients of an approximate curve representing the lanes in the road data.

[0019] The lane recognition reliability calculation unit 122 (first reliability setting unit) calculates the reliability of lane recognition (first reliability) using information acquired by the external recognition sensor 101 (rainfall information, external light information), the speed of the host vehicle (vehicle speed information) acquired by the vehicle speed sensor 103, the yaw rate of the host vehicle acquired by the yaw rate sensor 102, the position of the host vehicle (positioning result) acquired by the GNSS 104, and pre-stored map information.

[0020] The road data reliability calculation unit 123 (second reliability setting unit) calculates the reliability of the road data (second reliability) using the node point position coordinates and node point density of the driving route near the vehicle acquired by the road information acquisition unit, high-rise building information near the vehicle acquired by the city information acquisition unit 106, and the elapsed time from the update period timing.

[0021] The correlation degree determination unit 124 uses the first lane shape from the lane recognition unit 110 and the second lane shape from the road data lane shape calculation unit 121 to calculate the correlation degree between the first lane shape and the second lane shape.

[0022] The offset calculation unit 125 calculates an offset using the first lane shape from the lane recognition unit 110, the second lane shape from the road data lane shape calculation unit 121, the reliability of lane recognition (first reliability) from the lane recognition reliability calculation unit 122, the reliability of road data (second reliability) from the road data reliability calculation unit 123, and the correlation from the correlation determination unit 124, and outputs the offset to the lane shape correction unit 126. Furthermore, the offset calculation unit 125 outputs the first lane shape from the lane recognition unit 110 and the second lane shape from the road data lane shape calculation unit 121 to the lane shape correction unit 126 together with the calculated offset.

[0023] The lane shape correction unit 126 corrects the lane shape using the offset from the offset calculation unit 125, the first lane shape and the second lane shape obtained via the offset calculation unit 125, and the road data acquired by the road information acquisition unit 105.

[0024] The lane shape switching unit 127 switches the lane shape to be output to the lane keeping control unit 130 depending on the lane detection performance of the lane recognition unit 110. Specifically, if the lane detection performance of the lane recognition unit 110 deteriorates, the lane shape switching unit 127 outputs the lane shape corrected by the lane shape correction unit 126 to the lane keeping control unit 130;

[0025] The lane keeping control unit 130 has a control command value calculation unit 131. The control command value calculation unit 131 calculates a control command value for performing lane keeping control based on the lane shape (first lane shape or second lane shape) output from the lane shape switching unit 127 of the driving lane shape correction unit 120, and outputs the control command value to the VMC 140.

[0026] The VMC 140 has an actuator control unit 141. The actuator control unit 141 controls actuators that assist the steering, acceleration, braking, and other operations of the vehicle, and performs lane keeping control by controlling each actuator in accordance with a control command value from the lane keeping control unit 130.

[0027] 2 is a flowchart showing the flow of processing in the vehicle driving control system according to this embodiment. As mentioned above, the series of processing steps (steps S100 to S240) shown in FIG. 2 is performed in an ADAS controller mounted on the vehicle at an appropriate interval (for example, once every 50 ms).

[0028] First, the vehicle driving control system 100 reads the position, speed, yaw rate, road data, rainfall information, and external light information of the vehicle (step S100).

[0029] Next, the first lane shape (lateral position to lane: c1, yaw angle to lane: b1, lane curvature: a1) obtained by lane recognition unit 110 is read (step S110).

[0030] Next, using the vehicle position and road data, the second lane shape (lateral position of the approximate curve obtained from the lane on the road data: c2, yaw angle from the lane: b2, lane curvature: a2) is calculated (step S120).

[0031] Next, the reliability of lane recognition is calculated using rainfall information, external light information, the speed and yaw rate of the vehicle, the position of the vehicle, and road information (step S130).

[0032] The reliability of lane recognition is, for example, defined as the best value (highest value) with (reliability) = 1 and the lowest value with (reliability) = 0, and is defined as the percentage of times one or more predetermined conditions are satisfied (the percentage of all items defined as conditions that satisfy that condition). Here, the following three items (1) to (3) are set as conditions in advance, and the reliability is calculated by dividing the number of items that satisfy the conditions by the total number of items (here, three). That is, for example, if two of the conditions (1) and (3) among the conditions (1) to (3) are satisfied, the reliability of lane recognition is (reliability) = 2 / 3 ≈ 0.67.

[0033] Item (1): Roll angle < predetermined value and pitch angle < predetermined value The greater the change in the behavior of the vehicle, the more the shape of the road (the V-shape) captured by the lane recognition unit is distorted due to pitch and roll movements. Therefore, one of the conditions for calculating reliability is that roll angle < predetermined value and pitch angle < predetermined value. The roll angle can be estimated by multiplying the yaw rate by an appropriate gain, and the pitch angle can be estimated by multiplying the longitudinal G (longitudinal acceleration) by an appropriate gain.

[0034] Item (2): Rainfall < predetermined value, and amount of light from the outside < predetermined value. It is believed that the greater the rainfall and the stronger the outside light, the lower the accuracy of the lane shape determined by the lane recognition unit 110. Therefore, one of the conditions for calculating the reliability is set as rainfall < predetermined value, and amount of light from the outside < predetermined value. The rainfall is determined from the raindrop sensor, and the light amount is determined from the shutter speed.

[0035] Item (3): Distance to specific area > predetermined threshold When traveling near a specific area such as an intersection or a junction, there may be no lanes or there may be a possibility of misrecognition, so the reliability condition is set to "distance to specific area > threshold." The distance to the specific area can be calculated from the distance between the vehicle's position and a node point that has information about the specific area.

[0036] After the process of step S130 is completed, the DOP (Dilution of Precision) for evaluating the satellite arrangement is read (step S140).

[0037] Next, the reliability of the road data is calculated using the road data reliability calculation table with the DOP as an input (step S150).

[0038] FIG. 3 shows an example of a road data reliability calculation table (func1), with the horizontal axis representing the input DOP and the vertical axis representing the output road data reliability. As shown in FIG. 3 as (func1), the reliability of road data is, for example, set such that (reliability) = 1 indicates the best (highest value) and (reliability) = 0 indicates the lowest value. Here, considering that the fewer the number of satellites received or the closer the satellites are to each other, the greater the GNSS measurement error (in other words, the larger the DOP), func1 is set so that the reliability of road data decreases as the DOP increases. Specifically, when the DOP is between 0 (zero) and a predetermined value (e.g., DOP = 1), (reliability) = 1; when the DOP exceeds a predetermined value (e.g., DOP = 1), the reliability decreases as the DOP increases, and when the DOP exceeds 6, (reliability) = 0.

[0039] Returning to FIG. 2, after the process of step S150 is completed, the degree of correlation between the first lane shape and the second lane shape is calculated (step S160).

[0040] The correlation degree is an index that ranges from (correlation degree) = 0 (zero, no correlation) to (correlation degree) = 1 (maximum value), and for example, the correlation degree is determined to be higher as the difference between the first lane shape and the second lane shape is smaller, and is set so that the correlation degree does not become less than 0 (zero) when the difference exceeds a predetermined allowable value th. Specifically, the correlation degree is expressed by the following (Equation 1).

[0041]

[0042] In the above formula (1), x is the distance ahead of the vehicle and falls within the range of d1≦x≦d2.

[0043] After the process of step S160 is completed, it is then determined whether the reliability of the lane recognition is equal to or greater than a predetermined threshold and whether the reliability of the road data is equal to or greater than a predetermined threshold (step S170).

[0044] If the determination result in step S170 is YES, it is then determined whether the correlation is equal to or greater than a predetermined threshold (step S180), and if the determination result is YES, the offset is updated (step S190). The offset is updated by smoothing the difference between the first lane shape and the second lane shape using a low-pass filter or the like. At this time, the offset can be updated more accurately by increasing the time constant of the low-pass filter (i.e., stronger smoothing) as the correlation increases and decreasing the time constant (i.e., weaker smoothing) as the correlation decreases.

[0045] If the determination result in step S180 is NO, i.e., if the correlation is lower than the predetermined threshold, the offset is subsequently reset (step S181) by setting the offset to the current value of the difference between the first lane shape and the second lane shape.

[0046] Also, if the judgment result in step S170 is NO, that is, if at least one of the reliability of lane recognition is less than a predetermined threshold and the reliability of road data is less than a predetermined threshold, the offset value is maintained (step S171).

[0047] When the process of any one of steps S171, S181, and S190 is completed, it is then determined whether the lane detection performance of the lane recognition unit 110 is equal to or less than a predetermined threshold (step S200).

[0048] If the determination result in step S200 is YES, the lane shape is corrected by applying an offset value to the second lane shape (step S210), and the corrected lane shape is input to the lane keeping control unit 130 (step S220).

[0049] If the determination result in step S200 is NO, the first lane shape is input to the lane keeping control unit 130 (step S201).

[0050] After the processing of either step S201 or S220 is completed, the lane keeping control unit 130 subsequently calculates a lane keeping control command value based on the input lane shape and inputs it to the actuator control unit 141 (step S230).

[0051] Next, the actuator control unit 141 operates the steering actuator based on the control command value input from the lane keeping control unit 130 (step S240), and then ends the processing.

[0052] The effects of the present embodiment configured as above will be described.

[0053] Consider a case where, in a driving state in which there is a discrepancy between the actual lane shape around the vehicle detected by an external recognition sensor such as a stereo camera and the lane shape around the vehicle obtained based on the vehicle's driving position and map information obtained from a Global Navigation Satellite System (GNSS), the vehicle passes through a road (such as an intersection) that does not have or is unclear as a white line indicating the driving lane, using lane keeping control.

[0054] In such a case, for example, if commonly known odometry is used, since there are no lanes within the intersection, lane keeping control is continued by extrapolating road information from the map using odometry. However, since the vehicle travels according to the road alignment based on the map at the vehicle's current location, it is possible that the vehicle will deviate from its travel lane due to a discrepancy with the actual lane shape.

[0055] In other words, in the prior art, there were cases where cruise control could not be continued in a driving location where the offset amount, which is prior information related to lane keeping, was not stored in advance. Also, when a predetermined offset amount was used for each driving location, adjustment according to the situation was not possible, and there was room for improvement in reliability.

[0056] In contrast to this, in the present embodiment, the vehicle is configured to include a first reliability setting unit that sets a first reliability, which is the reliability of the recognition result of the external environment recognition unit installed in the vehicle; a second reliability setting unit that sets a second reliability, which is the reliability of road data in the vicinity of the vehicle obtained from map data; a correlation setting unit that sets the correlation between the lane shape obtained based on the recognition result and the lane shape obtained based on the road data; an adjustment unit that adjusts the difference between the lane shape based on the recognition result and the lane shape based on the road data based on the first reliability, the second reliability, and the correlation; and a lane keeping control unit that performs lane keeping control of the vehicle using the adjustment result by the adjustment unit.Therefore, it is possible to continue lane keeping even when traveling in a section where there is no prior information related to lane keeping.

[0057] In other words, by calculating and using an offset that absorbs the gap between the camera and the map, it is possible to continue to stay in the lane without deviating from the driving lane, even when lane recognition is impossible, such as at an intersection.

[0058] Second Embodiment A second embodiment of the present invention will be described with reference to Figures 4 and 8. In this embodiment, the same components as those in the first embodiment are denoted by the same reference numerals, and the description thereof will be omitted as appropriate.

[0059] This embodiment illustrates a case in which the vehicle uses lane keeping control to pass through a road (e.g., a road at an intersection) where there is no or unclear white line indicating the driving lane, in a driving state in which there is a discrepancy between the actual lane shape around the vehicle detected by an external recognition sensor such as a stereo camera and the lane shape obtained based on the vehicle's driving location determined from GNSS and map information. In the following description, it is assumed that the road information on the map is affected by GNSS multipath and matching errors, and that the GNSS is updated at an appropriate interval (e.g., every 50 ms). The series of processes in this embodiment are performed at an appropriate interval (e.g., once every 50 ms) within an ADAS controller installed in the vehicle.

[0060] 4 is a flowchart showing the flow of processing in the vehicle driving control system according to this embodiment. As mentioned above, the series of processing steps (steps S100 to S240) shown in FIG. 4 is performed in an ADAS controller mounted on the vehicle at an appropriate interval (for example, once every 50 ms).

[0061] First, the vehicle driving control system 100 performs the processes of steps S100 to S130. Note that the processes of steps S100 to S130 are the same as those in FIG. 2 (first embodiment), and therefore a description thereof will be omitted.

[0062] After the processing of step S130 is completed, the DOP (Dilution of Precision) for evaluating the satellite configuration and nearby high-rise building information from city information (e.g., 3D city model PLATEAU) are read (step S140A).

[0063] Next, at least one of the following items is input: the DOP, the interval between node points within a predetermined range acquired by the road information acquisition unit 105, the density of node points on nearby routes, and the distance to a high-rise building acquired from the vehicle's position coordinates and nearby high-rise building information, and the reliability of the road data is calculated using a road data reliability calculation table (step S150A). The reliability of the road data is calculated using, for example, one or more of the following road data reliability calculation tables (func1 to func5). When the reliability of road data is calculated using multiple conditions (road data reliability calculation tables), the average value of the reliabilities calculated based on each condition is used as the final reliability of the road data.

[0064] The road data reliability calculation table (func1) is the same as that in the first embodiment (FIG. 3), and therefore a description thereof will be omitted.

[0065] FIG. 5 shows an example of a road data reliability calculation table (func2), with the horizontal axis representing the average interval [m] of node points within a predetermined range as input and the vertical axis representing road data reliability as output. As shown as (func2) in FIG. 5, the reliability of road data is, for example, set such that (reliability) = 1 indicates the best (highest value) and (reliability) = 0 indicates the lowest value. Here, considering that the narrower the node interval, the more likely the road shape is to be complex, func2 is set so that the reliability of road data increases as the average node interval increases. Specifically, when the average node interval is between 0 (zero) [m] and 2 [m], (reliability) = 0 (zero), and when the average node interval exceeds 2 [m], the reliability increases as the average interval increases, and when the average node interval exceeds 5 [m], (reliability) = 1.

[0066] 6 is a diagram showing an example of a road data reliability calculation table (func3), in which the horizontal axis represents the distance [m] from the vehicle to the high-rise building, which is the input, and the vertical axis represents the road data reliability, which is the output. As shown as (func3) in FIG. 6, the reliability of the road data is, for example, set such that (reliability) = 1 represents the best value (highest value) and (reliability) = 0 represents the lowest value. In this example, in consideration of the fact that GNSS positioning errors increase due to multipath caused by GNSS radio waves being reflected by obstacles such as buildings, func3 is set so that the reliability of the road data increases as the distance to the high-rise building obtained from a 3D city model increases. Specifically, when the distance to the high-rise building is between 0 (zero) [m] and 40 [m], the reliability is set to 0 (zero), and when the distance to the high-rise building exceeds 40 [m], the reliability increases as the distance increases, and when the average distance exceeds 150 [m], the reliability is set to 1.

[0067] FIG. 7 is a diagram showing an example of a road data reliability calculation table (func4), in which the horizontal axis represents the density of node points [number / m ２] and the vertical axis indicates the reliability of road data, which is the output. As shown as (func4) in Fig. 7, the reliability of road data is, for example, set such that (reliability) = 1 is the best value (highest value) and (reliability) = 0 is the lowest value. Here, in consideration of the fact that the closer the location where the vehicle is traveling is to a side road, the higher the possibility of a matching error, func4 is set so that the reliability of road data decreases as the density of node points increases. Specifically, when the density of node points is 0 (zero) [number / m ２ ] to a predetermined value [pieces / m ２ ], the reliability is set to 1, and the density of node points is set to a predetermined value [ ２ ], the reliability decreases as the density of node points increases, and when the density of node points exceeds 0.3 [pieces / m ２ ], the reliability is set to 0.

[0068] FIG. 8 shows an example of a road data reliability calculation table (func5), with the horizontal axis representing the number of overhead lanes (as input) and the vertical axis representing road data reliability (as output). As shown in FIG. 8 as (func5), the reliability of road data is, for example, set such that (reliability) = 1 indicates the best (highest value) and (reliability) = 0 indicates the lowest value. Considering the increased possibility of multipath and matching errors when there is a highway or the like overhead, func5 is set so that the reliability of road data decreases as the number of overhead lanes increases. Specifically, when the number of overhead lanes is between 0 (zero) and a predetermined number, the reliability is set to 1. When the number of overhead lanes exceeds the predetermined number, the reliability decreases with the increase in the number of overhead lanes. When the number of overhead lanes exceeds 6, the reliability is set to 0.

[0069] After the process of step S150A is completed, the processes of steps S160 to S240 are performed. Note that the processes of steps S160 to S240 are the same as those in FIG. 2 (first embodiment), and therefore a description thereof will be omitted.

[0070] The other configurations are the same as those of the first embodiment.

[0071] The present embodiment configured as above can also achieve the same effects as the first embodiment.

[0072] Furthermore, in this embodiment, the reliability of road data is calculated to reduce the effects of GNSS multipath, matching errors, etc., and even in locations where the reliability of road information is likely to be reduced due to being close to high-rise buildings, close to side roads, or the presence of a highway above, an offset that appropriately absorbs the gap between the camera and the map can be calculated, making it possible to continue staying in lane.

[0073] <Third Embodiment> A third embodiment of the present invention will be described with reference to Figures 9 and 10. In this embodiment, the same components as those in the first and second embodiments are denoted by the same reference numerals, and descriptions thereof will be omitted as appropriate.

[0074] This embodiment illustrates a case in which the vehicle uses lane keeping control to pass through a road (e.g., a road at an intersection) without or with unclear white lines indicating the driving lane, in a driving state in which there is a discrepancy between the actual lane shape around the vehicle detected by an external recognition sensor such as a stereo camera and the lane shape obtained based on the vehicle's driving location determined from GNSS and map information. In the following description, it is assumed that the road information on the map is affected by GNSS multipath and matching errors, and that the GNSS needs to be extrapolated using odometry updated at an appropriate interval (e.g., every 1 s). The series of processes in this embodiment are performed at an appropriate interval (e.g., once every 50 ms) within an ADAS controller installed in the vehicle.

[0075] 9 is a flowchart showing the flow of processing in the vehicle driving control system according to this embodiment. As described above, the series of processing steps (steps S100 to S240) shown in FIG. 9 is performed in an ADAS controller mounted on the vehicle at an appropriate interval (for example, once every 50 ms).

[0076] First, the vehicle driving control system 100 performs the processes of steps S100 and S110. Note that the processes of steps S100 and S110 are similar to those in Fig. 2 (first embodiment) and Fig. 4 (second embodiment), and therefore description thereof will be omitted.

[0077] After the processing of step S110 is completed, the vehicle position and road data are used to extrapolate the vehicle position and orientation using odometry, and the second lane shape (the lateral position to the lane, the yaw angle from the lane, and the lane curvature, which are coefficients of an approximate curve representing the lane shape of the road data) is calculated (step S120B).

[0078] Next, the reliability of lane recognition is calculated using rainfall information, external light information, the speed and yaw rate of the vehicle, the position of the vehicle, and road information (step S130).

[0079] Next, the DOP (Dilution of Precision) for evaluating the satellite configuration and nearby high-rise building information from city information (e.g., 3D city model PLATEAU) are read (step S140A). Note that the processing in step S140A is the same as that in Fig. 4 (second embodiment), and therefore a description thereof will be omitted.

[0080] Next, at least one of the DOP, the interval between node points within a predetermined range acquired by the road information acquisition unit 105, the density of node points on nearby routes, and the distance to a high-rise building acquired from the vehicle's position coordinates and nearby high-rise building information is input, and the reliability of the road data is calculated using a road data reliability calculation table (step S150B). To calculate the reliability of the road data, for example, one or more of the multiple road data reliability calculation tables (func1 to func6) shown below are used. Note that when the reliability of road data is calculated using multiple conditions (road data reliability calculation tables), the average value of the reliabilities calculated based on each condition is used as the final reliability of the road data.

[0081] The road data reliability calculation table (func1) is the same as that in the first embodiment (FIG. 3), and the road data reliability calculation tables (func2 to func5) are the same as those in the second embodiment (FIGS. 5 to 8), so their explanations will be omitted.

[0082] 10 is a diagram showing an example of a road data reliability calculation table (func6), in which the horizontal axis indicates the elapsed time [s] from the time when GNSS information was acquired, which is the input, and the vertical axis indicates the road data reliability, which is the output. As shown as (func6) in FIG. 10, the reliability of the road data is, for example, set such that (reliability) = 1 indicates the best value (highest value) and (reliability) = 0 indicates the lowest value. Here, in consideration of the fact that the longer the time elapsed from the time when GNSS information was acquired, the larger the odometry error, func6 is set so that the reliability of the road data decreases as the elapsed time increases. Specifically, when the elapsed time from the acquisition of the GNSS information is between 0 (zero) [s] and a predetermined time [s], the reliability is set to 1; when the elapsed time exceeds the predetermined time [s], the reliability decreases at a predetermined rate (for example, slope α (< 0)) as the elapsed time increases; when the elapsed time exceeds 1 [s], the reliability decreases at another predetermined rate (for example, slope β: β < α < 0) as the elapsed time increases; and when the elapsed time exceeds 1.2 [s], the reliability is set to 0 (zero).

[0083] After the process of step S150B is completed, the processes of steps S160 to S240 are performed. Note that the processes of steps S160 to S240 are the same as those in Fig. 2 (first embodiment) and Fig. 4 (second embodiment), and therefore a description thereof will be omitted.

[0084] The other configurations are the same as those of the first and second embodiments.

[0085] The present embodiment configured as above can also achieve the same effects as the first and second embodiments.

[0086] Furthermore, in this embodiment, the vehicle position is updated using odometry to calculate an offset that appropriately absorbs the gap between the camera and the map, making it possible to continue maintaining the lane even if the GNSS update period is longer than the lane keep control period.

[0087] <Fourth embodiment> A fourth embodiment of the present invention will be described with reference to Figures 11 and 12. In this embodiment, the same components as those in the first to third embodiments are denoted by the same reference numerals, and descriptions thereof will be omitted as appropriate.

[0088] This embodiment illustrates a case in which the vehicle uses lane keeping control to pass through a road (e.g., a road at an intersection) where there is no or unclear white line indicating the driving lane, in a driving state in which there is a discrepancy between the actual lane shape around the vehicle detected by an external recognition sensor such as a stereo camera and the lane shape obtained based on the vehicle's driving location determined from GNSS and map information. In the following description, it is assumed that the road information on the map is affected by GNSS multipath and matching errors, and that the GNSS is updated at an appropriate interval (e.g., every 1 s). Furthermore, it is assumed that the series of processes in this embodiment are performed at an appropriate interval (e.g., once every 50 ms) within an ADAS controller installed in the vehicle.

[0089] FIG. 11 is a functional block diagram that shows a schematic configuration of a vehicle driving control system according to this embodiment.

[0090] 11, vehicle driving control system 100A is mounted on a vehicle driven by a driver (hereinafter referred to as the host vehicle), performs control related to driving assistance for the driver of the host vehicle, and includes a lane recognition unit 110, a driving lane shape correction unit 120A, and a vehicle control device (VMC: Vehicle Motion Controller) 140. Note that each functional unit of driving lane shape correction unit 120A is implemented in, for example, an ADAS controller (ECU).

[0091] The driving lane shape correction unit 120A calculates a control command value for performing lane keeping control based on information input from the external environment recognition sensor 101, the yaw rate sensor 102, the vehicle speed sensor 103, the GNSS (Global Navigation Satellite System), the road information acquisition unit 105, the city information acquisition unit 106, and the lane recognition unit 110, and outputs the calculated control command value to the VMC 140.

[0092] The driving lane shape correction unit 120A is roughly composed of a road data lane shape calculation unit 121, a lane recognition reliability calculation unit 122, a road data reliability calculation unit 123, a correlation determination unit 124, an offset calculation unit 125, a control command correction unit 126A, a control command switching unit 127A, a first control command value calculation unit 128A, and a second control command value calculation unit 129A.

[0093] The first control command value calculation unit 128A calculates a control command value (first control command value) for performing lane keeping control based on the first lane shape from the lane recognition unit 110.

[0094] The second control command value calculation unit 129A calculates a control command value (second control command value) for performing lane keeping control based on the second lane shape from the road data lane shape calculation unit 121.

[0095] The offset calculation unit 125A calculates an offset using the first lane shape from the lane recognition unit 110, the second lane shape from the road data lane shape calculation unit 121, the reliability of lane recognition (first reliability) from the lane recognition reliability calculation unit 122, the reliability of road data (second reliability) from the road data reliability calculation unit 123, the correlation from the correlation determination unit 124, the first control command value, and the second control command value, and outputs the offset to the control command correction unit 126A. Furthermore, the offset calculation unit 125A outputs the first lane shape from the lane recognition unit 110 and the second lane shape from the road data lane shape calculation unit 121 together with the calculated offset to the control command correction unit 126A.

[0096] The control command corrector 126A corrects the second control command value from the second control command value calculator 129A based on the offset from the offset calculator 125.

[0097] Control command switching unit 127A switches the control command value to be output to VMC 140 depending on the lane detection performance of lane recognition unit 110. Specifically, when the lane detection performance of lane recognition unit 110 deteriorates, control command switching unit 127A outputs the second control command value corrected by control command correction unit 126A to VMC 140, and otherwise outputs the first control command value from first control command value calculation unit 128A to VMC 140.

[0098] The VMC 140 has an actuator control unit 141. The actuator control unit 141 controls actuators that assist the steering, accelerator, brake, and other operations of the vehicle, and performs lane keeping control by controlling each actuator in accordance with a control command value from the driving lane shape correction unit 120A.

[0099] 12 is a flowchart showing the flow of processing in the vehicle driving control system according to this embodiment. As described above, the series of processing steps (steps S100 to S240) shown in FIG. 12 is performed in an ADAS controller mounted on the vehicle at an appropriate interval (for example, once every 50 ms).

[0100] The vehicle driving control system 100A first performs the processes of steps S100 to S160. Note that the processes of steps S100 to S160 are similar to those in Fig. 9 (third embodiment) and so will not be described further.

[0101] After the process of step S160 is completed, a control command value (first control command value) for performing lane keeping control is calculated based on the first lane shape (step S161C). The first control command value can be calculated using a general method as long as it uses the lateral position, yaw angle, and curvature of the road shape.

[0102] After the process of step S161C is completed, a control command value (second control command value) for performing lane keeping control is calculated based on the second lane shape (step S162C). The second control command value can be calculated using a general method as long as it uses the lateral position, yaw angle, and curvature of the road shape.

[0103] After the process of step S162C is completed, it is then determined whether the reliability of the lane recognition is equal to or greater than a predetermined threshold and whether the reliability of the road data is equal to or greater than a predetermined threshold (step S170).

[0104] If the determination result in step S170 is YES, it is then determined whether the correlation is equal to or greater than a predetermined threshold (step S180), and if the determination result is YES, the offset is updated (step S190C). The offset is updated by smoothing the difference between the first control command value and the second control command value using a low-pass filter or the like. At this time, the offset can be updated more accurately by increasing the time constant of the low-pass filter (i.e., stronger smoothing) as the correlation increases and decreasing the time constant (i.e., weaker smoothing) as the correlation decreases.

[0105] If the determination result in step S180 is NO, i.e., if the correlation is lower than the predetermined threshold, the offset is subsequently reset (step S181C) by setting the offset to the current value of the difference between the first control command value and the second control command value.

[0106] Also, if the judgment result in step S170 is NO, that is, if at least one of the reliability of lane recognition is less than a predetermined threshold and the reliability of road data is less than a predetermined threshold, the offset value is maintained (step S171).

[0107] When the process of any one of steps S171, S181C, and S190C is completed, it is then determined whether the lane detection performance of the lane recognition unit 110 is equal to or less than a predetermined threshold (step S200).

[0108] If the determination result in step S200 is YES, the second control command value is subsequently corrected by applying an offset value to the second control command value (step S210C), and the corrected lane shape is input to the actuator control unit 141 (i.e., the steering actuator) of the VMC 140 (step S220C).

[0109] If the determination result in step S200 is NO, the first control command value is subsequently input to the actuator control section 141 (i.e., the steering actuator) of the VMC 140 (step S201C).

[0110] When the process of either step S201C or S220C is completed, the VMC 140 subsequently operates the steering actuator based on the input control command value (step S240), and ends the process.

[0111] The other configurations are the same as those of the third embodiment.

[0112] The present embodiment configured as above can also achieve the same effects as the third embodiment.

[0113] <Notes> The present invention is not limited to the above-described embodiments, and includes various modifications and combinations within the scope of the gist of the present invention. Furthermore, the present invention is not limited to those having all of the configurations described in the above-described embodiments, and also includes those in which some of the configurations are omitted.

[0114] 100, 100A... Vehicle driving control system, 101... External environment recognition sensor, 102... Yaw rate sensor, 103... Vehicle speed sensor, 104... GNSS, 105... Road information acquisition unit, 106... City information acquisition unit, 110... Lane recognition unit, 120, 120A... Driving lane shape correction unit, 121... Road data lane shape calculation unit, 122... Lane recognition reliability calculation unit, 123... Road data reliability calculation unit, 124... correlation degree determination unit, 125, 125A... offset calculation unit, 126... lane shape correction unit, 126A... control command correction unit, 127... lane shape switching unit, 127A... control command switching unit, 128A... first control command value calculation unit, 129A... second control command value calculation unit, 130... lane keeping control unit, 131... control command value calculation unit, 140... vehicle control device, 141... actuator control unit

Claims

1. A vehicle driving control system comprising: a first reliability setting unit that sets a first reliability which is the reliability of the recognition result of an external environment recognition unit installed in the host vehicle; a second reliability setting unit that sets a second reliability which is the reliability of road data in the vicinity of the host vehicle obtained from map data; a correlation setting unit that sets a correlation between a lane shape obtained based on the recognition result and a lane shape obtained based on the road data; an adjustment unit that adjusts a difference between the lane shape based on the recognition result and the lane shape based on the road data based on the first reliability, the second reliability, and the correlation; and a lane keeping control unit that performs lane keeping control of the host vehicle using the adjustment result by the adjustment unit.

2. A vehicle driving control system according to claim 1, wherein the adjustment unit adjusts the difference in lane shapes as an offset.

3. A vehicle driving control system as described in claim 2, wherein the adjustment unit: increases the amount of change of the offset from its current value the higher the first reliability and the higher the second reliability; adjusts the offset to be smaller than its current value the higher the correlation; adjusts the offset to be larger than its current value the lower the correlation; and initializes the offset to its current value when the correlation is below a predetermined level.

4. A vehicle driving control system according to claim 2, characterized in that the adjustment unit adjusts the offset to a value that is set for the difference between the coefficients of an approximation curve that indicates the lane shape based on the road data and the coefficients of an approximation curve that indicates the lane shape based on the recognition results.

5. A vehicle driving control system according to claim 2, wherein the adjustment unit adjusts the offset to a value set for the difference between the driving steering angle in the lane shape and the lane shape based on the road data.

6. A vehicle driving control system according to claim 2, wherein the adjustment unit maintains the offset at a current value as the first reliability is lower and as the second reliability is lower.

7. A vehicle driving control system according to claim 1, wherein the first reliability setting unit sets the first reliability lower as the change in behavior of the host vehicle becomes greater.

8. A vehicle driving control system according to claim 1, wherein the first reliability setting unit sets the first reliability lower as the amount of rainfall / intensity of direct sunlight increases.

9. A vehicle driving control system according to claim 1, wherein the first reliability setting unit sets the first reliability to a low level when the vehicle is driving near an intersection or a junction.

10. A vehicle driving control system according to claim 1, wherein the second reliability setting unit sets the second reliability lower as the rate of decrease in the positioning accuracy of the host vehicle increases.

11. A vehicle driving control system as described in claim 1, characterized in that the second reliability setting unit sets the second reliability to a low level when the vehicle is driving in an area close to a high-rise building based on 3D city model data.

12. A vehicle driving control system according to claim 1, wherein the second reliability setting unit sets the second reliability lower as the density of nodes in the road data increases.

13. A vehicle driving control system as described in claim 1, wherein the second reliability setting unit sets the second reliability to a low level when node point information indicates that the vehicle is driving near a location with an expressway above.

14. A vehicle driving control system according to claim 1, wherein the second reliability setting unit sets the second reliability lower as the node interval of the road data becomes narrower.

15. A vehicle driving control system according to claim 1, wherein the second reliability setting unit sets the second reliability lower as the time that has elapsed since the update cycle timing becomes longer.

Images