JP2016218737A - Preceding vehicle estimation device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make it possible to estimate preceding vehicles even when lane marks are unclear or do not exist.SOLUTION: By a vehicle detection unit 36, preceding vehicle candidates are detected based on a stereo image generated by an imaging device 14. By an own drive lane estimation unit 32, a drive lane where an own vehicle drives is estimated based on a vehicle drive trajectory represented by time-series data on a location of the own vehicle detected by a location detection unit 20. By a lane allocation unit 40, a drive lane where the preceding vehicle drives is determined based on a vehicle drive trajectory represented by time-series data on locations of the detected preceding vehicle candidates. By a preceding vehicle determination unit 42, a tracking preceding vehicle is estimated from the preceding vehicle candidates, based on the vehicle drive trajectory of the own vehicle, the vehicle drive trajectory of the preceding vehicle candidate, the drive lane of the own vehicle, and the drive lane of the preceding vehicle candidate.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a preceding vehicle estimation device and a program.

  Conventionally, a follow-up traveling device that detects a vehicle ahead of the host vehicle in order to maintain a safe distance is known (Patent Document 1).

  Further, the vehicle ahead of the host vehicle on the curved road is detected based on the navigator map. A vehicle preceding vehicle detection device is known in which the vehicle is detected by a laser scanner (Patent Document 2).

  There is also known a vehicle travel control device that detects a white line and a preceding vehicle in the white line and determines the lateral position of the host vehicle based on the preceding vehicle (Patent Document 3).

  There is also known a preceding vehicle display device in which a preceding vehicle is determined by inter-vehicle communication (Patent Document 4).

  Also, an autonomous traveling device that performs route planning for autonomous traveling based on an environmental map is known (Patent Document 5).

  Non-Patent Document 1 describes a method of predicting a relatively long-term vehicle motion at an intersection.

JP 2008-150001 A JP 2000-187800 A JP 2013-180638 A JP 2014-89510 A JP 2014-10797 A

Petrich, Dominik, Thao Dang, Dietmar Kasper, Gabi Breuel, and Christoph Stiller. "Map-based Long Term Motion Prediction for Vehicles in Traffic Environments." In International IEEE Conference on Intelligent Transportation Systems (ITSC), pp. 2166-2172. 2013.

  In the technique described in Patent Document 1 described above, only the vehicle ahead of the host vehicle is detected, and the detection result is used only for maintaining a safe distance.

  The navigation map used in the method described in Patent Document 2 does not have travel lane information. This is also useful for curved roads, but in the case of multiple lanes or intersections, the preceding vehicle cannot be detected accurately.

  In the technique described in Patent Document 3, the own lane is determined by white line detection. However, in a small town or an intersection, the white line may be difficult to see or may not be present.

  In the technique described in Patent Document 4 described above, the preceding vehicle is detected on a communication basis, and therefore a specific communication unit is required for all surrounding vehicles. It does not work if the communication unit is not equipped in another car.

  In the technique described in Patent Document 5, the map is merely prior information and may be changed. In that case, route planning will not work.

  In the technique described in Non-Patent Document 1 above, a lane level map is used for movement prediction, but the allocation of lanes to vehicles is based on the fusion of many sensors, and travel path history and consistency issues are considered. Not. Therefore, if a position identification error occurs, it will not function easily.

  The present invention has been made in view of the above circumstances, and provides a preceding vehicle estimation device and program capable of accurately estimating a preceding vehicle even when the lane mark is unclear or does not exist. Objective.

  In order to achieve the above object, a preceding vehicle estimation device of the present invention is mounted on a host vehicle, and is based on an image generated by an imaging unit that images the periphery of the host vehicle and generates an image. Based on the vehicle travel locus represented by the time series data of the position of the host vehicle detected by the vehicle detection unit that detects the position of the host vehicle and the position detection unit that detects the position of the host vehicle, the travel lane on which the host vehicle is traveling is determined. A vehicle lane determining unit that determines the traveling lane in which the surrounding vehicle is traveling based on a vehicle traveling locus represented by time-series data of the position of the surrounding vehicle detected by the vehicle detecting unit. A lane determining means, a vehicle traveling locus of the host vehicle, a vehicle traveling locus of the surrounding vehicle, a traveling lane determined by the own vehicle traveling lane determining means, and the surrounding vehicle traveling lane determining means; Preceding vehicle estimation means for estimating a preceding vehicle of the host vehicle from the surrounding vehicles traveling in the same traveling lane as the traveling lane in which the host vehicle is traveling based on the traveling lane determined It consists of

  A program according to the present invention includes a vehicle detection unit that detects a surrounding vehicle based on an image that is mounted on the host vehicle and that generates an image by capturing an image of the periphery of the host vehicle. Self-vehicle travel lane determination means for determining a travel lane in which the self-vehicle is traveling based on a vehicle travel locus represented by time-series data of the position of the self-vehicle detected by the position detection means for detecting the position of the own vehicle; Based on the vehicle travel locus represented by the time series data of the position of the surrounding vehicle detected by the vehicle detecting means, the surrounding vehicle traveling lane determining means for determining the traveling lane in which the surrounding vehicle is traveling, and the own vehicle Vehicle travel trajectory, vehicle travel trajectory of the surrounding vehicle, travel lane determined by the own vehicle travel lane determination means, and determined by the surrounding vehicle travel lane determination means This is a program for functioning as preceding vehicle estimation means for estimating a preceding vehicle of the own vehicle from the surrounding vehicles that are traveling in the same traveling lane as the traveling lane in which the host vehicle is traveling based on a traveling lane. .

  According to the present invention, the vehicle detection unit detects the surrounding vehicle based on the image mounted on the host vehicle and generated by the imaging unit that captures an image of the periphery of the host vehicle and generates an image.

  Then, based on the vehicle travel locus represented by the time series data of the position of the host vehicle detected by the position detection unit that detects the position of the host vehicle by the host vehicle travel lane determining unit, the travel lane in which the host vehicle is traveling Determine. The surrounding vehicle traveling lane determining means determines the traveling lane in which the surrounding vehicle is traveling based on the vehicle traveling locus represented by the time series data of the position of the surrounding vehicle detected by the vehicle detecting means.

  Then, the preceding vehicle estimation means determines the vehicle traveling locus of the own vehicle, the vehicle traveling locus of the surrounding vehicle, the traveling lane determined by the own vehicle traveling lane determining means, and the surrounding vehicle traveling lane determining means. Based on the travel lane, a preceding vehicle of the host vehicle is estimated from the surrounding vehicles traveling in the same travel lane as the travel lane in which the host vehicle is traveling.

  As described above, the preceding vehicle of the host vehicle is estimated based on the vehicle traveling locus of the own vehicle and the vehicle traveling locus of the surrounding vehicle from the surrounding vehicles traveling in the same traveling lane as the traveling vehicle is traveling. Thus, even if the lane mark is unclear or does not exist, the preceding vehicle can be estimated with high accuracy.

  The preceding vehicle estimation device according to the present invention includes a physical boundary detection unit that detects a physical boundary of a traveling path on which the host vehicle is traveling based on an image generated by the imaging unit, and the physical boundary detection unit A travel route for planning a travel route of the host vehicle based on a physical boundary of the travel route detected by the vehicle and a vehicle travel locus of a surrounding vehicle estimated to be the preceding vehicle by the preceding vehicle estimation means. Planning means. This makes it possible to plan a safe travel route for the host vehicle.

  In addition, the preceding vehicle estimation device further includes driving behavior determining means for determining driving behavior of the own vehicle based on a vehicle traveling locus of a surrounding vehicle estimated to be the preceding vehicle by the preceding vehicle estimating means. The travel route planning means includes a physical boundary of the travel path detected by the physical boundary detection means, a vehicle travel locus of surrounding vehicles estimated by the preceding vehicle estimation means to be the preceding vehicle, Based on the driving behavior of the host vehicle determined by the driving behavior determining means, the travel route of the host vehicle can be planned. As a result, a safer travel route of the host vehicle can be planned.

  The own vehicle traveling lane determining means according to the present invention is a vehicle represented by time-series data of the lane information obtained in advance including the position of the lane and the connection relationship between the lanes, and the position of the own vehicle detected by the position detecting means. Based on the travel locus, a travel lane in which the host vehicle is traveling is determined, and the surrounding vehicle travel lane determination means is the time series data of the lane information and the position of the surrounding vehicle detected by the vehicle detection means. The travel lane in which the surrounding vehicle is traveling can be determined based on the vehicle travel locus represented by.

  The preceding vehicle estimation device according to the present invention further includes a tracking unit that tracks the surrounding vehicle detected by the vehicle detection unit, and the surrounding vehicle traveling lane determination unit includes a lane position and a connection relationship between the lanes. Based on the lane information obtained in advance and the vehicle travel locus represented by the time series data of the position of the surrounding vehicle obtained from the tracking result of the surrounding vehicle tracked by the tracking means, the surrounding vehicle travels. The running lane can be determined.

  The storage medium for storing the program of the present invention is not particularly limited, and may be a hard disk or a ROM. Further, it may be a CD-ROM, a DVD disk, a magneto-optical disk or an IC card. Furthermore, the program may be downloaded from a server or the like connected to the network.

  As described above, according to the preceding vehicle estimation device and the program of the present invention, the vehicle traveling locus and the surrounding vehicle of the own vehicle can be obtained from the surrounding vehicle traveling in the same traveling lane as the traveling vehicle is traveling. By estimating the preceding vehicle of the host vehicle based on the vehicle travel locus, it is possible to accurately estimate the preceding vehicle even when the lane mark is not clear or does not exist.

It is a figure which shows a mode that the preceding vehicle candidate was detected. (A) The figure which shows the example of the local map containing lane information, (B) It is a one part enlarged view of a local map. It is a block diagram which shows the driving assistance control apparatus which concerns on embodiment of this invention. It is a figure for demonstrating the allocation method of a lane. (A) A diagram for explaining a lane allocation method considering only the current position, (B) a diagram for explaining a lane allocation method considering both the current position and speed, and (C). It is a figure for demonstrating the method of this Embodiment. It is a figure which shows the allocation result by the lane allocation method. It is a flowchart which shows the content of the preceding vehicle determination process routine in the computer of the driving assistance control apparatus which concerns on embodiment of this invention. It is a figure which shows the estimation result of the tracking preceding vehicle when passing an intersection. It is a figure which shows the estimation result of the tracking preceding vehicle when passing an intersection. It is a figure which shows the detection result of the preceding vehicle in the conventional method.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the present embodiment, a case where the present invention is applied to a driving support control device that generates a vehicle traveling route for performing lane keeping control and outputs the driving route to a driving support unit will be described as an example.

<Outline of the present embodiment>
In the conventional method, a preceding vehicle is detected and determined, and a travel route is planned based on the observation and short-term prediction at that time. On the other hand, in this Embodiment, the travel locus of the vehicle which shows a road network is utilized as prior information and restrictions. Therefore, the prediction is longer and the uncertainty is much smaller. As a result, the detection of the preceding vehicle is optimal not only at the present time but also in the past history and future prospects over a long period of time. In this way, by using the vehicle travel locus, it is possible to estimate the following vehicle in the own lane even on a road where the boundary line such as a lane mark is deteriorated or does not exist.

  Further, the behavior (trajectory information) of the following vehicle is provided as high-level information, and is used in determining the driving behavior (stop, following, overtaking, etc.) of the own vehicle and planning the safe own vehicle travel route.

  Specifically, the tracking preceding vehicle in the own lane is estimated based on the traveling trajectory of the surrounding vehicle, the driving behavior of the own vehicle is determined using the estimated traveling trajectory of the following preceding vehicle, Plan the travel route of your vehicle.

  First, the lane information in the navigation map at the present time is acquired from the lane graph obtained from the average travel locus of the vehicle, the position of the host vehicle, and route information, and a local map including the lane information is acquired. A plurality of detected preceding vehicle candidates are projected and tracked on the local map by the extended Kalman filter. Next, by measuring the likelihood based on the standardized cross-correlation between the traveling trajectories of the plurality of vehicles and the plurality of traveling lanes, all the vehicles including the host vehicle are assigned to any of the lanes represented by the lane information. Then, not only the distance between the preceding vehicle candidate and the own vehicle travel lane at time T but also the similarity between the travel track of the own vehicle and the travel track of the preceding vehicle candidate is considered, and the following vehicle is estimated. Is called.

  When the following vehicle is estimated, the travel track of the following vehicle is used as altitude information used in the advanced driver assistance systems (ADAS) of the host vehicle, and the vehicle follows the current road conditions. In order to do so, the determination of the driving behavior in real time and the planning of the own vehicle travel route are performed.

  As shown in FIG. 10, in the conventional method, a road shoulder vehicle is detected as a preceding vehicle. This is because this vehicle is the latest vehicle detected on the estimated own lane. However, in the present embodiment, first, as shown in FIG. 1, a preceding vehicle candidate is detected, and as shown in FIG. 2, lane information is acquired in the navigation map, and on the local map including the lane information. The preceding vehicle candidate is detected and tracked. Then, the own vehicle and the preceding vehicle candidate are assigned to the travel lane by likelihood measurement. In this case, the own lane displayed with a broken line and the following vehicle preceding the vehicle displayed with a dotted line are also acquired. Thereafter, the traveling locus of the following vehicle that is indicated by a dotted line in the navigation map indicates that the following vehicle has overtaken the track. Accordingly, it is determined as a driving action that the own vehicle also overtakes the truck. Furthermore, the own vehicle travel route as the overtaking route shown by the one-dot chain line in FIGS. 1 and 2 is also planned based on the travel locus of the following vehicle.

<System configuration>
As shown in FIG. 3, the driving support control device 10 according to the embodiment of the present invention generates a stereo image by capturing a GPS receiver 12 that receives radio waves from GPS satellites and the front of the host vehicle. Based on the reception signals from the GPS satellites received by the imaging device 14 and the GPS receiver 12 and the stereo image captured by the imaging device 14, the following vehicle is estimated and the travel route of the host vehicle is planned. And a computer 16 that outputs to the driving support unit 18.

  The driving support unit 18 performs lane keeping control of the host vehicle, for example, based on the planned traveling route of the host vehicle.

  The GPS receiver 12 receives radio waves from a plurality of GPS satellites at each time and outputs received signals from all the received GPS satellites to the computer 16.

  The imaging device 14 repeatedly captures the front of the host vehicle at each time, repeatedly generates a stereo image, and outputs the stereo image to the computer 16.

  When the computer 16 is represented by functional blocks, as shown in FIG. 3, the GPS satellite information is obtained from the GPS receiver 12 for all GPS satellites that have received radio waves, and the absolute position of the vehicle at each time is obtained. Is generated in the past for a plurality of vehicles, a position detection unit 20 for calculating the vehicle position, a vehicle travel locus generation unit 22 for generating time series data of the absolute position of the host vehicle calculated by the position detection unit 20 as a vehicle travel locus. The absolute position of the host vehicle detected by the position detection unit 20 from the navigation database, the trajectory database 24 that stores the vehicle travel trajectory, and the lane graph generated from the vehicle travel trajectory, the map database 26 that stores the navigation map, and the navigation map. A lane graph acquisition unit 28 that generates a local map that includes information on the lane graph, Based on the travel route acquisition unit 30 that acquires route information of the host vehicle from the vigation system (not shown) and the vehicle travel locus and route information of the host vehicle, any lane included in the local map is traveled by the host vehicle. Generated by the own vehicle travel lane estimation unit 32 that estimates as a lane, the physical boundary detection unit 34 that detects the physical boundary of the travel path of the host vehicle, and the imaging device 14 from the stereo image generated by the imaging device 14. From the stereo image, the vehicle detection unit 36 that detects the preceding vehicle candidate, the vehicle tracking unit 38 that tracks the preceding vehicle candidate detected by the vehicle detection unit 36, and the vehicle travel locus obtained from the tracking result of the preceding vehicle candidate Based on the lane allocation unit 40 that allocates any lane included in the local map as the traveling lane of the preceding vehicle candidate, Based on the traveling lane, the vehicle traveling locus of the preceding vehicle candidate, and the traveling lane, the preceding vehicle determination unit 42 that determines any of the preceding vehicle candidates as the following vehicle and the vehicle traveling locus of the following vehicle. A behavior determination unit 44 that determines the driving behavior of the vehicle, and a route planning unit that plans the travel route of the host vehicle based on the vehicle travel locus of the following vehicle, the driving behavior of the host vehicle, and the physical boundary of the travel route. 46. The own vehicle travel lane estimation unit 32 and the lane allocation unit 40 are examples of the own vehicle travel lane determination unit and the surrounding vehicle travel lane determination unit, respectively. The preceding vehicle determination unit 42 is an example of a preceding vehicle estimation unit.

  The position detection unit 20 acquires GPS satellite information for all GPS satellites that have received radio waves from the GPS reception unit 12, and calculates the absolute position of the host vehicle based on the acquired GPS information. In the present embodiment, the absolute position of the host vehicle is obtained by using the information of GPS satellites and information obtained by an INS (Inertial Navigation System) sensor (not shown), as disclosed in JP-A-2013-130480. It is calculated by the same method as the positioning system described in the reference). In the positioning system method described in Patent Document 3, since the GPS and INS (Inertial Navigation System) sensors are used, the absolute position of the host vehicle has an error. In general, the absolute position error ranges from 1 m to 3 m. As the INS sensor, for example, an acceleration sensor, a geomagnetic sensor, a gyro sensor, or the like may be used.

  The vehicle travel locus generation unit 22 collects time-series data of the absolute position of the host vehicle calculated by the position detection unit 20 using the same method as the positioning system described in Patent Document 3, and generates a vehicle travel locus.

  The trajectory database 24 stores in advance vehicle travel trajectories collected during travel for each of the plurality of vehicles. For example, in each of a plurality of vehicles, similar to the position detection unit 20 and the vehicle travel locus generation unit 22, the absolute value of the vehicle is detected, the vehicle travel locus is generated, and the vehicle travel locus of the plurality of vehicles is collected. Good.

  The trajectory database 24 stores a lane graph generated in advance from a vehicle travel trajectory generated for each of a plurality of vehicles. For example, a plurality of travel tracks belonging to the same travel lane are averaged and registered as corresponding travel lanes in the lane graph. Thereby, a lane graph indicating lane information including the position and the topology is generated. The method developed by the present inventors for generating a lane graph from the traveling trajectories of a plurality of vehicles may also be used (Non-Patent Documents (Guo, Chunzhao, Jun-ichi Meguro, Yoshiko Kojima, and Takashi Naito. " (See Automatic lane-level map generation for advanced driver assistance systems using low-cost sensors. "In Robotics and Automation (ICRA), 2014 IEEE International Conference on, pp. 3975-3982. IEEE, 2014.))

  The map database 26 stores a navigation map representing road network information.

  When the vehicle position is given from the position detection unit 20, the lane graph acquisition unit 28 defines a navigation map having a certain size (for example, 30 m × 100 m), and obtains lane information in the defined area from the lane graph. Extract and generate a local map including lane information in the defined area (see FIG. 2A).

  The travel route acquisition unit 30 acquires route information to a set destination from a navigation system (not shown).

  Next, the principle of assigning a vehicle to a travel lane will be described. Assigning a vehicle to a driving lane is crucial for situational awareness and motion prediction. In the conventional method, the Euclidean distance or Mahalanobis distance is measured, and the travel lane closest to the vehicle is assigned. However, due to the specific error of the vehicle detection position, it is erroneously assigned especially on roads and intersections of multiple lanes. In addition, lane assignments are easily mistaken for old digital maps. Thus, such methods typically require high-precision GPS and up-to-date digital maps of predetermined courses. On the other hand, in the method of the present embodiment, an inexpensive sensor is used to construct a lane graph. This can secure a large number of search vehicles, can cover a wider area and at a higher speed, and can provide the latest information.

  In order to deal with potential errors due to positioning of the own vehicle and vehicle detection, the method according to the present embodiment considers not only the current position of the vehicle but also the past travel locus in the lane allocation scoring process. .

  More specifically, with respect to the allocation target vehicle, each neighboring lane is regarded as a travel lane candidate, and a score function shown in the following equation is calculated between the travel locus of the vehicle and the travel lane candidate.

  here

Represents the degree of similarity between the vehicle travel locus and the travel lane candidate, and is calculated as follows using the standardized cross-correlation method (NCC).

Here, t _c is the current time, Δt is a time interval, and defines a section of a traveling locus behind the vehicle (see the dotted frame in FIG. 4).

Represents the position of the vehicle in that section,

Is the nearest point in the lane candidate.

When

Are respectively the average position of the vehicle in the section and the average position of the travel lane candidates. It is preferable that white line recognition is also performed online and the information on white line recognition is used for specifying the position of the host vehicle in the lane.

Represents the total distance between the vehicle travel locus and the travel lane candidate, and is calculated as follows.

  Here, λ is a weighting parameter, and σ is a variance. The allocation target vehicle is allocated to the driving lane candidate having the maximum score.

  The lane allocation method in consideration of the travel trajectory of the present embodiment can overcome problems caused by detection position errors, multiple travel lanes, and irregular driving behavior. For example, FIG. 5 shows a two-lane road where a parked vehicle (solid hatched line), an overtaking vehicle (broken dashed line), a normally running vehicle (dot), and a host vehicle (open) are present. A method that considers only the current position (see FIG. 5A), a method that considers both the current position and velocity (see FIG. 5B), and the method of the present embodiment (FIG. 5C FIG. 6 shows a comparison of driving lane allocation methods by different methods including (see FIG. 6). As can be seen from FIG. 6, the method according to the present embodiment can perform stable and rational lane assignment in a complicated situation.

  In the present embodiment, the host vehicle travel lane estimation unit 32 relates to the host vehicle and uses each of the nearby lanes included in the local map generated by the lane graph acquisition unit 28 as a travel lane candidate as a vehicle travel trajectory generation unit. The score function shown in the above equation (5) is calculated between the travel trajectory of the host vehicle generated by 22 and the travel lane candidate, and the travel lane candidate with the maximum score is estimated as the host vehicle travel lane.

  The physical boundary detection unit 34 detects the physical boundary of the traveling path of the host vehicle from the stereo image generated by the imaging device 14. As a method for detecting a physical boundary, a conventionally known method may be used, and thus description thereof is omitted. For example, non-patent literature (Guo, Chunzhao, Junichi Meguro, Yoshiko Kojima, and Takashi Naito. "CADAS: A multimodal advanced driver assistance system for normal urban streets based on road context understanding." In Intelligent Vehicles Symposium (IV), 2013 IEEE , pp. 228-235. IEEE, 2013.).

  The vehicle detection unit 36 detects a preceding vehicle candidate present ahead based on the stereo image captured by the imaging device 14. For vehicle detection, for example, the references proposed by the present inventors (K. Goto, K. Kidono, Y. Kimura and T. Naito, "Pedestrian Detection and Direction Estimation by Cascade Detector with Multi-Classifiers Utilizing Feature Interaction Descriptor ", in proc. of 2011 IEEE Intelligent Vehicles Symposium (IV), pp. 224-229, 2011). This is a cascade combination of a plurality of classifiers using feature correlation descriptors (FIND).

  For each of the preceding vehicle candidates detected by the vehicle detection unit 36, the vehicle tracking unit 38 tracks the preceding vehicle candidate based on the position of the preceding vehicle candidate at each time detected by the vehicle detection unit 36. The travel history of the preceding vehicle candidate is output.

For example, in order to estimate the dynamic characteristics of the vehicle, the two-dimensional position (u, v) and velocity ( ^· u, ^· v) in the local map, and the size (w, h) of the bounding box of the detection area in the image Each preceding vehicle candidate is tracked individually using an extended Kalman filter (EKF) having a state consisting of: That is, the state S ⁱ _t of the preceding vehicle candidate i at time t, expressed as follows.

  The two-dimensional vehicle position (u, v) is calculated from a disparity (parallax) map. This is obtained based on an SGM (semi-global matching) algorithm using a real-time GPU implementation. First, motion prediction is performed using a dynamic motion model. This is calculated based on the past and current model state of the vehicle. Further, since the vehicle is normally driven along the traveling lane, when the traveling lane is allocated by the lane allocation unit 40, the traveling lane is used to give a moving direction at the time of motion prediction.

  For each of the preceding vehicle candidates detected by the vehicle detection unit 36, the lane allocation unit 40 uses each of the nearby lanes included in the local map generated by the lane graph acquisition unit 28 as a driving lane candidate, The score function shown in the above equation (5) is calculated between the travel locus of the preceding vehicle candidate obtained in step 38 and the travel lane candidate, and the travel lane candidate having the maximum score is determined as the travel lane of the preceding vehicle candidate. Assign as.

  Next, the principle of determining the following vehicle will be described.

  When the traveling lane is assigned to the preceding vehicle candidate, the vehicle is classified. Specifically, the preceding vehicle candidate belonging to the traveling lane of the own vehicle is inspected, and if it corresponds, it is determined as the following vehicle. Since the following vehicle is for control purposes, the preceding vehicle candidate in the own vehicle lane whose traveling locus coincides with the own vehicle is determined as the following vehicle, not the vehicle immediately before the own vehicle. It should be. Therefore, the likelihood score as the following vehicle is calculated for each preceding vehicle candidate belonging to the host vehicle lane by the following equation.

  here

Is a score between the preceding vehicle candidate and the vehicle lane.

Is a score between the host vehicle and the preceding vehicle candidate and is defined by the following equation.

  here,

Represents the similarity between the driving track of the host vehicle and the driving track of the preceding vehicle candidate,

Represents the total distance between the traveling locus of the host vehicle and the traveling locus of the preceding vehicle candidate. That is, it is expressed by the following formula.

Here, ^ph _t represents the position in the section of the travel locus of the host vehicle (see the dotted frame in FIG. 5B), and p ^v _{c, t} is the closest point in the travel locus of the preceding vehicle quantity candidate. is there. ¯p ^h _t and ¯p ^v _{c, t,} respectively, the average position of the travel locus of the preceding vehicle weight candidate and the average position of the travel locus of the vehicle in that section.

  As can be seen from these equations, the score of the preceding vehicle candidate whose traveling locus overlaps with the own vehicle becomes high. Therefore, after calculating the likelihood score for all the preceding vehicle candidates, if the maximum score is larger than the threshold value, the preceding vehicle candidate having the maximum score is determined as the follow-up preceding vehicle. Otherwise, there is no follow-up preceding vehicle determined for the host vehicle.

  As described above, in the present embodiment, the preceding vehicle determination unit 42 determines the preceding vehicle candidate by the own vehicle traveling lane estimated by the own vehicle traveling lane estimation unit 32, the traveling locus of the own vehicle, and the lane allocation unit 40. A likelihood score calculated according to the above equation (8) among the preceding vehicle candidates to which the same lane as the host vehicle traveling lane is assigned based on the lane assigned for each and the traveling trajectory of each preceding vehicle candidate. The preceding vehicle candidate having the maximum value and the likelihood score equal to or greater than the threshold is determined as the following vehicle.

  The action determination unit 44 determines whether the driving behavior of the host vehicle is steady driving, overtaking, lane changing, stopping, braking, and acceleration based on the driving trajectory of the vehicle determined by the preceding vehicle determining unit 42 as the following vehicle. Decide either. For example, using machine learning, a learning model for determining driving behavior is learned on the basis of a running locus from actual data, and the learning model is used to automatically learn from a running locus of the following vehicle. Determine the driving behavior of the vehicle.

  For example, if there is a vehicle that is parked in the vehicle lane in front of your vehicle and it is a parked car (you must pass it in this case), or the last car in a traffic jam Suppose you do not know (in this case you have to wait behind the car and wait). However, if a tracking preceding vehicle is detected and the following vehicle overtakes the vehicle, it is known that it is a parked car, and the driving action is determined so that the own vehicle also overtakes the parked vehicle. Is done.

  The route plan unit 46 includes a travel locus of the vehicle determined as the follow preceding vehicle by the preceding vehicle determination unit 42, the driving behavior of the host vehicle determined by the behavior determination unit 44, and the travel detected by the physical boundary detection unit 34. Based on the physical boundary of the road, the safe travel route and speed of the host vehicle are planned so as to imitate the travel trajectory of the following vehicle, and output to the driving support unit 18.

  For example, a safe and appropriate route and speed of the overtaking process are planned using the travel locus of the following vehicle. It is confirmed that this route is safe by the following vehicle. Also, because road conditions can change, route planning requires physical boundaries. In short, it is possible to plan a safe traveling route in an appropriate speed region based on the physical boundary detected by the host vehicle.

<Operation of the driving support control device>
Next, the operation of the present embodiment will be described.

  When the vehicle equipped with the driving support control device 10 is running, the imaging device 14 captures the front of the host vehicle, sequentially generates a stereo image, and the GPS receiver 12 generates a signal from a GPS satellite. Are sequentially received, the driving support control device 10 executes a preceding vehicle determination processing routine shown in FIG.

  In step 100, the GPS signal received by the GPS receiver 12 is acquired.

  In step 102, the absolute position of the host vehicle is detected based on the GPS signal acquired in step 102 and the INS information output from the INS sensor (not shown).

  In the next step 104, the travel route of the host vehicle from the navigation system to the destination is acquired.

  In step 106, lane information corresponding to the absolute position of the host vehicle is acquired from the lane graph stored in the trajectory database 24 based on the absolute position of the host vehicle detected in step 102, and Map information corresponding to the absolute position is acquired, and a local map including lane information is generated.

  In step 108, a stereo image generated by the imaging device 14 is acquired. In step 110, the GPS signal received by the GPS receiver 12 is acquired.

  In step 112, a preceding vehicle candidate is detected from the stereo image acquired in step 108. In step 114, each of the preceding vehicle candidates is tracked based on the detection result in step 112 and the previous tracking result in step 114.

  In step 116, the absolute position of the host vehicle is detected based on the GPS signal acquired in step 110 and the INS information output from the INS sensor (not shown).

  In step 118, the physical boundary of the travel path is detected from the stereo image acquired in step 108.

  In step 120, it is determined whether or not the processing in steps 108 to 118 has been repeated until a predetermined period has elapsed. If the process is repeated until the predetermined period elapses, the process proceeds to step 112. On the other hand, if the predetermined period has not elapsed, the process returns to step 108.

  In step 122, a travel locus of the host vehicle is generated based on the absolute position of the host vehicle detected in step 116. In step 124, based on the tracking result of each preceding vehicle candidate in step 114, a traveling locus of each preceding vehicle candidate is generated.

  In step 126, the host vehicle travel lane is estimated based on the lane information obtained in step 106 and the travel track of the host vehicle generated in step 122.

  In step 130, a lane is assigned to each preceding vehicle candidate based on the lane information obtained in step 106 and the travel locus of each preceding vehicle candidate generated in step 124.

  In step 132, the traveling locus of the host vehicle generated in step 122, the traveling locus of each preceding vehicle candidate generated in step 124, the own vehicle traveling lane estimated in step 126, and the step 130. The following vehicle is determined based on the lane assigned to each preceding vehicle candidate.

  In step 134, the driving behavior of the host vehicle is determined based on the traveling locus of the preceding vehicle candidate determined as the following vehicle in step 132.

  In step 136, the driving trajectory of the preceding vehicle candidate determined as the following vehicle in step 132, the driving behavior of the host vehicle determined in step 134, and the physical boundary detected in step 118 are determined. Based on this, the travel route of the host vehicle is planned, output to the driving support unit 18, and the process returns to step 100.

  By repeatedly executing the preceding vehicle determination processing routine, the planned traveling route of the own vehicle is output to the driving support unit 18 and is based on the planned traveling route of the own vehicle by the driving support unit 18. Driving assistance is provided.

  As described above, according to the driving support control apparatus according to the embodiment of the present invention, the traveling lanes of the host vehicle and the preceding vehicle candidate are determined, and the same traveling lane as the traveling lane on which the host vehicle is traveling is determined. This is a case where the lane mark is not clear or does not exist by estimating the following vehicle of the own vehicle based on the vehicle traveling locus of the own vehicle and the vehicle traveling locus of the preceding vehicle candidate from the traveling preceding vehicle candidate. However, the following vehicle can be estimated with high accuracy. In addition, a safe travel route of the host vehicle can be planned based on the vehicle travel locus of the following vehicle and the physical boundary of the travel route.

  Further, in the conventional method, there is no lane mark indicating the own lane at the intersection, so that the processing at the intersection cannot be performed. Even if the own lane is known, the conventional preceding vehicle estimation method does not work well because it only considers the relationship between the vehicle and the current lane at that time. This is because a plurality of surrounding vehicles are driven in different directions closely, and at some point, another vehicle is closer to the own lane than the preceding vehicle. On the other hand, in the method of the present embodiment, since the past travel locus of the vehicle is taken into consideration in addition to the current information, the following vehicle can be correctly determined. For example, as shown in FIG. 8, even when the host vehicle turns right at an intersection, the following vehicle can be accurately determined (see the vehicle in the hatched area). In addition, as shown in FIG. 9, when the follow-up preceding vehicle travels in a direction different from that of the host vehicle, it is possible to prevent the follow-up preceding vehicle from being erroneously determined.

  In addition, in the conventional detection of the preceding vehicle, the vehicle immediately before the host vehicle is detected because of adaptive cruise control (ACC) and pre-crash system (PCS). In this embodiment, the following vehicle in the own lane that can provide the most effective information for the own vehicle for determining the travel route, the speed range, the driving behavior, and the like is estimated. Therefore, the follow preceding vehicle estimated in the present embodiment is not for following the vehicle immediately before the own vehicle with respect to the following vehicle, but with advanced information that best matches the driving of the own vehicle, Driving behavior decision (lane maintenance, obstacle avoidance, lane change, etc.) and situation understanding (e.g., judging whether a car parked in front of the vehicle is parked or the last car in a traffic jam) Can be provided.

  Further, in the conventional system, inter-vehicle communication is used for cooperative operation of the preceding vehicle and the following vehicle. However, some surrounding cars do not have an inter-vehicle communication unit. Furthermore, although the travel locus transmitted from the preceding vehicle is acquired from the GPS, the GPS signal may not be acquired or the accuracy may be very low in a tunnel or a place with many high-rise buildings. On the other hand, in the present embodiment, a safe travel route of the host vehicle can be stably planned based on the travel locus of the following vehicle and the physical boundary of the travel route.

  Further, the following vehicle is determined not only based on the current positional relationship but also based on the past position / travel locus and the like. Therefore, the following vehicle can be estimated more stably and accurately. Also, surrounding vehicles are assigned to the corresponding travel lanes by the same method. As a result, it is possible to understand the movement of surrounding vehicles and to make a longer-term prediction. On the other hand, the travel trajectory of the following vehicle can support the planning of a longer travel trajectory of the host vehicle. In this case, the ADAS system can perform “early detection, determination, determination” (prefetching).

  In the above embodiment, the case where the trajectory database and the map database are provided in the driving support control device of the own vehicle has been described as an example, but the present invention is not limited to this, and the trajectory database and the map are not limited thereto. A database may be provided in another device, and the trajectory database and map database of the other device may be accessed via a network.

DESCRIPTION OF SYMBOLS 10 Driving assistance control apparatus 12 GPS receiving part 14 Imaging device 16 Computer 18 Driving assistance part 20 Position detection part 22 Vehicle travel locus generation part 24 Track database 26 Map database 28 Lane graph acquisition part 30 Travel route acquisition part 32 Own vehicle travel lane estimation Unit 34 physical boundary detection unit 36 vehicle detection unit 38 vehicle tracking unit 40 lane allocation unit 42 preceding vehicle determination unit 44 action determination unit 46 route planning unit

Claims (6)

Vehicle detection means for detecting a surrounding vehicle based on an image mounted on the own vehicle and generated by an imaging means for capturing an image of the periphery of the own vehicle and generating an image;
A host vehicle travel lane determination unit that determines a travel lane in which the host vehicle is traveling based on a vehicle travel locus represented by time-series data of the position of the host vehicle detected by the position detection unit that detects the position of the host vehicle; ,
Peripheral vehicle travel lane determining means for determining a travel lane in which the peripheral vehicle is traveling based on a vehicle travel locus represented by time series data of the position of the surrounding vehicle detected by the vehicle detection means;
Based on the vehicle traveling locus of the host vehicle, the vehicle traveling locus of the surrounding vehicle, the traveling lane determined by the own vehicle traveling lane determining unit, and the traveling lane determined by the surrounding vehicle traveling lane determining unit Preceding vehicle estimation means for estimating a preceding vehicle of the host vehicle from the surrounding vehicles traveling in the same traveling lane as the traveling lane in which the vehicle is traveling;
A preceding vehicle estimation device. Physical boundary detection means for detecting a physical boundary of a traveling path on which the host vehicle is traveling based on the image generated by the imaging means;
Based on the physical boundary of the travel path detected by the physical boundary detection means and the vehicle travel trajectory of surrounding vehicles estimated by the preceding vehicle estimation means as the preceding vehicle, the travel route of the host vehicle Travel route planning means for planning,
The preceding vehicle estimation device according to claim 1, further comprising: Driving behavior determining means for determining the driving behavior of the host vehicle based on the vehicle travel locus of the surrounding vehicle estimated to be the preceding vehicle by the preceding vehicle estimating means;
The travel route planning means includes a physical boundary of the travel path detected by the physical boundary detection means, a vehicle travel trajectory of a surrounding vehicle estimated by the preceding vehicle estimation means as the preceding vehicle, The preceding vehicle estimation device according to claim 2, wherein the traveling route of the host vehicle is planned based on the driving behavior of the host vehicle determined by the driving behavior determining means. The own vehicle traveling lane determining means includes lane information obtained in advance including the position of the lane and the connection relationship between the lanes, and a vehicle traveling locus represented by time series data of the position of the own vehicle detected by the position detecting means. Based on the travel lane in which the vehicle is traveling,
The surrounding vehicle travel lane determining means is a travel lane in which the surrounding vehicle is traveling based on the lane information and a vehicle travel locus represented by time series data of the position of the surrounding vehicle detected by the vehicle detection means. The preceding vehicle estimation device according to any one of claims 1 to 3, wherein Tracking means for tracking the surrounding vehicle detected by the vehicle detection means;
The surrounding vehicle travel lane determining means is configured to determine the position of the surrounding vehicle obtained from the lane information obtained in advance including the position of the lane and the connection relationship between the lanes and the tracking result of the surrounding vehicle tracked by the tracking means. The preceding vehicle estimation device according to any one of claims 1 to 4, wherein a traveling lane in which the surrounding vehicle is traveling is determined based on a vehicle traveling locus represented by time-series data. Computer
Vehicle detection means for detecting surrounding vehicles based on an image mounted on the own vehicle and generated by an imaging means for imaging the periphery of the own vehicle and generating an image;
Self-vehicle travel lane determination means for determining a travel lane in which the self-vehicle is traveling based on a vehicle travel locus represented by time-series data of the position of the self-vehicle detected by the position detection means for detecting the position of the own vehicle;
The surrounding vehicle traveling lane determining means for determining the traveling lane in which the surrounding vehicle is traveling based on the vehicle traveling locus represented by the time series data of the position of the surrounding vehicle detected by the vehicle detecting means; and The host vehicle travels based on a vehicle traveling locus, a vehicle traveling locus of the surrounding vehicle, a traveling lane determined by the own vehicle traveling lane determining unit, and a traveling lane determined by the surrounding vehicle traveling lane determining unit. A program for functioning as preceding vehicle estimation means for estimating a preceding vehicle of the host vehicle from the surrounding vehicles traveling in the same traveling lane as the traveling lane of the vehicle.