Track estimation method and track estimation device

Abstract

To suppress a decrease in the accuracy of estimating a runway due to that the runway of the host vehicle is estimated using road shapes of map data, when the road shapes of map data differ from the actual lane shapes.SOLUTION: The present invention includes: extracting a first road shape that represents a first road along which the host vehicle travels from map data in which road shapes are expressed in road units by arrangement of a plurality of nodes (S3); detecting a road boundary of the first road (S4); calculating the road width of the first road (S5); calculating the position of the road boundary in a map of map data on the basis of the vehicle position, and estimating a runway which is a travel track along the first road on the basis of the first road shape, when the road width is constant and the position of node of the first road shape is at center of road boundary of both edges (S8); and estimating a runway without using the first road shape, when the road width varies and the position of node of the first road shape is at center of road boundary of both edges (S11).SELECTED DRAWING: Figure 6

Description

Technical Field

[0001] The present invention relates to a travel route estimation method and a travel route estimation device.

Background Art

[0002] Patent Document 1 describes a technique in which position data of nodes, which are road marking points located before and after the current position of a vehicle, is read from a road map database, and the travel route shape of the vehicle is estimated from a plurality of nodes to perform irradiation control of a headlamp.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the case of map data in which the road shape is represented in road units by an array of a plurality of nodes, each node often indicates a position near the center of the road. For this reason, at a place where the number of lanes increases, such as near the entrance of an intersection, the position of the node in the lane width direction moves in the direction in which the number of lanes increases, so the shape represented by the array of nodes is different from the actual lane shape. In such a case, if the travel route of the host vehicle is estimated based on the road shape of the map data in autonomous driving control or driving support control, there is a risk that the estimation accuracy will be low. An object of the present invention is to suppress a decrease in the estimation accuracy of a travel route by estimating the travel route of the host vehicle using the road shape of map data when the road shape of the map data is different from the actual lane shape.

Means for Solving the Problems

[0005] In one aspect of the present invention, a route estimation method detects the vehicle's current position, extracts a first road shape from map data in which road shapes are represented by an array of nodes, which represents the road shape of the first road on which the vehicle is traveling, detects road boundaries extending along the direction of the first road at both ends in the road width direction, calculates the position of the road boundary on the map data based on the vehicle's position, calculates the road width of the first road from the distance between the road boundaries at both ends in a direction perpendicular to the array direction of the nodes of the first road shape, estimates a route that follows the first road based on the first road shape if the road width is constant and the position of the node of the first road shape is in the center of the road boundaries at both ends within a predetermined distance in front of the vehicle, and estimates a route without using the first road shape if the road width changes and the position of the node of the first road shape is in the center of the road boundaries at both ends within a predetermined distance in front of the vehicle. [Effects of the Invention]

[0006] According to the present invention, when the road shape in map data differs from the actual lane shape, the accuracy of the vehicle's route can be suppressed by using the road shape in map data to estimate the vehicle's route. [Brief explanation of the drawing]

[0007] [Figure 1] This figure shows an example of a schematic configuration of a vehicle equipped with the driving control device of the embodiment. [Figure 2] (a) and (b) are schematic diagrams illustrating the track estimation method of the first embodiment. [Figure 3] Figure 1 is a block diagram showing an example of the controller's functional configuration. [Figure 4] This is a schematic diagram illustrating an example of a method for estimating road layouts within and near intersections. [Figure 5] This is a schematic diagram of an example of road markings that warn of a divergence between a straight lane and a right-turn lane. [Figure 6] This is a flowchart of an example of the path estimation method of the first embodiment. [Figure 7]This is a schematic diagram illustrating the track estimation method of the second embodiment. [Figure 8] This is a flowchart of an example of the route estimation method of the second embodiment. [Figure 9] This is a schematic diagram illustrating the track estimation method of the third embodiment. [Figure 10] This is a schematic diagram illustrating the track estimation method of the fourth embodiment. [Modes for carrying out the invention]

[0008] Embodiments of the present invention will be described below with reference to the drawings. Note that the drawings are schematic and may differ from actual ones. Furthermore, the embodiments of the present invention described below are illustrative examples of devices and methods for realizing the technical concept of the present invention, and the technical concept of the present invention is not limited to the structure, arrangement, etc., of the components described below. The technical concept of the present invention can be modified in various ways within the technical scope defined by the claims described in the patent claims.

[0009] (First Embodiment) (composition) The vehicle 1 is equipped with a driving control device 10 that assists in the driving of the vehicle 1. The driving control device 10 detects the driving environment around the vehicle 1 and assists in the driving of the vehicle 1 by automatically controlling the driving of the vehicle 1 based on the detected driving environment. The driving control device 10 is an example of the "road estimation device" described in the claims. For example, the driving assistance of the vehicle 1 by the driving control device 10 may include autonomous driving control that automatically drives the vehicle 1 without the involvement of an occupant (e.g., driver). Alternatively, the driving assistance of the vehicle 1 by the driving control device 10 may include automatically controlling at least the steering angle of the vehicle 1. Alternatively, the driving assistance of the vehicle 1 by the driving control device 10 may include automatically controlling the driving force or braking force in addition to automatically controlling the steering angle of the vehicle 1.

[0010] The driving control device 10 includes a positioning device 11, a map database 12, an external sensor 14, a vehicle sensor 15, a controller 16, and an actuator 17. In the drawings, the map database is referred to as "Map DB". The positioning device 11 measures the current position of the vehicle 1. In the following description, the current position of the vehicle 1 will be referred to as "vehicle position". The positioning device 11 may include, for example, a Global Navigation Satellite System (GNSS) receiver. A GNSS receiver is, for example, a GPS receiver, which measures the vehicle's position by receiving radio waves from multiple navigation satellites.

[0011] Map database 12 stores map data. The map data stored in map database 12 may be, for example, navigation map data that includes information on a road-by-road basis. Navigation map data may be data that represents road shapes on a road-by-road basis using a series of nodes. Refer to Figure 2(a). Each of the circular plots N drawn in the figure represents a node in the map data. The same applies to Figures 2(b), 4, 7, 9, and 10. In the example in Figure 2(a), the road shape RS1 of road R1 (hereinafter referred to as "Road 1 R1") on which vehicle 1 is traveling is represented on a road-by-road basis by a sequence of nodes N.

[0012] In this specification, a node column may be, for example, an array of multiple nodes N, or data that includes an array of multiple nodes N and line segments (links L) connecting these nodes N with curves or straight lines. In other words, a node column represents an array of multiple nodes N, i.e., the shape of a road on a map, and is not limited to but may include an array of multiple nodes N.

[0013] Refer to FIG. 1. The external sensor 14 detects various information (surrounding environment information) about the surrounding environment around the host vehicle 1. For example, the external sensor 14 detects an object around the host vehicle 1. The external sensor 14 detects the surrounding environment of the host vehicle 1, such as an object existing around the host vehicle 1, the relative position between the host vehicle 1 and the object, the distance between the host vehicle 1 and the object, and the direction in which the object exists. The external sensor 14 outputs the detected surrounding environment information as surrounding environment information to the controller 16. For example, the external sensor 14 detects the relative position of other vehicles and targets around the host vehicle 1 with respect to the host vehicle 1. Here, the target means, for example, a traffic signal provided on the road on which the host vehicle 1 travels, a line on the road surface (white line, stop line, lane boundary line, lane division line, etc.), a curb of the road shoulder, a guardrail, a pole, a fence, a cat's eye or a block, etc.

[0014] The external sensor 14 may include a monocular camera such as a full HD resolution color camera. The camera captures an image including the recognition target of the surrounding environment of the host vehicle 1, and outputs the captured image as surrounding environment information to the controller 16. In addition, the external sensor 14 may include a distance measuring device such as a laser range finder (LRF), a radar, or a LiDAR (Light Detection and Ranging) laser radar. The distance measuring device detects, for example, a relative position determined by the relative distance and direction from an object existing around the host vehicle. The distance measuring device outputs the detected distance measuring data as surrounding environment information to the controller 16.

[0015] The vehicle sensor 15 detects various information (vehicle information) obtained from the host vehicle 1. The vehicle sensor 15 includes, for example, a vehicle speed sensor that detects the traveling speed (vehicle speed) of the host vehicle 1, a wheel speed sensor that detects the rotational speed of each tire provided on the host vehicle 1, a three-axis acceleration sensor (G sensor) that detects the acceleration in the three-axis directions of the host vehicle 1 (including deceleration), a steering angle sensor that detects the steering angle of the steering wheel, a steering angle sensor that detects the steering angle of the steering wheel, a gyro sensor that detects the angular velocity generated in the host vehicle 1, a yaw rate sensor that detects the yaw rate, an accelerator sensor that detects the operation amount of the accelerator pedal of the host vehicle 1, and a brake sensor that detects the brake operation amount by the driver.

[0016] The controller 16 is an electronic control unit (ECU: Electronic Control Unit) that performs the driving control of the host vehicle 1. During the driving control of the host vehicle 1, the controller 16 automatically controls the driving of the host vehicle 1 based on the surrounding environment. The controller 16 may be configured as a single electronic control unit or as a collection of a plurality of electronic control units. The controller 16 includes a processor 20 and peripheral components such as a storage device 21. The processor 20 may be, for example, a CPU (Central Processing Unit) or an MPU (Micro-Processing Unit). <OO00099>The storage device 21 may include a semiconductor storage device, a magnetic storage device, an optical storage device, etc. The storage device 21 may include memories such as registers, cache memories, a ROM (Read Only Memory) and a RAM (Random Access Memory) used as a main storage device. The functions of the controller 16 described below are realized, for example, when the processor 20 executes a computer program stored in the storage device 21.

[0017] The controller 16 may be formed by dedicated hardware for performing the information processing described below. For example, the controller 16 may include functional logic circuits set in a general-purpose semiconductor integrated circuit. For example, the controller 16 may have a programmable logic device (PLD) such as a field-programmable gate array (FPGA).

[0018] The actuator 17 operates the accelerator opening and brake system of the vehicle 1 in response to a control signal from the controller 16 to generate a driving force to propel the vehicle 1 or a braking force to brake the vehicle 1. The actuator 17 comprises an accelerator opening actuator and a brake control actuator. The accelerator opening actuator controls the accelerator opening of the vehicle 1. The brake control actuator controls the braking operation of the brake system of the vehicle 1. Furthermore, the actuator 17 may also include a steering actuator that controls the steering direction and steering amount of the steering mechanism of the vehicle 1. The actuator 17 may also operate the steering mechanism of the vehicle 1 in response to a control signal from the controller 16.

[0019] Next, the driving control of the vehicle 1 by the controller 16 will be described. The driving control by the controller 16 may include automatic control that estimates the driving path of the vehicle 1 along the first road R1 on which the vehicle 1 travels, and controls at least the steering angle of the vehicle 1 based on the estimated path. For example, the driving control by the controller 16 may include lane departure prevention control (lane keeping control) that controls the steering angle so that the vehicle 1 travels along the estimated path.

[0020] In this specification, "roadway" may refer to, for example, the area on which the vehicle 1 travels. That is, the roadway may be an area with a width in the direction of the road width. In this case, for example, the controller 16 may estimate the left and right boundaries of the area that forms the roadway. For example, the controller 16 may estimate the lane on which the vehicle 1 travels (the vehicle's lane) as the roadway. Alternatively, the controller 16 may estimate a line or sequence of points representing the trajectory on which its own vehicle 1 travels as the "path." For example, the controller 16 may estimate a line or sequence of points representing the trajectory on which the reference point (e.g., the center of gravity) of its own vehicle 1 moves as the "path."

[0021] In areas where there are no lanes, such as within an intersection, the controller 16 may estimate the area on which its own vehicle 1 travels, or lines or sequences of points representing the trajectory of its own vehicle 1, as the "road." The controller 16 obtains the road shape RS1 of the first road R1 on which its vehicle 1 travels (hereinafter sometimes referred to as "first road shape RS1") from the map data stored in the map database 12 and estimates the route.

[0022] As described above, the map data stored in the map database 12 represents road shapes on a road-by-road basis using a node sequence. In this type of map data, as shown in Figure 2(a), each node N often indicates a position near the center of the road. Therefore, as the number of road lanes increases, the position of node N in the lane width direction shifts in the direction of the increase in lanes.

[0023] In the example shown in Figure 2(a), the number of lanes increases as the lane of the first road R1 branches into a straight lane LN1 and a right-turn lane LN2 in front of the vehicle 1's path. As a result, the position of node N in the lane width direction shifts towards the right-turn lane LN2. Therefore, the first road shape RS1, represented by the node sequence, differs from the actual lane shape of the first road R1. In the example in Figure 2(a), the position of node N shifts by half a lane's width towards the right-turn lane LN2 as the number of lanes increases by one. As a result, the first road shape RS1 differs from the actual shapes of the straight-ahead lane LN1 and the right-turn lane LN2. As a result, estimating the route Ld that vehicle 1 should travel on based on the road shape of the map data may lead to low estimation accuracy.

[0024] Therefore, in this invention, it is determined whether or not the vehicle's track Ld can be estimated using the first road shape RS1, based on the respective distances between node N of the first road shape RS1 and the left and right road boundaries EL and ER of the first road R1, and the road width W of the first road R1. Specifically, the controller 16 detects road boundaries EL and ER extending along the direction of the first road R1 at both ends of the road width direction of the first road R1 based on the output signal of the external sensor 14, and converts these positions into positions on map data. Furthermore, the controller 16 calculates the road width W of the first road R1 from the distance between the road boundaries EL and ER at both ends in a direction perpendicular to the arrangement direction of nodes N of the first road shape RS1.

[0025] The controller 16 estimates the track Ld, which is a driving trajectory along the first road R1, based on the first road shape RS1, when the road width W is constant and the position of node N of the first road shape RS1 is in the center of the road boundaries EL and ER (i.e., the distance LL from node N to the road boundaries EL and ER is equal to LR). On the other hand, as shown in Figure 2(b), within a predetermined distance d in front of the vehicle 1, the road width W changes and the position of node N of the first road shape RS1 is in the center of the road boundaries EL and ER. In In this case, the controller 16 does not estimate the track Ld based on the first road shape RS1. In this case, the track Ld is estimated without using the first road shape RS1, as described later.

[0026] According to this embodiment, it is possible to determine whether the first road shape RS1 obtained from map data differs from the actual lane shape of the first road R1. If the first road shape RS1 differs from the actual lane shape, it is possible to choose to estimate the road Ld without using the first road shape RS1. This makes it possible to suppress the decrease in accuracy of road estimation by estimating the vehicle's road path using the first road shape RS1.

[0027] Next, the driving control device 10 in the embodiment will be described in more detail. Figure 3 is a block diagram of an example of the functional configuration of the controller 16 in Figure 1. The controller 16 includes a vehicle position estimation unit 40, a road shape acquisition unit 41, a road boundary detection unit 42, a road width boundary distance calculation unit 43, a road course estimation unit 44, and a vehicle control unit 45. The vehicle position estimation unit 40 estimates the vehicle's position in a fixed coordinate system (hereinafter sometimes referred to as "map coordinate system") of map data stored in the map database 12, based on the measurement results from the positioning device 11. Methods for estimating the vehicle's position include using a Kalman filter with GNSS, but any vehicle position estimation method that can obtain the general shape of the road the vehicle is traveling on is acceptable.

[0028] The road shape acquisition unit 41 identifies the first road R1 on which the vehicle 1 is currently traveling, based on the estimated position of the vehicle itself, from among the roads listed in the map data stored in the map database 12. The road shape acquisition unit 41 may also identify the first road R1 based on the planned driving route information of the vehicle 1 set by the navigation system (not shown). The road shape acquisition unit 41 extracts the first road shape RS1 of the first road R1 from the map database 12.

[0029] The road boundary detection unit 42 detects road boundary EL and ER extending along the direction of the first road R1 at both the left and right ends in the road width direction of the first road R1 by detecting road shoulder curbs, median strips, white lines (solid lines), road boundary lines, guardrails, poles, fences, cat's eyes, or blocks, etc. that indicate the road boundary. Although methods using recognition cameras or LiDAR are common for detection, any method that can achieve the above objective is acceptable.

[0030] The road width boundary distance calculation unit 43 calculates the positions of road boundaries EL and ER detected by the road boundary detection unit 42 in the map coordinate system based on the vehicle's position in the map coordinate system estimated by the vehicle position estimation unit 40. The road width boundary distance calculation unit 43 calculates the road width W of the first road R1 using the first road shape RS1 acquired by the road shape acquisition unit 41 and the road boundaries EL and ER. The road width boundary distance calculation unit 43 also calculates the distances LL and LR from each node N of the first road shape RS1 to the road boundaries EL and ER. For example, the road width boundary distance calculation unit 43 may extend perpendicular lines at each node N of the first road shape RS1 in a direction perpendicular to the connection direction (link direction) with adjacent preceding and succeeding nodes, find the intersection points of the road boundaries EL and ER with the perpendicular lines, and calculate the distances between the intersection points and node N as distances LL and LR. In other words, the road width boundary distance calculation unit 43 calculates the distances between node N and road boundaries EL and ER in a direction perpendicular to the arrangement direction of node N as distances LL and LR. Note that it is not necessary to calculate distances LL and LR for every node N included in road shape RS1; distances LL and LR may be calculated for every other node N or every few nodes N.

[0031] Furthermore, the road width boundary distance calculation unit 43 calculates the sum of distances LL and LR as road width W = LL + LR. The road width boundary distance calculation unit 43 then sets a predetermined distance d and determines whether the road width W is constant or changing within the predetermined distance d in front of the vehicle 1. For example, the road width boundary distance calculation unit 43 may determine whether the road width W is constant or not based on whether the amount of change in the road width W within the predetermined distance d is within a predetermined value.

[0032] Furthermore, the road width boundary distance calculation unit 43 determines whether the distances LL and LR from node N to the left and right road boundaries EL and ER are equal. That is, it determines whether the position of node N is in the center of the road boundaries EL and ER at both ends of the first road R1. For example, the road width boundary distance calculation unit 43 may determine whether distances LL and LR are equal to each other based on whether the sum or average of the differences between distances LL and LR calculated for each node is within a predetermined value Th (|LR|≦Th). The predetermined distance d can be set to a range in which the road width boundary distance calculation unit 43 can reliably calculate the road width W and distances LL and LR, for example, by using the maximum detectable distance of road boundaries EL and ER by the road boundary detection unit 42 as the maximum value.

[0033] The road width boundary distance calculation unit 43 determines whether or not to estimate the road Ld on which the vehicle 1 travels based on the first road shape RS1, according to these determination results. Specifically, the road width boundary distance calculation unit 43 determines that, within a predetermined distance d in front of its own vehicle 1, if the road width W is constant and the position of the node of the first road shape RS1 is in the center of the road boundaries EL and ER, it estimates the road Ld based on the first road shape RS1.

[0034] On the other hand, if the road width W changes within a predetermined distance d in front of the vehicle 1 and the position of the node of the first road shape RS1 is in the middle of the road boundaries EL and ER at both ends, it is determined to estimate the road Ld without using the first road shape RS1. The road width boundary distance calculation unit 43 of the first embodiment determines that the road width W is changing within a predetermined distance d in front of the vehicle 1 and that the position of the node of the first road shape RS1 is not in the center of the road boundaries EL and ER at both ends, and that it estimates the road Ld based on the first road shape RS1.

[0035] The road estimation unit 44 estimates the road Ld based on the determination result from the road width boundary distance calculation unit 43. If it is determined that the road Ld should be estimated based on the first road shape RS1, the road estimation unit 44 may estimate the road Ld by, for example, aligning the position of the first road shape RS1 with the current position of the vehicle 1. For example, the first road shape RS1 may be translated such that the position of the node in the vicinity of the vehicle's position (for example, the nearest node) moves to the vehicle's position, and the line represented by the first road shape RS1 after the translation or a predetermined width region including the first road shape RS1 may be estimated as the road Ld.

[0036] On the other hand, if it is determined that the road shape RS1 should not be used to estimate the road Ld, the road estimation unit 44 may estimate the road Ld based on the surrounding environment of the vehicle 1 detected by the external sensor 14. As a road estimation based on the surrounding environment, for example, the road estimation unit 44 may estimate the road Ld based on the detection results of white lines (road boundary lines, lane marking lines, lane boundary lines) around the vehicle 1. Alternatively, as a road estimation based on the surrounding environment, the road estimation unit 44 may estimate the road Ld based on the driving trajectories of other vehicles around the vehicle 1 (for example, a preceding vehicle traveling in front of the vehicle 1).

[0037] For example, if the route estimation unit 44 determines that the number of lanes on the first road R1 increases ahead of the vehicle 1's path, it may estimate the route Ld based on the shape of the road boundary EL and ER on the left and right of the first road R1 that is in line with the planned route of the vehicle 1. For example, the route estimation unit 44 may determine that the number of lanes on the first road R1 increases ahead of the vehicle 1's path based on the recognition results of road signs and road markings around the vehicle 1. Refer to Figure 4. For example, the road estimation unit 44 may detect road markings Ts that indicate a divergence between a straight lane and a right-turn lane by analyzing the image of the road surface in the lane ahead of the vehicle 1, which is captured by the camera of the external sensor 14.

[0038] Figure 5 is a schematic diagram of an example of a road marking Ts that warns of a divergence between a straight lane and a right-turn lane. The road marking Ts is a road marking in which an arrow indicating straight ahead and an arrow indicating right turn are placed within a predetermined distance of each other in the direction of the road width, indicating that the lane in which the road marking Ts is installed is diverging into a straight lane and a right-turn lane ahead. For example, if the planned route of vehicle 1 is a route that goes straight through an intersection, the route Ld may be estimated using the road boundary shape of the left road boundary EL. If the planned route of vehicle 1 is a route that turns right at an intersection, the route Ld may be estimated using the road boundary shape of the right road boundary ER.

[0039] Refer to Figure 4. Now, let's assume that there is an intersection section Sc ahead of the path of vehicle 1, and that in the section Su before the intersection section Sc (i.e., the section closer to vehicle 1 than the intersection section Sc), the path Ld1 of vehicle 1 is estimated without using the first road shape RS1 (i.e., using the surrounding environment). Thus, if the road Ld1 is estimated in the section Su before the intersection without using the first road shape RS1, the road estimation unit 44 may also estimate the road Ld2 in the intersection section Sc without using the first road shape RS1.

[0040] This is because, within intersections such as crossroads, there is a high probability that road boundaries EL and ER do not exist to indicate the direction of travel of vehicle 1, and therefore it is not possible to determine whether or not it is appropriate to estimate the road path Ld using the first road shape RS1 based on the road boundaries EL and ER. Furthermore, whether or not a section Su, where the road Ld1 is estimated without using the first road shape RS1, is before an intersection can be determined in the map database 12 based on whether or not it is within a predetermined distance from the intersection node Nc where the road shape of the intersecting road intersects with the first road shape RS1.

[0041] Furthermore, when estimating the road path Ld2 within the intersection section Sc without using the first road shape RS1, the road path estimation unit 44 may, for example, estimate the road path Ld2 based on the driving trajectories of other vehicles around its own vehicle 1 as part of the surrounding environment. This is because road path estimation based on white lines is highly unlikely to be possible in the intersection section Sc. Furthermore, in the distant section Sf ahead of the vehicle 1's path (hereinafter sometimes referred to as "distant section Sf"), a temporary track Ld3 may be estimated using the first road shape RS1, and a trajectory that smoothly connects the track Ld1 estimated based on the surrounding environment in the section Su before the intersection with the temporary track Ld3 may be estimated as the track Ld2 within the intersection section Sc.

[0042] This makes it possible to suppress errors in the road length (Ld) caused by estimating based on an incorrect road shape in sections where there is no recognition of the surrounding environment. For example, the distant section Sf may be a section that is at least a predetermined distance d away from the vehicle 1. For example, the distant section Sf may be a section where it has not yet been determined whether or not the road Ld can be estimated using the first road shape RS1, based on the road boundaries EL and ER and the road width W of the first road R1. For example, as in the example in Figure 4, if the section Su from which the road Ld1 is estimated without using the first road shape RS1 is the section before the intersection, then the distant section Sf may be a section further than the intersection from the perspective of the vehicle 1. Furthermore, the track estimation unit 44 may calculate, for example, a clothoid curve or a polynomial curve of degree 3 or higher as a track that smoothly connects the track Ld1 and the temporary track Ld3.

[0043] Furthermore, if the road width W is constant but the distances LL and LR are not equal (i.e., the position of node N in the first road shape RS1 is not in the center of the road boundaries EL and ER), it is possible that the accuracy of the first road shape RS1 is poor. In this case, the road estimation unit 44 may estimate the road Ld based on the center line of the road boundary. Furthermore, even if the road width boundary distance calculation unit 43 determines that the road path estimation unit 44 should estimate the road path Ld based on the first road shape RS1, the road path estimation unit 44 may estimate the road that is smoothly connected overall as the road path Ld of the vehicle 1, based on the white line shape detected as the surrounding environment and the driving trajectory of other vehicles. The vehicle control unit 45 drives the actuator 17 to control the accelerator and brakes, as well as the steering direction and amount of the steering mechanism, so that the vehicle 1 travels along the track Ld estimated by the track estimation unit 44.

[0044] (operation) Figure 6 is a flowchart of an example of the path estimation method of the first embodiment. In step S1, the vehicle position estimation unit 40 acquires map data from the map database 12. In step S2, the vehicle position estimation unit 40 estimates the vehicle's position in the map coordinate system based on the measurement results from the positioning device 11. In step S3, the road shape acquisition unit 41 extracts the first road shape RS1 of the first road R1 on which the vehicle 1 is currently traveling. In step S4, the road boundary detection unit 42 detects road boundaries EL and ER at both the left and right ends in the road width direction of the first road R1, which extend along the direction of extension of the first road R1.

[0045] In step S5, the road width boundary distance calculation unit 43 calculates the road width W of the first road R1 and the distances LL and LR from each node N of the first road shape RS1 to the road boundaries EL and ER. In step S6, the road width boundary distance calculation unit 43 determines whether the road width W is constant within a predetermined distance d in front of the vehicle 1. If the road width W is constant (step S6:Y), the process proceeds to step S7. If the road width W is not constant (step S6:N), the process proceeds to step S10.

[0046] In step S7, the road width boundary distance calculation unit 43 determines whether distances LL and LR are equal. If distances LL and LR are equal (step S7:Y), the process proceeds to step S8. If distances LL and LR are not equal (step S7:N), the process proceeds to step S9. In step S8, the track estimation unit 44 estimates the track Ld based on the first road shape RS1. The process then proceeds to step S12. In step S9, the track estimation unit 44 estimates the track Ld without using the first road shape RS1. The process then proceeds to step S12.

[0047] In step S10, the road width boundary distance calculation unit 43 determines whether distances LL and LR are equal. If distances LL and LR are equal (step S10:Y), the process proceeds to step S11. If distances LL and LR are not equal (step S10:N), the process proceeds to step S8. In this case, the track estimation unit 44 estimates the track Ld based on the first road shape RS1. The process then proceeds to step S12. In step S11, the track estimation unit 44 estimates the track Ld without using the first road shape RS1. The process then proceeds to step S12. In step S12, the vehicle control unit 45 controls the accelerator and brakes, as well as the steering direction and amount of the steering mechanism, so that the vehicle 1 travels along the track Ld estimated by the track estimation unit 44. The process then ends.

[0048] (Second Embodiment) The road width boundary distance calculation unit 43 of the second embodiment determines that it will estimate the road Ld based on the first road shape RS1 only if the road width W is changing within a predetermined distance d in front of the vehicle 1 and the position of the node of the first road shape RS1 is not in the center of the road boundaries EL and ER at both ends, and the distance between either of the road boundaries EL and ER at both ends and the first road shape RS1 does not change before and after the point where the change in road width W begins. Figure 7 is a schematic diagram illustrating the track estimation method of the second embodiment.

[0049] The road width boundary distance calculation unit 43, when the predetermined distance d in front of the vehicle 1 includes a section where the road width W is constant and a section where the road width W changes, sets up multiple sections within the predetermined distance d that include a section where the road width W is constant and a section where the road width W changes. For example, the road width boundary distance calculation unit 43 sequentially calculates the road width W for each node N within a predetermined distance d in front of the vehicle 1, determines whether the road width W has changed based on whether the difference in the road width W calculated at adjacent nodes is greater than or equal to a predetermined value, and sets up multiple sections by dividing the section at the nodes where it is determined that the road width W has changed.

[0050] In the example shown in Figure 7, the road width W is constant in section Sa, while the road width W changes in section Sb. The road width boundary distance calculation unit 43 determines whether or not to estimate the road Ld based on the first road shape RS1 in section Sb, and determines whether or not the distance LL or LR has changed between section Sa and section Sb. That is, in section Sb where the road width W changes, it determines whether or not the distance LL or LR has changed compared to section Sa before the change in road width W.

[0051] If either distance LL or LR does not change, the road width boundary distance calculation unit 43 determines to estimate the road Ld based on the first road shape RS1. If both distance LL or LR changes, the road width boundary distance calculation unit 43 determines to estimate the road Ld without using the first road shape RS1. In the example in Figure 7, LL and LR are not equal in section Sb, but since the distance LL is constant between section La and section Lb, the shape does not differ from the actual lane shape due to the increase in lanes. Therefore, it can be determined that the lane Ld can be estimated based on the first road shape RS1.

[0052] Figure 8 is a flowchart of an example of the path estimation method in the second embodiment. The processes from steps S21 to S24 are the same as the processes from steps S1 to S4 in Figure 6. In step S25, the road width boundary distance calculation unit 43 sets up a plurality of sections within a predetermined distance d in front of the vehicle 1, including sections where the road width W is constant and sections where the road width W changes. In step S26, the road width boundary distance calculation unit 43 calculates the road width W and the distances LL and LR to the road boundary for each of the multiple sections.

[0053] In step S27, the road width boundary distance calculation unit 43 selects one of the multiple sections set in step S25 as target sections, either in order of proximity to the vehicle 1 or in order of distance from the vehicle 1. In step S28, the road width boundary distance calculation unit 43 determines whether the road width W is constant in the target section. If the road width W is constant (step S28:Y), the process proceeds to step S29. If the road width W is not constant (step S28:N), the process proceeds to step S32.

[0054] In step S29, the road width boundary distance calculation unit 43 determines whether distances LL and LR are equal in the target section. If distances LL and LR are equal (step S29:Y), the process proceeds to step S30. If distances LL and LR are not equal (step S29:N), the process proceeds to step S31. In step S30, the road width boundary distance calculation unit 43 sets the target section to the section for which the road Ld is estimated based on the first road shape RS1. The process then proceeds to step S35. In step S31, the road width boundary distance calculation unit 43 sets the target section to a section in which the road Ld is estimated without using the first road shape RS1. The process then proceeds to step S35.

[0055] In step S32, the road width boundary distance calculation unit 43 determines whether distances LL and LR are equal in the target section. If distances LL and LR are equal (step S32:Y), the process proceeds to step S34. If distances LL and LR are not equal (step S32:N), the process proceeds to step S33. In step S33, the road width boundary distance calculation unit 43 determines whether either distance LL or LR is the same between the previous target section and the current target section. If either distance LL or LR is the same (step S33:Y), the process proceeds to step S30. If both distance LL and LR change (step S33:N), the process proceeds to step S34.

[0056] In step S34, the road width boundary distance calculation unit 43 sets the target section to a section in which the road Ld is estimated without using the first road shape RS1. The process then proceeds to step S35. In step S35, the road width boundary distance calculation unit 43 determines whether there are any sections remaining from the multiple sections set in step S25 that have not yet been selected in step S27. If there are sections that have not been selected (step S35:N), the process returns to step S27. In step S27, the section following the section that was immediately selected is selected as the target section, and steps S28 to S35 are repeated.

[0057] If there are no unselected intervals remaining (step S35:Y), the process proceeds to step S37. In step S36, the road estimation unit 44 estimates the road Ld for each of the multiple sections based on the settings made by the road width boundary distance calculation unit 43. The process in step S37 is the same as the process in step S12 in Figure 6. The process then terminates.

[0058] (Third embodiment) Figure 9 is a schematic diagram illustrating the track estimation method of the third embodiment. In the third embodiment, the road shape acquisition unit 41 extracts the second road shape RS2, which is the road shape of the opposing road R2 (hereinafter sometimes referred to as "second road R2"), which is the road on which the vehicle traveling opposite to the vehicle 1 is traveling, from the map database 12. If the road width boundary distance calculation unit 43 determines that the road path Ld should not be estimated using the first road shape RS1, the road path estimation unit 44 estimates the road path Ld of the vehicle 1 using the second road shape RS2. This is because, in places where the number of lanes increases, such as before an intersection, the road width W changes, and if the first road shape RS1 differs from the actual lane shape, the second road shape RS2 of the opposing road (second road R2) may be usable for road path estimation without being affected by the increase in lanes.

[0059] When estimating the track Ld without using the first road shape RS1, the track estimation unit 44 may determine whether or not to estimate the track Ld using the second road shape RS2. When determining whether or not to estimate the road Ld using the second road shape RS2, the road estimation unit 44, for example, moves the positions of the nodes of the second road shape RS2 in the road width direction to the positions where the nodes of the first road shape RS1 are arranged. For example, the road estimation unit 44 may move the positions of the nodes of the second road shape RS2 in the road width direction such that the position of node Nc2 near the midpoint of the second road R2 in the section between intersection C1 behind the vehicle 1 and intersection C2 in front of the vehicle 1 moves to the position of node Nc1 near the midpoint of the first road R1 in the section between intersection C1 and intersection C2. Alternatively, for example, the position of the nodes of the second road shape RS2 may be moved in the road width direction such that the node Nn2 of the second road shape RS2 closest to the vehicle 1 moves to the position of node Nn1 of the first road shape RS1 closest to the vehicle 1.

[0060] The road estimation unit 44 uses the second road shape RS2, similar to the case of the first road shape RS1, to calculate the distances LL and LR from the nodes of the second road shape RS2 to the left and right road boundaries EL and ER, respectively, and the road width W of the first road R1. The track estimation unit 44 may estimate the track Ld based on the second road shape RS2 after movement if the road width W is constant and the position of the node of the second road shape RS2 after movement is in the center of the road boundary EL, ER, or if the road width W changes and the position of the node of the second road shape RS2 after movement is not in the center of the road boundary EL, ER.

[0061] Furthermore, if the road width W changes and the position of the node of the second road shape RS2 after movement is in the center of the road boundaries EL and ER, the road Ld may be estimated without using the second road shape RS2 after movement. In this case, for example, the track estimation unit 44 may estimate the track Ld based on the surrounding environment of the vehicle 1.

[0062] (Fourth Embodiment) Figure 10 is a schematic diagram illustrating the route estimation method of the fourth embodiment. In the fourth embodiment, the route estimation unit 44 corrects the first road shape RS1 based on the amount of change in road width W in section S2 when the road width W changes as in section S2 and the position of the node of the first road shape is in the center of the road boundaries EL and ER (i.e., when the distances LL and LR are equal). The route estimation unit 44 estimates the route Ld based on the corrected first road shape RS1. In Figure 10, the rectangular plots represent the positions of node N in the corrected first road shape RS1, showing how the position of node N in the road width direction has been corrected.

[0063] For example, consider a case where, in section S1 adjacent to section S2 (i.e., on the opposite side of the direction of travel of vehicle 1), the road width W is a constant value W1 and the distances LL and LR are equal, and in section S2, the road width W changes from W1 to W2. The track estimation unit 44 corrects the road width direction position of the nodes of the first road shape RS1 in section S2 based on the difference |W2-W1| between road widths W1 and W2, and estimates the track Ld based on the corrected first road shape RS1. For example, the road estimation unit 44 may set the maximum correction amount Δm (Δ3 in the example in Figure 10) to half of the difference in road width W (|W2-W1| / 2), and set the correction amounts Δ1, Δ2, etc. for node positions within section S2 so that the correction amount gradually increases from 0 to Δm as the road width W changes from W1 to W2 in section S2.

[0064] The direction of correction for the node position may be set, for example, according to whether the road boundary along the planned travel path of the vehicle 1 is road boundary EL or ER, and whether the change in road width W is increasing or decreasing. For example, if the planned route of vehicle 1 is a route that goes straight through an intersection and the road width W increases in section S2, the position of node N in the road width direction may be corrected so that the distance between the left road boundary EL and the node position decreases. That is, as shown in Figure 10, the position of node N in the road width direction may be corrected to move to the left. Furthermore, for example, if the planned route of vehicle 1 involves turning right at an intersection and the road width W increases in section S2, the road width position of node N may be corrected so that the distance between the right-side road boundary ER and the node position decreases. In other words, the road width position of node N may be corrected to move to the right.

[0065] Furthermore, for example, if the planned route of vehicle 1 is a route that goes straight through an intersection and the road width W decreases in section S2, the road width position of node N may be corrected so that the distance between the left road boundary EL and the node position increases. In other words, the road width position of node N may be corrected to move to the right. Furthermore, for example, if the planned route of vehicle 1 involves turning right at an intersection and the road width W decreases in section S2, the road width position of node N may be corrected so that the distance between the right-side road boundary ER and the node position increases. In other words, the road width position of node N may be corrected to move to the left.

[0066] As shown in Figure 10, if section S2 is sandwiched between section S1 and section S3, and the road width W in section S1 is a constant value W1, and the road width W in section S3 is a constant value W2 (≠W1), then the position of the node of the first road shape RS1 in section S3 may be corrected based on the difference |W2-W1|. For example, the correction amount for the road width direction position of the node in section S3 may be set to a constant value Δ3 = |W2 - W1| / 2, as shown in Figure 10.

[0067] However, as shown in Figure 4, if the road width W increases along with the increase in the number of lanes of the first road R1 before the intersection section Sc, the number of lanes will decrease to the original number in section Sf after passing the intersection section Sc. In such cases, the node positions may be set so that the first road shape RS1 of the intersection section Sc and the section Su before it are smoothly connected to the first road shape RS1 after passing the intersection section Sc. For example, in section S3 of Figure 10, the position of node N in the road width direction may be set such that the distance LL between node N and the left road boundary EL gradually decreases. Therefore, in section S3, the correction amount for the road width direction position of node N may be set to decrease as it moves away from section S2.

[0068] (Effects of the embodiment) (1) The positioning device 11 detects the vehicle's current position. The controller 16 extracts the first road shape from map data, which is the road shape of the first road on which the vehicle 1 travels, from map data in which road shapes are represented by an array of multiple nodes. The controller 16 detects the road boundaries extending along the direction of the first road at both ends in the road width direction of the first road, calculates the position of the road boundaries on the map data based on the vehicle's position, calculates the road width of the first road from the distance between the road boundaries at both ends in a direction perpendicular to the array direction of the nodes of the first road shape, estimates the driving trajectory along the first road based on the first road shape if the road width is constant and the position of the node of the first road shape is in the center of the road boundaries at both ends within a predetermined distance in front of the vehicle 1, and estimates the driving trajectory without using the first road shape if the road width changes and the position of the node of the first road shape is in the center of the road boundaries at both ends within a predetermined distance in front of the vehicle 1.

[0069] Locations where the road width changes are considered to be sections where the road structure of the first road on which vehicle 1 is traveling has changed, such as an increase in lanes. In such locations, if the first road shape obtained from map data is in the center of the left and right road boundaries, it is considered that the first road shape is also changing in the same direction as the change in road width, and does not represent the actual shape of the lane on which vehicle 1 is traveling. In such cases, the driving path is estimated without using the first road shape, which can suppress misestimation of the driving path (decreased accuracy) based on an incorrect first road shape.

[0070] (2) The controller 16 may estimate the route based on the first road shape if the road width changes within a predetermined distance in front of its own vehicle 1 and the position of the node of the first road shape is not in the center of the road boundary at both ends. As a result, if the road width changes and the first road shape in the map data is not in the center of the left and right road boundaries, the first road shape is not affected by the factors causing the change in road width, so the road can be estimated using the first shape. (3) The controller 16 may estimate the road based on the first road shape only if the road width changes and the position of the node of the first road shape is not in the center of the road boundary at both ends, and the distance between either of the road boundary at both ends and the first road shape does not change before and after the point where the change in road width begins. This allows the route to be estimated based on the correct first road shape, provided that the distance from the first road shape in the map data to either the left or right road boundary has not changed since the road width was constant.

[0071] (4) When the controller 16 estimates the route without using the first road shape, it may extract the second road shape, which is the road shape of the second road on which the vehicle oncoming the controller 1 travels, from the map data and estimate the route using the second road shape. The controller 16 moves the position of the node of the second road shape in the road width direction to the position where the node of the first road shape is arranged, calculates the road width of the first road from the distance between the road boundaries at both ends in a direction perpendicular to the arrangement direction of the node of the second road shape after the move, and estimates the road based on the second road shape after the move if the road width is constant and the position of the node of the second road shape after the move is in the center of the road boundaries at both ends, or if the road width changes and the position of the node of the second road shape after the move is not in the center of the road boundaries at both ends, within a predetermined distance in front of the vehicle 1. The first road shape may be incorrect due to the increase in lanes near the intersection, but the opposing road does not have an increase in lanes, making the second road shape likely to be correct. Therefore, the correct route can be estimated by using the second road shape.

[0072] (5) The controller 16 may correct the first road shape based on the amount of change in road width when the road width changes and the position of the node of the first road shape is in the middle of the road boundary at both ends, and estimate the road based on the corrected first road shape. This allows for the correction of an incorrect first road shape by predicting the amount of change in the first road shape due to changes in road width. (6) The controller 16 may estimate the route based on the center line of the road boundary when the road width is constant and the position of the node of the first road shape is not in the center of the road boundary at both ends. When the road width is constant and the position of the node is not in the center of the left and right boundaries from the first road shape, there is a high possibility that the first road shape is incorrect. In that case, estimating the route based on the center line of the left and right boundaries can be expected to lead to the estimation of a correct route.

[0073] (7) When the controller 16 estimates the route without using the first road shape in the section before the intersection, it may also estimate the route within the intersection without using the first road shape. In many cases, there are no road boundaries such as curbs to explain the direction of travel of the vehicle within a four-way intersection, and if the road shape is incorrect before the intersection, it is likely that the road shape will also be incorrect within the intersection. For this reason, estimating the route without using the first road shape can suppress incorrect route estimation within the intersection. (8) The controller 16 may detect the surrounding environment of its own vehicle 1 and estimate the route without using the first road shape, based on the detected surrounding environment. For example, the surrounding environment may include lane markings around the own vehicle 1 or the driving trajectories of other vehicles around the own vehicle 1. This allows for a more accurate estimation of the track by utilizing the surrounding environment, such as white lines and vehicle trajectories. [Explanation of symbols]

[0074] 1...Vehicle, 10...Driving control device, 12...Map database, 14...External environment sensor, 15...Vehicle sensor, 16...Controller, 17...Actuator, 20...Processor, 21...Storage device, 40...Vehicle position estimation unit, 41...Road shape acquisition unit, 42...Road boundary detection unit, 43...Road width boundary distance calculation unit, 44...Road estimation unit, 45...Vehicle control unit

Claims

1. The vehicle's current position is detected, From map data in which road shapes are represented on a road-by-road basis by an array of multiple nodes, the first road shape, which is the road on which the vehicle is traveling, is extracted. The road boundaries extending along the direction of extension of the first road at both ends in the road width direction of the first road are detected, Based on the vehicle's position, the position of the road boundary on the map data is calculated. The road width of the first road is calculated from the distance between the road boundaries at both ends in a direction perpendicular to the arrangement direction of the nodes of the first road shape. Within a predetermined distance in front of the vehicle, depending on whether the road width is constant and the position of the node of the first road shape is at the center of the road boundary at both ends, or whether the road width changes within the predetermined distance in front of the vehicle and the position of the node of the first road shape is at the center of the road boundary at both ends, it is determined whether or not to estimate a road that is a driving trajectory along the first road based on the first road shape. If it is determined that a road, which is a driving trajectory along the first road, is not to be estimated based on the first road shape, then the road shape of the second road, which is adjacent to the first road in the road width direction and is the road on which the vehicle traveling opposite to the first road travels, is extracted from the map data, and the road shape of the second road, which is represented by an array of nodes different from the array of nodes representing the first road shape, is used to estimate the road. A method for estimating a running path, characterized by the features described above.

2. The vehicle's current position is detected, From map data in which road shapes are represented on a road-by-road basis by an array of multiple nodes, the first road shape, which is the road on which the vehicle is traveling, is extracted. The road boundaries extending along the direction of extension of the first road at both ends in the road width direction of the first road are detected, Based on the vehicle's position, the position of the road boundary on the map data is calculated. The road width of the first road is calculated from the distance between the road boundaries at both ends in a direction perpendicular to the arrangement direction of the nodes of the first road shape. Within a predetermined distance in front of the vehicle, depending on whether the road width is constant and the position of the node of the first road shape is at the center of the road boundary at both ends, or whether the road width changes within the predetermined distance in front of the vehicle and the position of the node of the first road shape is at the center of the road boundary at both ends, it is determined whether or not to estimate a road that is a driving trajectory along the first road based on the first road shape. If it is determined that a road, which is a driving trajectory along the first road, is not to be estimated based on the first road shape, then the lane markings around the vehicle or the driving trajectories of other vehicles around the vehicle are detected as the surrounding environment of the vehicle, and the road is estimated based on the detected surrounding environment. A method for estimating a running path, characterized by the features described above.

3. The road estimation method according to claim 1 or 2, characterized in that, within the predetermined distance in front of the vehicle, the road width changes and the position of the node of the first road shape is not the center of the road boundary at both ends, the road is estimated based on the first road shape.

4. The method for estimating a road based on the first road shape, characterized in that, if the road width changes and the position of the node of the first road shape is not at the center of the road boundary at both ends, the distance between either of the road boundaries at both ends and the first road shape does not change before and after the point where the change in road width begins.

5. Move the position of the node of the second road shape to the position where the nodes of the first road shape are arranged in the road width direction. The width of the first road is calculated from the distance between the road boundaries at both ends in a direction perpendicular to the arrangement direction of the nodes of the second road shape after movement. Within a predetermined distance in front of the vehicle, if the road width is constant and the position of the node of the second road shape after movement is at the center of the road boundary at both ends, or if the road width changes and the position of the node of the second road shape after movement is not at the center of the road boundary at both ends, the road is estimated based on the second road shape after movement. The track estimation method according to feature 1.

6. A method for estimating a road, characterized in that, when the road width changes from a first road width to a second road width in a certain section and the position of the node of the first road shape is at the center of the road boundary at both ends, half of the difference obtained by subtracting the first road width from the second road width is calculated as the maximum correction amount, the first road shape is corrected by moving the position of the node of the first road shape in the road width direction by a correction amount that changes from 0 to the maximum correction amount as the road width in the section changes from the first road width to the second road width, and the road is estimated based on the corrected first road shape.

7. A method for estimating a road route according to any one of claims 1 to 6, characterized in that, when the road width is constant and the position of the node of the first road shape is not the center of the road boundary at both ends, the road route is estimated based on the center line of the road boundary.

8. A method for estimating a road path according to any one of claims 1 to 4, characterized in that if it is determined that the road path should not be estimated based on the first road shape in the section before the intersection, it is determined that the road path should not be estimated based on the first road shape even within the intersection.

9. It comprises a positioning device that detects the vehicle's current position, a sensor that detects objects around the vehicle, and a controller. The aforementioned controller, From map data in which road shapes are represented on a road-by-road basis by an array of multiple nodes, the first road shape, which is the road on which the vehicle is traveling, is extracted. Based on the output of the sensor, the road boundaries extending along the direction of the first road at both ends in the road width direction of the first road are detected. Based on the vehicle's position, the position of the road boundary on the map data is calculated. The road width of the first road is calculated from the distance between the road boundaries at both ends in a direction perpendicular to the arrangement direction of the nodes of the first road shape. Within a predetermined distance in front of the vehicle, depending on whether the road width is constant and the position of the node of the first road shape is at the center of the road boundary at both ends, or whether the road width changes within the predetermined distance in front of the vehicle and the position of the node of the first road shape is at the center of the road boundary at both ends, it is determined whether or not to estimate a road that is a driving trajectory along the first road based on the first road shape. If it is determined that a road, which is a driving trajectory along the first road, is not to be estimated based on the first road shape, then the road shape of the second road, which is adjacent to the first road in the road width direction and is the road on which the vehicle traveling opposite to the first road travels, is extracted from the map data, and the road shape of the second road, which is represented by an array of nodes different from the array of nodes representing the first road shape, is used to estimate the road. A track estimation device characterized by the following features.

10. A positioning device that detects the current position of the vehicle, a sensor that detects objects around the vehicle, and a controller, The aforementioned controller, From map data in which road shapes are represented on a road-by-road basis by an array of multiple nodes, the first road shape, which is the road on which the vehicle is traveling, is extracted. Based on the output of the sensor, the road boundaries extending along the direction of the first road at both ends in the road width direction of the first road are detected. Based on the vehicle's position, the position of the road boundary on the map data is calculated. The road width of the first road is calculated from the distance between the road boundaries at both ends in a direction perpendicular to the arrangement direction of the nodes of the first road shape. Within a predetermined distance in front of the vehicle, depending on whether the road width is constant and the position of the node of the first road shape is at the center of the road boundary at both ends, or whether the road width changes within the predetermined distance in front of the vehicle and the position of the node of the first road shape is at the center of the road boundary at both ends, it is determined whether or not to estimate a road that is a driving trajectory along the first road based on the first road shape. If it is determined that a road, which is a driving trajectory along the first road, is not to be estimated based on the first road shape, then the lane markings around the vehicle or the driving trajectories of other vehicles around the vehicle are detected as the surrounding environment of the vehicle, and the road is estimated based on the detected surrounding environment. A track estimation device characterized by the following features.

Images