JP7563263B2 - Track estimation method and track estimation device - Google Patents

Description

The present invention relates to a route estimation method and a route estimation device.

There is a known invention that recognizes the route that the vehicle is planning to take when passing through an intersection (Patent Document 1). The invention described in Patent Document 1 sets a processing area within a captured image, and recognizes the route that the vehicle is planning to take within the set processing area.

JP 2019-70895 A

However, the invention described in Patent Document 1 is an invention that recognizes the white lines beyond the intersection exit, so if the intersection is large and has a long section without white lines, there is little information beyond the intersection exit, and there is a risk that the lane within the intersection cannot be recognized.

The present invention was made in consideration of the above problems, and its purpose is to provide a roadway estimation method and a roadway estimation device that can estimate the shape of the road on which the vehicle is traveling.

A lane estimation method according to one aspect of the present invention estimates the position of the vehicle on a map, calculates the degree of deviation between a lane shape model and the general shape of the road, and the degree of deviation between the lane shape model and the curvature of a right-turn or left-turn path in an intersection section, sets a weight value so that the weight value related to the deviation of the general shape of the road is greater than the weight value related to the deviation of the curvature when the vehicle travels straight through the intersection section, sets a weight value so that the weight value related to the deviation of the curvature is greater than the weight value related to the deviation of the general shape of the road when the vehicle travels right or left when the vehicle travels through the intersection section, multiplies each deviation by the set weight value, estimates a lane shape model such that the combined deviation degree obtained by adding these is equal to or less than a predetermined value, and estimates the road shape based on the estimated lane shape model.

The present invention makes it possible to estimate the shape of the road on which the vehicle is traveling.

FIG. 1 is a configuration diagram of a roadway estimation device 1 according to an embodiment of the present invention. FIG. 2 is a diagram illustrating an intersection section. FIG. 3 is a diagram for explaining a lane shape model. FIG. 4 is a diagram for explaining the degree of deviation between the curvature detection result and the lane shape model. FIG. 5 is a diagram for explaining the degree of deviation between the detection result of lane markings and a lane shape model. FIG. 6 is a diagram for explaining the degree of deviation between the detection result of the general shape of the road and the lane shape model. FIG. 7 is a diagram for explaining the degree of deviation between the detection result of the arrangement of road structures and the lane shape model. FIG. 8 is a diagram for explaining the degree of deviation between the detection result of the movement trajectory of another vehicle and the lane shape model. FIG. 9 is a flowchart illustrating an example of the operation of the travel path estimation device 1.

Embodiments of the present invention will be described below with reference to the drawings. In the description of the drawings, identical parts are given the same reference numerals and the description will be omitted.

An example of the configuration of the path estimation device 1 will be described with reference to FIG. 1. The path estimation device 1 may be mounted on a vehicle with an automatic driving function, or on a vehicle without an automatic driving function. The path estimation device 1 may also be mounted on a vehicle that is capable of switching between automatic driving and manual driving. The automatic driving function may also be a driving assistance function that automatically controls only some of the vehicle control functions, such as steering control, braking force control, and driving force control, to assist the driver in driving. In this embodiment, the path estimation device 1 will be described as being mounted on a vehicle with an automatic driving function.

As shown in FIG. 1, the roadway estimation device 1 includes a GNSS receiver 10, a sensor 11, a map database 12, a controller 20, and various actuators 13. The various actuators 13 include, for example, a steering actuator, an accelerator pedal actuator, and a brake actuator.

The GNSS receiver 10 detects the position information of the vehicle on the ground by receiving radio waves from artificial satellites. A Kalman filter may be used for the detection. The position information of the vehicle detected by the GNSS receiver 10 includes latitude information and longitude information. The method of detecting the position information of the vehicle is not limited to the GNSS receiver 10. For example, the position may be estimated using a method called odometry. Odometry is a method of estimating the position of the vehicle by determining the amount and direction of movement of the vehicle according to the rotation angle and rotation angular velocity of the vehicle. The GNSS receiver 10 outputs the detected position information to the controller 20.

The sensor 11 is used to detect objects around the vehicle (pedestrians, bicycles, motorcycles, other vehicles, etc.) and information about the vehicle's surroundings (division lines, traffic lights, signs, pedestrian crossings, intersections, etc.). The sensor 11 includes a camera, lidar, radar, laser range finder, sonar, etc. The sensor 11 outputs the detected information to the controller 20.

The map database 12 is a database stored in a car navigation device or the like, and stores map information necessary for route guidance, such as road information and facility information. The map database 12 in this embodiment does not have high-precision information. High-precision information is, for example, information on a lane-by-lane basis with an accuracy of centimeters. However, the map database 12 in this embodiment does not have such high-precision information. The information held by the map database 12 is, for example, information on the general shape of roads composed of nodes and links.

The controller 20 is a general-purpose microcomputer equipped with a CPU (Central Processing Unit), a memory, and an input/output unit. A computer program for functioning as the road estimation device 1 is installed in the microcomputer. By executing the computer program, the microcomputer functions as multiple information processing circuits equipped in the road estimation device 1. Note that here, an example is shown in which the multiple information processing circuits equipped in the road estimation device 1 are realized by software, but it is of course possible to configure the information processing circuits by preparing dedicated hardware for executing each information process shown below. The multiple information processing circuits may also be configured by individual hardware. The controller 20 includes, as an example of multiple information processing circuits, a vehicle position estimation unit 21, a curvature estimation unit 22, a lane marking detection unit 23, a structure detection unit 24, an object detection and tracking unit 25, a section setting unit 26, a weight setting unit 27, a model parameter estimation unit 28, and a road shape estimation unit 29.

The vehicle position estimation unit 21 estimates the position of the vehicle on the map database 12 based on the information acquired from the GNSS receiver 10. The vehicle position estimation unit 21 also acquires the general shape of the roads around the vehicle based on the acquired vehicle position.

The curvature estimation unit 22 estimates the curvature of the right-turn or left-turn route that the vehicle is scheduled to travel in the intersection section. Whether the planned travel route of the vehicle is a right turn or a left turn can be obtained from the travel direction of the route information to the destination set in advance by the user in the navigation device. As an example of a method for estimating the curvature, it can be obtained from the radius of a circle that is tangent to the entrance point and exit point of the vehicle in the intersection section and a local straight line calculated from the general shape of the road (see radius r in FIG. 4). For the entrance point and the exit point, the intersection section can be obtained, and the entrance point and the exit point can be calculated from the position when the vehicle enters the intersection section, the general shape of the road, and the route. When calculating the radius of the circle, the entrance point and the exit point can be set at a position offset by a predetermined value from the intersection of the general shape of the road of the straight road and the right and left turns, and the lateral position of the entrance point can be set based on the position when the vehicle enters the intersection section. Alternatively, a stop line and a crosswalk can be detected using a camera, a lidar, etc., and the stop line near the entrance of the intersection can be set as the entrance point. Similarly, an exit point may be set from a crosswalk near the intersection exit or a stop line on the oncoming lane near the intersection exit.

The lane marking detection unit 23 detects lane markings around the vehicle using information acquired from the sensor 11. Lane markings are lines used to separate multiple lanes on roads with two or more lanes on each side, and are also called lane boundary lines or dividing lines.

The structure detection unit 24 detects the arrangement of structures around the vehicle using information acquired from the sensor 11. In this embodiment, structures refer to road boundaries such as curbs, medians, and side strips.

The object detection and tracking unit 25 uses information acquired from the sensor 11 to detect and track other vehicles (preceding vehicles) traveling ahead of the vehicle. The object detection and tracking unit 25 also acquires the movement trajectory of the other vehicles. One example of a tracking method is a filtering method such as a particle filter, but tracking is not limited to this method and can be performed using any known method.

The section setting unit 26 sets intersection sections on roads on the map database 12. In this embodiment, an intersection section is a section having a predetermined size, and is set on roads on the map database 12. When setting an intersection section, for example, if information on intersections or traffic lights installed at intersections is stored in the map database 12, then that location may be set as the intersection section. Alternatively, the intersection section may be set by enlarging the map database 12 at a predetermined magnification, with the intersection point at which the general shapes of the roads intersect as the center. When setting an intersection section, the section setting unit 26 first sets an estimated route range on the guidance route, and sets the intersection section within that range. The guidance route refers to the route to the destination input by the user.

Now, the intersection section will be described with reference to FIG. 2. In this embodiment, the intersection section refers to the area surrounded by the reference symbol 30 in FIG. 2. That is, in this embodiment, the intersection section refers to an area that includes the entire intersection. One definition of "the entire intersection" is as follows: if a line is drawn between the stop line and the crosswalk as shown in FIG. 2, it is defined as the area inside that line. However, this definition is merely an example, and another definition would be to draw a line on the crosswalk and define the area inside that line.

The weight setting unit 27 sets the weight depending on whether the vehicle is traveling straight ahead, turning right, or turning left when traveling through the intersection section.

The model parameter estimation unit 28 estimates parameters of a preset lane shape model. In the lane shape model, the shape is expressed by, for example, a cubic function. Specifically, it is expressed by f(x)= ^ax3 + ^bx2 +cx. However, this is only an example, and the lane shape model may be expressed by other functions. Here, the lane shape model will be described with reference to FIG. 3. As shown in FIG. 3, the lane shape model is composed of start points ( _V1 , _V2 , V3, _V4 , _V9 , _V10 ) of lanes ( _L1 , _L2 , _L3 , _L4 ) in a section (section set by the section setting unit ₂₆ ), goal points ( _V5 , _V6 , _V7 , _V8 , _V11 ), and shapes ( _f1 , _f2 , _f3 , _f4 , _f5 ). The start points and goal points are expressed in a local coordinate system. The finish point in the section is expressed as the start point of the next section in the road travel direction. If each point V _i has its coordinate values x _s , y _s and shape parameters a, b, c, and the coordinate values x _e , _ye of the finish point as parameters, these parameters are estimated using information acquired from the sensor 11. Note that x _s and x _e may be initial values or fixed values corresponding to the start and finish of the set section. The cubic function may be expressed as f(x) = ax ³ + bx ² + cx + d. In this case, the parameter d is the y coordinate. The x coordinate can be obtained by offsetting the coordinate system of the link. The number of lanes may be obtained from the map database 12 and the detection results of lane markings. Alternatively, the number of lanes may be set as a parameter, and the number of multiple lanes (one lane or more) may be set as an initial value and estimated together with the lane shape model using information acquired from the sensor 11.

The parameter estimation method will be described with reference to Figs. 4 to 8. In Fig. 4, the input data is the curvature of the intersection section. Reference numeral 50 in Fig. 4 is the detection result of the curvature of the right-turn route in the intersection section. Note that Fig. 4 shows the curvature of the right-turn route and the curvature of the left-turn route, but the curvature of the right-turn route will be described here, and the curvature of the left-turn route will be omitted. Of course, the same control as that for turning right can be applied when turning left. Based on the detection result of the curvature of the intersection section, parameters (specifically, _xs , _ys , a, b, c, _xe , _ye , hereinafter simply referred to as parameters) are estimated, and a lane shape model (lane markings estimated using the lane shape model) is generated. The generated lane shape model is indicated by reference numeral 40. In this embodiment, the difference between reference numeral 50 and reference numeral 40 is called the deviation degree. The smaller the deviation degree, the more accurately the road shape is estimated. If the degree of deviation between the code 50 and the code 40 is _Cc , _Cc is expressed by the following equation 1.

ここでＥ_C（Ｐ_ｍ，ｐ_C）は最近傍点間のノルムである。

Here, E _C (P _m , p _C ) is the norm between the nearest neighbors.

In FIG. 5, the input data are lane markings detected by the lane marking detection unit 23. Reference numeral 41 in FIG. 5 denotes the detection result of the lane markings. Parameters (specifically, _xs , _ys , a, b, c, _xe , _ye , hereinafter simply referred to as parameters) are estimated based on the detection result of the lane markings, and a lane shape model (lane markings estimated using the lane shape model) is generated. The generated lane shape model is indicated by reference numeral 40. In this embodiment, the difference between reference numeral 41 and reference numeral 40 is called the deviation degree. The smaller this deviation degree, the more accurately the road shape is estimated. If the deviation degree between reference numeral 41 and reference numeral 40 is _CRP , _CRP is expressed by Equation 2.

ここでＥ_ＲＰ（Ｐ_ｍ，ｐ_ｄ）は最近傍点間のノルムである。

Here, E _RP (P _m , p _d ) is the norm between the nearest neighbors.

In Fig. 6, the input data is the general shape of the road. Reference numeral 42 in Fig. 6 indicates the result of obtaining the general shape of the road from the map database 12 and generating the general shape at each start point. Based on the result of generating the general shape of the road, parameters are estimated and a lane shape model is generated. The generated lane shape model is indicated by reference numeral 40. If the degree of deviation between reference numeral 43 and reference numeral 40 is _CSD , _CSD is expressed by Equation 3. Note that here, the general shape of the road is usually the shape of a link connecting node points stored in the map database 12 as a line segment along the center of the road width direction, and reference numeral 42 in Fig. 6 indicates a line segment obtained by offsetting the link shape to match the start points ( _V1 , _V2 , _V3 , _V4 ).

ここでＥ_ＳＤ（Ｐ_ｍ，ｐ_s）は最近傍点間のノルムである。

Here, E _SD (P _m , p _s ) is the norm between the nearest neighbors.

In Fig. 7, the input data is the arrangement of road structures (arrangement of curbs, median strips, etc.). That is, as shown in Fig. 7, it is an arrangement of structures arranged along the road (arrangement that can be represented by line segments), and may be simply referred to as road structures or structures below. Reference numeral 43 in Fig. 7 is the detection result of the arrangement of road structures. Parameters are estimated based on the detection result of the arrangement of road structures, and a lane shape model is generated. The generated lane shape model is indicated by reference numeral 40. If the degree of deviation between reference numeral 43 and reference numeral 40 is _CRF , _CRF is expressed by Equation 4.

ここでＥ_ＲＦ（Ｐ_ｍ，ｐ_f）は最近傍点間のノルムである。

Here, E _RF (P _m , p _f ) is the norm between the nearest neighbors.

In Fig. 8, the input data is the movement trajectory of another vehicle. Reference numeral 45 in Fig. 8 is the detection result of the movement trajectory of another vehicle. The movement trajectory of the other vehicle is offset in the lane width direction based on the vehicle width of the other vehicle or a predetermined lane width to calculate a point cloud of reference numeral 44. Parameters are estimated based on the detection result of the movement trajectory of the other vehicle, and a lane shape model is generated. The generated lane shape model is indicated by reference numeral 40. If the degree of deviation between reference numeral 44 and reference numeral 40 is _CTL , _CTL is expressed by Equation 5.

ここでＥ_ＴＬ（Ｐ_ｍ，ｐ_tｌ）は最近傍点間のノルムである。

Here, E _TL (P _m , p _tl ) is the norm between the nearest neighbors.

The model parameter estimation unit 28 optimizes the parameters to minimize these deviations. The sum of each deviation is calculated, using the weights set by the weight setting unit 27. If the sum of each deviation is C, the integrated deviation C is expressed by Equation 6.

λ_１，λ_２，λ_３，λ_４，λ_５は重み設定部２７によって設定される重みの値である。

λ ₁ , λ ₂ , λ ₃ , λ ₄ , and λ ₅ are weight values set by the weight setting unit 27 .

An example of the setting of λ ₁ , λ ₂ , λ ₃ , λ ₄ , and λ ₅ will be described. As a concept of the setting, the higher the reliability or importance of the information, the larger the weight value. For example, when the traveling direction of the vehicle when traveling through the intersection section is straight, since the general shape of the road is generally straight, λ ₁ is set to 0.95 and λ ₂ is set to 0.05. Even when the general shape of the road is emphasized, if it is known in advance by learning or the like that a higher weight of the curvature element such as a Y-shaped straight line can be used to estimate a route with higher accuracy, the value of λ ₂ may be increased, such as λ ₁ to 0.8 and λ ₂ to 0.2. In addition, when the traveling direction of the vehicle when traveling through the intersection section is a right turn or a left turn, since the general shape of the road is generally curved, the value of the curvature is emphasized, and λ ₂ is set to 0.9 and λ ₁ to 0.1. In addition, when it is known in advance that the curvature values obtained by the curvature estimation tend to be large, the weighting value of the general shape of the road can be increased to represent straight portions when turning right or left, and _λ2 can be set to 0.7 and _λ1 can be set to 0.3.

In addition to the above, when the movement trajectory of another vehicle is acquired, the information is useful for identifying the road shape for both going straight and turning right and left. Therefore, when the traveling direction of the vehicle when traveling through the intersection section is going straight, λ ₃ is set to 0.7, λ ₁ is set to 0.25, and λ ₂ is set to 0.05. Alternatively, when it is known in advance that the accuracy of the general shape of the road acquired from the map database 12 is high, the weight value of the general shape of the road may be increased while maintaining the weight relationship, and λ ₃ may be set to 0.5, λ ₁ to 0.45, and λ ₂ to 0.05. When the traveling direction of the vehicle when traveling through the intersection section is turning right or left, λ ₃ is set to 0.7, λ ₂ to 0.2, and λ ₁ to 0.1. When the reliability of the curvature estimation is high like going straight, λ ₃ may be set to 0.5, λ ₂ to 0.4, and λ ₁ to 0.1.

The structures in the intersection section are useful for identifying the shape of the roadway taking into account the entrance and exit of the intersection section. On the other hand, the structures in the intersection section are the curbs and median strips that correspond to the road boundaries, and the general shape of the road is more useful information for the shape inside the intersection. Based on the above, when the traveling direction of the host vehicle when traveling through the intersection section is straight, _λ3 is set to 0.6, _λ1 is set to 0.2, _λ4 is set to 0.15, and _λ2 is set to 0.05. When the traveling direction of the host vehicle when traveling through the intersection section is a right or left turn, _λ3 is set to 0.5, _λ2 is set to 0.25, _λ1 is set to 0.15, and _λ4 is set to 0.1.

There is a possibility that there are no lane markings in the intersection section, or that false detection occurs due to a pedestrian crossing. However, since there is a possibility that guiding lines exist in the intersection section, lane markings such as guiding lines can also be taken into consideration by adopting a small value for the weight of the lane markings. Therefore, when the traveling direction of the host vehicle when traveling in the intersection section is straight, λ ₃ is set to 0.6, λ ₁ is set to 0.25, λ ₅ is set to 0.1, and λ ₂ is set to 0.05. When the traveling direction of the host vehicle when traveling in the intersection section is a right turn or a left turn, λ ₃ is set to 0.5, λ ₂ is set to 0.35, λ ₁ is set to 0.1, and λ ₅ is set to 0.05. When the traveling direction of the host vehicle when traveling in the intersection section is straight, λ ₃ may be set to 0.4, λ ₁ to 0.3, λ ₄ to 0.15, λ ₅ to 0.1, and λ ₂ to 0.05. When the vehicle is turning right or left when traveling through an intersection section, λ ₃ may be set to 0.4, λ ₂ to 0.3, λ ₁ to 0.15, λ ₄ to 0.1, and λ ₅ to 0.05. Note that the weights are set to values between 0 and 1 for ease of use, but are not limited to these values.

The model parameter estimation unit 28 uses the weight values set in this way to estimate parameters so as to minimize the integrated deviation degree C. Note that minimization techniques are well known and will not be described in detail. An example of a minimization technique is the least squares method. Note that it is not necessary to estimate parameters so as to minimize them. It is sufficient to estimate parameters so that they are equal to or less than a predetermined value. The predetermined value can be obtained through experiments and simulations.

The road shape estimation unit 29 estimates the road shape based on the model parameter estimation results for each section of the road estimation range. The road shape estimation unit 29 estimates the road shape at a predetermined discrete interval for each section, and controls the actuator 13 so that the vehicle travels along the estimated road shape.

Next, an example of the operation of the roadway estimation device 1 will be described with reference to the flowchart in FIG. 9.

In step S101, the controller 20 acquires map information from the map database 12. The process proceeds to step S103, where the vehicle position estimation unit 21 estimates the position of the vehicle on the map database 12 based on the information acquired from the GNSS receiver 10. The process proceeds to step S105, where the curvature estimation unit 22 estimates the curvature of the right-turn route or left-turn route along which the vehicle is planned to travel in the intersection section. The process proceeds to step S107, where the section setting unit 26 sets an estimated route range on the guidance route. The process proceeds to step S109, where the section setting unit 26 sets an intersection section within the estimated route range. The process proceeds to step S111, where the controller 20 acquires whether the location where the vehicle is traveling is an intersection section or not.

If the location where the vehicle is traveling is an intersection section (YES in step S113), it is determined whether the traveling direction of the vehicle when traveling through the intersection section is straight, turns right, or turns left. If the traveling direction of the vehicle when traveling through the intersection section is straight (YES in step S115), the process proceeds to step S117, where the weight setting unit 27 sets the weight value so that the weight value related to the deviation degree of the general shape of the road is greater than the weight value related to the deviation degree of the curvature. On the other hand, if the traveling direction of the vehicle when traveling through the intersection section is a right turn or a left turn (NO in step S115), the process proceeds to step S119, where the weight setting unit 27 sets the weight value so that the weight value related to the deviation degree of the curvature is greater than the weight value related to the deviation degree of the general shape of the road. If the location where the vehicle is traveling is not an intersection section (NO in step S113), the process proceeds to step S123.

In step S121, the model parameter estimation unit 28 multiplies each deviation degree by the set weight value and estimates parameters so that the integrated deviation degree obtained by adding these is minimized. The process proceeds to step S123, where the controller 20 acquires the next section on the guidance route. If the section is within the estimation range (YES in step S125), the process returns to step S113. On the other hand, if the section is not within the estimation range (NO in step S125), the process proceeds to step S127, where the road shape estimation unit 29 estimates the road shape using the model parameters estimated in step S121.

(Action and Effect)
As described above, the path estimation method of the path estimation device 1 according to this embodiment provides the following advantageous effects.

The roadway estimation device 1 includes a map database 12 in which the general shapes of the roads connecting the nodes are stored, and a controller 20 that estimates the shape of the road along which the vehicle travels based on a preset lane shape model. The controller 20 estimates the position of the vehicle on the map, and calculates the degree of deviation between the lane shape model and the general shape of the road, and the degree of deviation between the lane shape model and the curvature of the right-turn or left-turn path in the intersection section. When the traveling direction of the vehicle when traveling through the intersection section is straight, the controller 20 sets a weight value so that the weight value related to the deviation degree of the general shape of the road is greater than the weight value related to the deviation degree of the curvature, and when the traveling direction of the vehicle when traveling through the intersection section is a right turn or a left turn, the controller 20 sets a weight value so that the weight value related to the deviation degree of the curvature is greater than the weight value related to the deviation degree of the general shape of the road. The controller 20 multiplies each deviation degree by a set weight value, estimates a lane shape model such that the combined deviation degree obtained by adding these is equal to or less than a predetermined value, and estimates the road shape based on the estimated lane shape model. When going straight through an intersection section, the road is generally linear, so the general shape of the road is close to the road shape. On the other hand, when turning right or left at an intersection section, the road is curved, so the shape calculated from the curvature is closer to the road shape than the combination of the general shapes of the two intersecting roads. Therefore, when going straight through an intersection section, the general shape of the road is weighted more heavily than the curvature, and when turning right or left at an intersection section, the curvature is weighted more heavily than the general shape of the road. This makes it possible to weight the more important information in each case, making it possible to estimate the road shape.

The controller 20 also acquires the movement trajectory of another vehicle traveling in the same direction as the traveling direction of the vehicle, and calculates the degree of deviation between the lane shape model and the movement trajectory of the other vehicle. When the traveling direction of the vehicle traveling in the intersection section is straight, the controller 20 sets the weight value so that the weight value related to the deviation degree of the general shape of the road is larger than the weight value related to the deviation degree of the curvature, and the weight value related to the deviation degree of the movement trajectory of the other vehicle is larger than the weight value related to the deviation degree of the general shape of the road. When the traveling direction of the vehicle traveling in the intersection section is a right turn or a left turn, the controller 20 sets the weight value so that the weight value related to the deviation degree of the curvature is larger than the weight value related to the deviation degree of the general shape of the road, and the weight value related to the deviation degree of the movement trajectory of the other vehicle is larger than the weight value related to the curvature. In intersections where there are no lane markings or where the detection results of lane markings are unreliable, the movement trajectories of other vehicles are important information. Therefore, by weighting the movement trajectories of other vehicles more heavily than the general shape and curvature of the road, both in going straight and in turning right or left, it is possible to estimate the road shape with high accuracy.

In addition, when calculating the degree of deviation between the lane shape model and the movement trajectory of another vehicle, the controller 20 calculates the degree of deviation between the lane shape model and the movement trajectory of the other vehicle that entered the intersection section from a direction intersecting the traveling direction of the own vehicle and proceeded to the exit of the intersection section in the same traveling direction as the own vehicle.
This makes it possible to use only the movement trajectories of other vehicles that have already traveled the route that the vehicle is scheduled to travel in the future to calculate the deviation degree. In other words, the movement trajectories of other vehicles preceding the vehicle are used to calculate the deviation degree, and since the movement trajectories of the preceding vehicles are highly reliable, it becomes possible to estimate the road shape with high accuracy.

The controller 20 also detects the arrangement of structures related to the road boundary around the vehicle and calculates the degree of deviation between the lane shape model and the arrangement of the structures. When the traveling direction of the vehicle when traveling through the intersection section is straight, the controller 20 sets the weighting value so that the weighting value related to the degree of deviation of the arrangement of the structures is greater than the weighting value related to the degree of deviation of the curvature, and the weighting value related to the degree of deviation of the general shape of the road is greater than the weighting value related to the degree of deviation of the arrangement of the structures. When the traveling direction of the vehicle when traveling through the intersection section is a right turn or a left turn, the controller 20 sets the weighting value so that the weighting value related to the degree of deviation of the general shape of the road is greater than the weighting value related to the degree of deviation of the arrangement of the structures, and the weighting value related to the degree of deviation of the curvature is greater than the weighting value related to the degree of deviation of the general shape of the road. The structures are curbs, medians, etc., and the information indicates the entrance and exit points of the intersection section and the drivable area. Because this is partial information, the weight of the arrangement of structures is smaller than the general shape of the road in the intersection section, but by weighting and taking it into consideration, it is possible to estimate the road shape with high accuracy.

In addition, when the direction of travel of the vehicle when traveling through an intersection section is straight, the controller 20 sets the weighting values so that the weighting value related to the degree of deviation of the arrangement of structures is greater than the weighting value related to the degree of deviation of the curvature, the weighting value related to the degree of deviation of the arrangement of structures is greater than the weighting value related to the degree of deviation of the general shape of the road, and the weighting value related to the degree of deviation of the movement trajectory of other vehicles is greater than the weighting value related to the degree of deviation of the general shape of the road. When the traveling direction of the vehicle when traveling through an intersection section is a right turn or a left turn, the controller 20 sets the weighting values so that the weighting value related to the deviation degree of the general shape of the road is greater than the weighting value related to the deviation degree of the arrangement of structures, the weighting value related to the deviation degree of the curvature is greater than the weighting value related to the deviation degree of the general shape of the road, and the weighting value related to the deviation degree of the movement trajectory of other vehicles is greater than the weighting value related to the curvature deviation. By adding the elements of the movement trajectories of other vehicles, it becomes possible to estimate the road shape with high accuracy.

The controller 20 also detects lane markings around the vehicle and calculates the degree of deviation between the lane shape model and the lane markings. When the vehicle travels straight through the intersection, the controller 20 sets the weighting values so that the weighting value related to the deviation of the lane markings is greater than the weighting value related to the deviation of the curvature, and the weighting value related to the deviation of the general shape of the road is greater than the weighting value related to the deviation of the lane markings. When the vehicle travels right or left through the intersection, the controller 20 sets the weighting values so that the weighting value related to the deviation of the general shape of the road is greater than the weighting value related to the deviation of the lane markings, and the weighting value related to the deviation of the curvature is greater than the weighting value related to the deviation of the general shape of the road. There may be no lane markings in the intersection, or lane markings may be erroneously detected due to a pedestrian crossing or the like. However, when there are guiding lines at an intersection, the information on lane markings is useful. Therefore, by incorporating lane markings with a weight smaller than the general shape of the road, a road shape that matches the lane markings can be estimated even when the general shape of the road is highly consistent. Therefore, road shapes in which the lane markings do not match the general shape of the road can be rejected, and when the lane markings, such as guiding lines, are thought to be of a correct shape, they can be used to estimate the road shape with high accuracy. Note that when driving straight, the road is generally linear, so guiding lines can be taken into account by weighting the lane markings higher than the curvature.

In addition, when the traveling direction of the vehicle when traveling through an intersection section is straight, the controller 20 sets the weighting values so that the weighting value related to the degree of deviation of the lane markings is greater than the weighting value related to the degree of deviation of the curvature, the weighting value related to the degree of deviation of the general shape of the road is greater than the weighting value related to the degree of deviation of the lane markings, and the weighting value related to the degree of deviation of the movement trajectory of other vehicles is greater than the weighting value related to the degree of deviation of the general shape of the road. When the traveling direction of the vehicle when traveling through an intersection section is a right turn or a left turn, the controller 20 sets the weighting values so that the weighting value related to the deviation degree of the general shape of the road is greater than the weighting value related to the deviation degree of the lane markings, the weighting value related to the deviation degree of the curvature is greater than the weighting value related to the deviation degree of the general shape of the road, and the weighting value related to the deviation degree of the movement trajectory of other vehicles is greater than the weighting value related to the curvature deviation. By adding the elements of the movement trajectories of other vehicles, it becomes possible to estimate the road shape with high accuracy.

In addition, when the vehicle's direction of travel when traveling through an intersection section is straight, the controller 20 sets the weighting values so that the weighting value related to the degree of deviation of the lane markings is greater than the weighting value related to the degree of deviation of the curvature, the weighting value related to the degree of deviation of the arrangement of structures is greater than the weighting value related to the degree of deviation of the lane markings, and the weighting value related to the degree of deviation of the general shape of the road is greater than the weighting value related to the degree of deviation of the arrangement of structures. When the vehicle is turning right or left while traveling through an intersection section, the controller 20 sets the weighting values so that the weighting value related to the deviation of the structure arrangement is greater than the weighting value related to the deviation of the lane markings, the weighting value related to the deviation of the general shape of the road is greater than the weighting value related to the deviation of the structure arrangement, and the weighting value related to the deviation of the curvature is greater than the weighting value related to the deviation of the general shape of the road. By adding the elements of the structures and lane markings, it becomes possible to estimate the road shape with high accuracy.

In addition, when the traveling direction of the vehicle when traveling through an intersection section is straight, the controller 20 sets the weighting values so that the weighting value related to the degree of deviation of the lane markings is greater than the weighting value related to the degree of deviation of the curvature, the weighting value related to the degree of deviation of the arrangement of structures is greater than the weighting value related to the degree of deviation of the lane markings, the weighting value related to the degree of deviation of the general shape of the road is greater than the weighting value related to the degree of deviation of the arrangement of structures, and the weighting value related to the degree of deviation of the movement trajectory of other vehicles is greater than the weighting value related to the degree of deviation of the general shape of the road. When the traveling direction of the vehicle when traveling through an intersection section is a right turn or a left turn, the controller 20 sets the weighting values so that the weighting value related to the degree of deviation of the arrangement of structures is greater than the weighting value related to the degree of deviation of lane markings, the weighting value related to the degree of deviation of the general shape of the road is greater than the weighting value related to the degree of deviation of the arrangement of structures, the weighting value related to the degree of deviation of the curvature is greater than the weighting value related to the degree of deviation of the general shape of the road, and the weighting value related to the degree of deviation of the movement trajectory of other vehicles is greater than the weighting value related to the degree of deviation of the curvature. By adding the elements of the movement trajectories of other vehicles, it becomes possible to estimate the road shape with high accuracy.

Each of the functions described in the above embodiments may be implemented by one or more processing circuits. Processing circuits include programmed processing devices, such as processors that include electrical circuitry. Processing circuits also include devices, such as application specific integrated circuits (ASICs) or circuit components, arranged to perform the described functions.

In the above embodiment, the curvature of the right-turn or left-turn path, the general shape of the road, lane markings, structures, and the movement trajectories of other vehicles are taken into consideration. However, the parameters may be estimated based on the degree of deviation (C _c ) between the lane shape model and the curvature of the right-turn or left-turn path and the degree of deviation (C _SD ) between the lane shape model and the general shape of the road, by taking into consideration only the curvature of the right-turn or left-turn path and the general shape of the road. In other words, using only the degree of deviation ( _Cc ) between the lane shape model and the curvature of the right-turn or left-turn path, and the degree of deviation ( _CSD ) between the lane shape model and the approximate shape of the road, weighting coefficients ( _λ1 , _λ2 ) can be set to multiply each degree of deviation depending on whether the vehicle is traveling straight, turning right, or turning left, and the _deviation degrees multiplied by the set weighting coefficients can be added to calculate the integrated deviation degree C (C = _λ1CSD + _{λ2Cc )} , and parameters can be estimated so that the integrated deviation degree C is less than or equal to a predetermined value. However, as in the above-described embodiment, by calculating the integrated deviation degree C taking into account not only the degree of deviation between the lane shape model and the curvature of the right-turn or left-turn path and the general shape of the road, but also the degree of deviation between the lane shape model and the arrangement of structures (C _RF ), the degree of deviation between the lane shape model and the movement trajectory of other vehicles (C _TL ), and the degree of deviation between the lane shape model and lane markings (C _RP ), more accurate parameter estimation is possible.

1 Roadway estimation device, 10 GNSS receiver, 11 sensor, 12 map database, 13 actuator, 20 controller, 21 vehicle position estimation unit, 22 curvature estimation unit, 23 lane marking detection unit, 24 structure detection unit, 25 object detection and tracking unit, 26 section setting unit, 27 setting unit, 28 model parameter estimation unit, 29 roadway shape estimation unit

Claims (10)

A lane estimation method for a lane estimation device including a map database storing an outline of a road connecting nodes and a controller estimating a shape of a lane on which a vehicle travels based on a preset lane shape model, comprising:
The controller:
Estimate the vehicle's position on a map;
Calculating a degree of deviation between the lane shape model and an approximate shape of the road, and a degree of deviation between the lane shape model and a curvature of a right-turn path or a left-turn path in an intersection section,
setting a weight value such that a weight value related to a deviation degree of the general shape of the road is greater than a weight value related to a deviation degree of the curvature when the vehicle travels in a straight line through the intersection section;
setting a weight value such that a weight value related to a degree of deviation of the curvature is greater than a weight value related to a degree of deviation of the general shape of the road when the traveling direction of the vehicle when traveling through the intersection section is a right turn or a left turn;
multiplying each deviation degree by a weighting value set, and estimating a lane shape model such that an integrated deviation degree obtained by adding these deviation degrees is equal to or smaller than a predetermined value;
A lane estimating method comprising: estimating a lane shape based on the estimated lane shape model. The controller:
Acquire a movement trajectory of another vehicle traveling in the same direction as the traveling direction of the host vehicle;
Calculating a degree of deviation between the lane shape model and the movement trajectory of the other vehicle;
setting a weight value such that, when the traveling direction of the vehicle when traveling through the intersection section is straight, a weight value related to a degree of deviation of the general shape of the road is greater than a weight value related to a degree of deviation of the curvature, and a weight value related to a degree of deviation of the movement trajectory of the other vehicle is greater than a weight value related to a degree of deviation of the general shape of the road;
The lane estimation method according to claim 1, characterized in that, when the direction of travel of the vehicle when traveling through the intersection section is a right turn or a left turn, the weighting value related to the degree of deviation of the curvature is set to be greater than the weighting value related to the degree of deviation of the general shape of the road, and the weighting value related to the degree of deviation of the movement trajectory of the other vehicle is set to be greater than the weighting value related to the degree of deviation of the curvature. The controller:
The lane estimation method according to claim 2, characterized in that, when calculating the degree of deviation between the lane shape model and the movement trajectory of the other vehicle, the movement trajectory of the other vehicle that entered the intersection section from a direction intersecting the traveling direction of the host vehicle and proceeded to an exit of the intersection section in the same traveling direction as the host vehicle is excluded from the calculation. The controller:
Detecting an arrangement of structures related to road boundaries around the vehicle;
Calculating a degree of deviation between the lane shape model and an arrangement of the structures;
setting a weight value such that, when the traveling direction of the vehicle when traveling through the intersection section is straight, a weight value related to a degree of deviation of the arrangement of the structures is greater than a weight value related to a degree of deviation of the curvature, and a weight value related to a degree of deviation of the general shape of the road is greater than a weight value related to a degree of deviation of the arrangement of the structures;
The lane estimation method according to claim 1, characterized in that, when the direction of travel of the vehicle when traveling through the intersection section is a right turn or a left turn, the weighting value related to the degree of deviation of the general shape of the road is set to be greater than the weighting value related to the degree of deviation of the arrangement of the structures, and the weighting value related to the degree of deviation of the curvature is set to be greater than the weighting value related to the degree of deviation of the general shape of the road. The controller:
Acquire a movement trajectory of another vehicle traveling in the same direction as the traveling direction of the host vehicle;
Calculating a degree of deviation between the lane shape model and the movement trajectory of the other vehicle;
when the traveling direction of the vehicle when traveling through the intersection section is straight, a weight value is set so that a weight value related to the degree of deviation of the arrangement of the structures is greater than a weight value related to the degree of deviation of the curvature, a weight value related to the degree of deviation of the arrangement of the structures is greater than a weight value related to the degree of deviation of the general shape of the road, and a weight value related to the degree of deviation of the movement trajectory of the other vehicle is greater than a weight value related to the degree of deviation of the general shape of the road;
The lane estimation method according to claim 4, characterized in that, when the direction of travel of the vehicle when traveling through the intersection section is a right turn or a left turn, the weighting value related to the degree of deviation of the general shape of the road is greater than the weighting value related to the degree of deviation of the arrangement of the structures, the weighting value related to the degree of deviation of the curvature is greater than the weighting value related to the degree of deviation of the general shape of the road, and the weighting value related to the degree of deviation of the movement trajectory of the other vehicle is greater than the weighting value related to the degree of deviation of the curvature. The controller:
Detecting lane markings around the vehicle;
Calculating a degree of deviation between the lane shape model and the lane markings;
setting weight values such that, when the traveling direction of the vehicle when traveling through the intersection section is straight, a weight value related to the deviation degree of the lane markings is greater than a weight value related to the deviation degree of the curvature, and a weight value related to the deviation degree of the general shape of the road is greater than a weight value related to the deviation degree of the lane markings;
The lane estimation method according to claim 1, characterized in that, when the direction of travel of the vehicle when traveling through the intersection section is a right turn or a left turn, the weighting value related to the degree of deviation of the general shape of the road is set to be greater than the weighting value related to the degree of deviation of the lane markings, and the weighting value related to the degree of deviation of the curvature is set to be greater than the weighting value related to the degree of deviation of the general shape of the road. The controller:
Acquire a movement trajectory of another vehicle traveling in the same direction as the traveling direction of the host vehicle;
Calculating a degree of deviation between the lane shape model and the movement trajectory of the other vehicle;
when the traveling direction of the host vehicle when traveling through the intersection section is straight, weight values are set so that a weight value related to the degree of deviation of the lane markings is greater than a weight value related to the degree of deviation of the curvature, a weight value related to the degree of deviation of the general shape of the road is greater than a weight value related to the degree of deviation of the lane markings, and a weight value related to the degree of deviation of the movement trajectory of the other vehicle is greater than a weight value related to the degree of deviation of the general shape of the road;
The lane estimation method according to claim 6, characterized in that, when the direction of travel of the vehicle when traveling through the intersection section is a right turn or a left turn, the weighting value related to the degree of deviation of the general shape of the road is greater than the weighting value related to the degree of deviation of the lane markings, the weighting value related to the degree of deviation of the curvature is greater than the weighting value related to the degree of deviation of the general shape of the road, and the weighting value related to the degree of deviation of the movement trajectory of the other vehicle is greater than the weighting value related to the degree of deviation of the curvature. The controller:
Detecting an arrangement of structures related to road boundaries around the vehicle;
Detecting lane markings around the vehicle;
Calculating a degree of deviation between the lane shape model and an arrangement of the structures, and a degree of deviation between the lane shape model and the lane markings;
when the traveling direction of the vehicle when traveling through the intersection section is straight, setting weight values so that a weight value related to the degree of deviation of the lane markings is greater than a weight value related to the degree of deviation of the curvature, a weight value related to the degree of deviation of the arrangement of the structures is greater than a weight value related to the degree of deviation of the lane markings, and a weight value related to the degree of deviation of the general shape of the road is greater than a weight value related to the degree of deviation of the arrangement of the structures;
The lane estimation method according to claim 1, characterized in that, when the direction of travel of the vehicle when traveling through the intersection section is a right turn or a left turn, the weighting value related to the degree of deviation of the arrangement of the structures is set to be greater than the weighting value related to the degree of deviation of the lane markings, the weighting value related to the degree of deviation of the arrangement of the structures is set to be greater than the weighting value related to the degree of deviation of the general shape of the road, and the weighting value related to the degree of deviation of the curvature is set to be greater than the weighting value related to the degree of deviation of the general shape of the road. The controller:
Acquire a movement trajectory of another vehicle traveling in the same direction as the traveling direction of the host vehicle;
Calculating a degree of deviation between the lane shape model and the movement trajectory of the other vehicle;
when the traveling direction of the host vehicle when traveling through the intersection section is straight, weight values are set so that the weight value related to the degree of deviation of the lane markings is greater than the weight value related to the degree of deviation of the curvature, the weight value related to the degree of deviation of the arrangement of the structures is greater than the weight value related to the degree of deviation of the lane markings, the weight value related to the degree of deviation of the general shape of the road is greater than the weight value related to the weight value related to the degree of deviation of the arrangement of the structures, and the weight value related to the degree of deviation of the movement trajectory of the other vehicle is greater than the weight value related to the degree of deviation of the general shape of the road;
The lane estimation method according to claim 8, characterized in that, when the traveling direction of the vehicle when traveling through the intersection section is a right turn or a left turn, the weighting value related to the degree of deviation of the arrangement of the structures is greater than the weighting value related to the degree of deviation of the lane markings, the weighting value related to the degree of deviation of the arrangement of the structures is greater than the weighting value related to the degree of deviation of the general shape of the road, the weighting value related to the degree of deviation of the curvature is greater than the weighting value related to the degree of deviation of the general shape of the road, and the weighting value related to the degree of deviation of the movement trajectory of the other vehicle is greater than the weighting value related to the curvature. A map database that stores the general shapes of roads that connect the nodes;
A controller that estimates a road shape on which the host vehicle is traveling based on a preset lane shape model,
The controller:
Estimate the vehicle's position on a map;
Calculating a degree of deviation between the lane shape model and an approximate shape of the road, and a degree of deviation between the lane shape model and a curvature of a right-turn path or a left-turn path in an intersection section,
setting a weight value such that a weight value related to a deviation degree of the general shape of the road is greater than a weight value related to a deviation degree of the curvature when the vehicle travels in a straight line through the intersection section;
setting a weight value such that a weight value related to a degree of deviation of the curvature is greater than a weight value related to a degree of deviation of the general shape of the road when the traveling direction of the vehicle when traveling through the intersection section is a right turn or a left turn;
multiplying each deviation degree by a weighting value set, and estimating a lane shape model such that an integrated deviation degree obtained by adding these deviation degrees is equal to or smaller than a predetermined value;
A lane estimating device which estimates a lane shape based on the estimated lane shape model.

Images