Passing location generation device

Abstract

The present disclosure relates to a passing location generation device for generating a passing location to be reached by a moving body, the passing location generation device comprising: a target state candidate generation unit for calculating a plurality of target state candidates including at least a state quantity of a position on the basis of surrounding environment information of the moving body and a state quantity of the moving body; and a local target state generation unit that calculates a local target state, which is a local target state that can reach an arbitrary target state candidate, at an intermediate point up to the periphery of the plurality of target state candidates, and outputs the local target state as a passing point.

Description

Technical Field

[0001] The present disclosure relates to a passage point generation device for generating a passage point for realizing automatic driving of a vehicle or the like. Background Art

[0002] In recent years, the development of autonomous mobile bodies such as self-driving cars and autonomous transport vehicles has been developed. These autonomous mobile bodies are controlled to reach a target state including a determined target position. However, depending on the situation, sometimes it is not possible to recognize that the target position cannot be passed until the target position is approached based on information obtained using external sensors. For example, in the scenario of passing through a toll station gate, consider the following scenario: when the target toll station gate is selected at a stage when the target vehicle is still a sufficient distance away from the toll station gate, it is not until the toll station gate that is initially targeted is approached that it is learned that the toll station gate is not in operation or a faulty vehicle is parked in the toll station. In such a scenario, it is necessary to change the target position. For example, as disclosed in Patent Document 1, a technology has been developed to correct the selected gate based on dynamic information obtained using sensors, etc., thereby achieving passage through a more suitable gate.

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: International Publication No. 2018 / 142561 Summary of the invention

[0006] Patent document 1 proposes: based on static information, a first gate is selected from multiple gates, and then, based on dynamic information, the selection result of the first gate is corrected to a second gate, and the gate selection result is corrected to select a more suitable gate technology and improve the probability of passing the gate. A vehicle control system is proposed.

[0007] However, when the vehicle is corrected to a gate far away from the first gate while traveling to the first gate, or when the vehicle is corrected to another gate just before reaching the first gate, the vehicle may need to control the steering wheel sharply to correct the track, or may not be able to find a track to reach the corrected gate.

[0008] The present disclosure is made to solve the above-mentioned problems, and an object of the present disclosure is to provide a passing point generating device that can reach the changed target state without causing a sudden trajectory correction even when the target state has to be changed.

[0009] The passing point generation device disclosed in the present invention is a passing point generation device that generates a passing point that a mobile body should reach, and comprises: a target state candidate generation unit, which calculates a plurality of target state candidates including at least a state quantity of a position based on the surrounding environment information of the mobile body and the state quantity of the mobile body; and a local target state generation unit, which calculates a local target state that can reach any target state candidate, i.e., a local target state, at an intermediate point around the plurality of target state candidates, and outputs the local target state as the passing point.

[0010] According to the passing point generating device disclosed in the present invention, as preparation for the change of the target state, the target state that can reach any target state candidate, that is, the local target state, is calculated at an intermediate point around a plurality of target state candidates, and the local target state is set as the passing point. Thus, even if the target state changes, it is possible to cope with it without making a sharp trajectory correction and reach the changed target state. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 This is a block diagram showing the structure of a mobile object equipped with the passing point generation device according to the first embodiment of the present disclosure.

[0012] Figure 2 The figure schematically shows a state in which a moving object attempts to enter a driving lane of a toll booth.

[0013] Figure 3 This is a flowchart showing an example of the operation of the passing point generating device according to the first embodiment of the present disclosure.

[0014] Figure 4 is a conceptual diagram illustrating a process of generating a target state candidate set.

[0015] Figure 5 is a conceptual diagram illustrating target state candidates.

[0016] Figure 6 is a conceptual diagram illustrating the reachable boundary.

[0017] Figure 7 It is a block diagram showing the structure of a modified example of the target state candidate generating unit.

[0018] Figure 8 This is a conceptual diagram for explaining the process of generating a virtual track until the vehicle is ready for travel.

[0019] Fig. 9 It is a conceptual diagram for explaining the processing in a modified example of the target state candidate calculation unit.

[0020] Fig.10This is a conceptual diagram for explaining the process of calculating the local target state in the local target state generation unit.

[0021] Fig.11 This is a conceptual diagram for explaining the process of calculating the local target state in the local target state generation unit.

[0022] Fig.12 This is a conceptual diagram for explaining the process of calculating the local target state in the local target state generation unit.

[0023] Fig.13 This is a conceptual diagram of the case where the local target state is set as a region.

[0024] Fig.14 This is a conceptual diagram of the case where the local target state is set as a probability distribution.

[0025] Fig.15 This is a conceptual diagram for explaining the processing of calculating a local target state when weighting target state candidates.

[0026] Fig.16 This is a conceptual diagram for explaining the processing of calculating the local target state when weighting the virtual track.

[0027] Fig.17 This is a conceptual diagram for explaining a process of generating a target trajectory of a moving object in a trajectory generating unit.

[0028] Fig.18 This is a conceptual diagram for explaining the process of generating a target trajectory from target state candidates in the trajectory generating unit.

[0029] Fig.19 This is a conceptual diagram showing an example of a sudden change in a local target state.

[0030] Fig. 20 This is a conceptual diagram showing an example of a sudden change in a local target state.

[0031] Fig.21 This is a conceptual diagram for explaining a process of imposing restrictions on abrupt changes in a local target state.

[0032] Fig. 22 This is a block diagram showing the structure of a mobile object equipped with the passing point generation device according to Embodiment 2 of the present disclosure.

[0033] Fig.23 This is a conceptual diagram for explaining a process of generating a wide-area target state in a wide-area target state generating unit.

[0034] Fig.24 This is a conceptual diagram for explaining the processing in the target state candidate calculation unit.

[0035] Fig.25 It is a conceptual diagram for explaining processing in a modified example of the target state candidate calculation unit.

[0036] Fig.26 This is a block diagram showing the structure of a control server equipped with a passage point generation device according to Embodiment 3 of the present disclosure.

[0037] Fig. 27 This is a conceptual diagram showing a system in which a local target state is transmitted from a control server to a mobile body through communication between the control server and the mobile body.

[0038] Fig.28 This is a diagram showing a hardware configuration for realizing the passage point generation device according to the first to third embodiments.

[0039] Fig.29 This is a diagram showing a hardware configuration for realizing the passage point generation device according to the first to third embodiments. DETAILED DESCRIPTION

[0040] <Implementation Method 1>

[0041] Figure 1 This is a block diagram showing the structure of the mobile object 1 equipped with the passing point generation device according to the first embodiment of the present disclosure.

[0042] Figure 1 The mobile body 1 shown in the figure includes: an autonomous sensor information acquisition unit 100 for acquiring information about the surrounding environment, the destination, and the state of the mobile body 1; a passing point generation device 200 for generating a passing point to be reached by the mobile body 1 based on the information obtained from the autonomous sensor information acquisition unit 100; and a motion control unit 300 for controlling the motion of the mobile body 1 so as to reach the local target state based on the local target state generated by the passing point generation device 200. In addition, the information obtained by the autonomous sensor information acquisition unit 100 is input to the passing point generation device 200 via the pre-processing unit 10, and the function of the pre-processing unit 10 will be described later. In addition, the local target state generated by the passing point generation device 200 is input to the motion control unit 300 via the post-processing unit 20, and the function of the post-processing unit 20 will be described later.

[0043] The autonomous sensor information acquisition unit 100 includes a surrounding environment information acquisition unit 110 and an autonomous state acquisition unit 120 .

[0044] The surrounding environment information acquisition unit 110 acquires the relative position, speed, azimuth, and obstacle-free traversable space information of moving obstacles such as walls and other moving objects around the moving object, for example, using a millimeter wave radar, camera, LiDAR (Light Detection And Ranging), or sonar installed on the moving object.

[0045] The surrounding environment information acquisition unit 110 also acquires information such as at least one possible drivable position, driving speed, and driving azimuth of the mobile body 1 as a driving target. They acquire information from, for example, information pre-specified by a user, a predetermined position of map information possessed by the mobile body 1, and the like. Here, the map information possessed by the mobile body 1 refers to a high-precision map, a vehicle navigation map, a point cloud map generated using SLAM (Simultaneous Localization and Mapping), and the like.

[0046] The drivable position may include, for example, the entrance of each gate or the position of the gate pole in the toll booth, the front wheel of each aircraft in the case of a towing tractor, at least one position of the mobile body 1 specified by the user, etc. The driving speed may include, for example, the legal speed, the speed specified in advance by the user, etc. The driving azimuth is the target angle when passing through the drivable position, and may include, for example, the direction in the vertical direction relative to the gate when passing through the toll booth gate, etc.

[0047] Figure 2 Schematically shows a state in which the moving body 1 attempts to enter a driving lane defined by left and right lane boundaries LB in front of the toll booth FS. Figure 2 An example of the possible travel position, travel speed, and travel azimuth when passing through the toll gate GT is shown in FIG. Figure 2 In the case where i is set as the suffix, (x pi ,y pi ) is the drivable position, v pi is the driving speed at each drivable position, θ pi is the driving azimuth at each drivable position.

[0048] The set of drivable positions is expressed by the following formula (1).

[0049] [Formula 1]

[0050] P p ={(x p1 ,y p1 ), (x p2 ,y p2 ), (x p3 ,yp3 )…(x pf ,y pf )}…(1)

[0051] The set of traveling speeds is expressed by the following formula (2).

[0052] [Formula 2]

[0053] V p = {v p1 , v p2 , v p3 …v pf}…(2)

[0054] The set of driving direction angles is expressed by the following formula (3).

[0055] [Formula 3]

[0056]

[0057] Here, the suffix f is the number at the end of the environment. Figure 2 In this example, the last digit is 7.

[0058] exist Figure 2 In the example of , when there are multiple drivable positions, there are also multiple driving speeds and driving azimuths, but they can also be set to one and do not need to correspond to the number of drivable positions.

[0059] In the following, the drivable state is defined based on each drivable position, driving speed, and driving azimuth, and the set of drivable positions, the set of driving speeds, and the set of driving azimuths are collectively referred to as a drivable state set.

[0060] In addition, Figure 2 In the figure, the ☆ mark schematically indicates the drivable position, the direction of the arrow schematically indicates the driving azimuth, and the length of the arrow schematically indicates the driving speed, which is also shared in other top views.

[0061] The self-state acquisition unit 120 has a plurality of sensors for acquiring the current state of the mobile body. Examples of the plurality of sensors include a velocity sensor, an acceleration sensor, an inertial measurement unit, a steering angle sensor, a steering torque sensor, a yaw rate sensor, and a global navigation satellite system (GNSS) sensor. Hereinafter, the inertial measurement unit is referred to as an IMU (Inertial Measurement Unit) sensor.

[0062] exist Figure 2 In the example, use (x e ,ye ,θ e , v e ) represents the current state of the moving body 1, x e is the x coordinate of the position of the moving body 1, y e is the y coordinate of the position of the moving body 1, θ e is the azimuth of mobile body 1, v e is the speed of the mobile body 1. In addition, the state quantity of the mobile body 1 is not limited to the position, azimuth and speed, and it is sufficient to include at least the state quantity of the position. In addition, it can also include other state quantities of the mobile body 1 such as acceleration, jerk, altitude, yaw rate, rotation angular velocity, and steering angle.

[0063] Here, return to Figure 1 Description. Figure 1 As shown, the passing point generation device 200 includes a target state candidate generation unit 210 and a local target state generation unit 220 .

[0064] When the mobile body 1 cannot travel in the target state (i.e., gate selected as the initial travel target from the set of passing states, i.e., all toll gates GT in the environment where the mobile body 1 exists), the target state candidate generation unit 210 generates a target state newly determined based on at least one drivable state obtained from the surrounding environment information acquisition unit 110 and the current state of the mobile body 1 obtained from the own state acquisition unit 120 as a target state candidate for changing the target state selected by the mobile body 1. Here, the target state refers to the state of one target that the mobile body 1 wants to pass through eventually.

[0065] The mobile body 1 can change the target state among the target state candidates and keep the changeable target state candidates. Thus, the target state is restricted from being changed to a passing state other than the target state candidate, thereby preventing the target state from being changed to a target state other than the target state originally intended and greatly deviating from the position.

[0066] The target state is not limited to the position, speed and azimuth of the mobile body 1, and it is sufficient to include at least the state quantity of the position. In addition, it can also include other state quantities of the mobile body 1 such as acceleration, jerk, altitude, yaw rate, rotation angular velocity, and steering angle.

[0067] The local target state generation unit 220 calculates one local target state reachable by any target state candidate at an intermediate point when moving to the periphery of at least one target state candidate obtained from the target state candidate generation unit 210 .

[0068] The motion control unit 300 includes a trajectory generation unit 310 , a control amount calculation unit 320 , and an actuator control unit 330 .

[0069] The trajectory generation unit 310 generates a trajectory, and outputs the generated trajectory to the control amount calculation unit 320 so that the mobile body 1 is controlled along the generated trajectory, wherein the trajectory includes at least a path and a speed that the mobile body 1 should travel until the local target state generated by the local target state generation unit 220 is reached. At this time, the trajectory generated by the trajectory generation unit 310 can also generate a path that does not include speed information and time information.

[0070] The control amount calculation unit 320 uses the trajectory generated by the trajectory generation unit 310 as a target trajectory, calculates a target control value for the moving body 1 for traveling along the target trajectory, and outputs the target control value to the actuator control unit 330 .

[0071] The actuator control unit 330 is a controller mounted on the moving body 1 and operates the actuator so that the moving body 1 follows the target control value calculated by the control amount calculation unit 320. Examples of the actuator include a steering wheel, a drive motor, and a brake.

[0072] Next, refer to Figure 2 ,use Figure 3 The flowchart shown in the figure explains an example of the operation of the passage point generating device 200. Next, the case of traveling through a toll gate will be explained.

[0073] First, before the processing in the location generating device 200, the pre-processing unit 10 determines whether the area where the mobile body 1 is currently traveling is an area near a toll booth based on the surrounding environment information acquired by the surrounding environment information acquiring unit 110 of the autonomous sensor information acquiring unit 100 (step S101).

[0074] exist Figure 2 The area near the toll station FS can be, for example, an area from a position at an arbitrary distance in front of the point where the demarcation line BL disappears to a drivable position (i.e., passing through the toll station gate GT), or an area without a demarcation line BL or an area within an arbitrary distance from the drivable position.

[0075] In determining whether the area where the mobile body 1 is currently traveling is an area near a toll station, the location information of the toll station in the map information such as the high-precision map and vehicle navigation map possessed by the mobile body 1 can be used, or image processing can be performed on the image obtained from the front camera mounted on the mobile body 1 to make the judgment.

[0076] When it is determined in step S101 that the mobile body 1 is traveling near a toll booth (in the case of "Yes"), the location generating device 200 obtains obstacle information and a set of drivable states around the mobile body from the surrounding environment information acquiring unit 110, and obtains the current state (own state) of the mobile body 1 from the own state acquiring unit 120 (step S102). On the other hand, when the mobile body 1 is not traveling at a toll booth (in the case of "No"), the track generating unit 310 generates a track on which the mobile body 1 should travel based on the information of the lanes of the road acquired by the surrounding environment information acquiring unit 110.

[0077] In step S102, after obtaining the own state, surrounding environment information, and a set of drivable states, in the target state candidate generation unit 210, multiple target state candidates (target state candidate set) are generated based on the information obtained from the surrounding environment information acquisition unit 110 and the own state acquisition unit 120 (step S103). Figure 4 A conceptual diagram showing this process.

[0078] <Processing in Target State Candidate Generating Unit>

[0079] Figure 4 The schematic diagram shows a state in which the moving body 1 intends to enter the driving lane in front of the toll station FS. Figure 4 In the case where i is set as the suffix, (x ci ,y ci ) is the target position candidate, v ci is the target velocity candidate, θ ci is the target azimuth candidate.

[0080] The set of target position candidates is expressed by the following formula (4).

[0081] [Formula 4]

[0082] P c ={(x c1 ,y c1 ), (x c2 ,y c2 ), (x c3 ,y c3 )…(x ce ,y ce )}…(4)

[0083] The set of target speed candidates is expressed by the following formula (5).

[0084] [Formula 5]

[0085] V c = {v c1 , v c2 , v c3 …vce}…(5)

[0086] The set of target azimuth angle candidates is expressed by the following formula (6).

[0087] [Formula 6]

[0088]

[0089] Here, the suffix e is the number at the end of the environment. Figure 4 In this example, the last digit is 3.

[0090] exist Figure 4 In the example of , when there are multiple target position candidates, there are also multiple target speed candidates and target azimuth angle candidates, but they can also be set to 1, and do not need to correspond to the number of target position candidates.

[0091] Hereinafter, target state candidates are defined based on respective target position candidates, target speed candidates, and target azimuth angle candidates, and a collection of target position candidate sets, target speed candidate sets, and target azimuth angle candidate sets is referred to as a target state candidate set.

[0092] exist Figure 4 In the example, the target state candidate is from Figure 2 The three drivable states selected from the drivable state set obtained by the surrounding environment information acquisition unit 110 are indicated by blackening the ☆ mark. Figure 2 The result of reducing the set of drivable states to 3 drivable states is shown as the target state candidate set.

[0093] In addition, if Figure 4 For example, for each target state candidate, the state quantities of the target state candidate set are blocks of values ​​close to each other in such a manner that adjacent drivable states are selected. The block is a set of state quantities such that adjacent target states are within a predetermined range set in advance.

[0094] Since the target state candidates are all located in close proximity, trajectory correction by changing the target state can be performed more smoothly.

[0095] In addition, if Figure 5 As shown, the target state candidates are all set to be state quantities that are within the reachable region R0 that is kinematically reachable by the mobile body 1 and that guarantee the possibility of reaching the target state.

[0096] That is, in Figure 5In the figure, a plurality of reachable boundaries R0B defining the reachable region R0 are indicated by circle marks. Reachable boundaries R0B located in front and behind each other are connected to form reachable boundary lines, and a region sandwiched by two reachable boundary lines becomes the reachable region R0.

[0097] Since the possibility of reaching any target state candidate is guaranteed, the possibility of the moving object 1 reaching any target state becomes high.

[0098] exist Figure 5 In the example, there are four target state candidates included in the reachable area R0, which constitute the target state candidate set. Figure 5 In the example of , all of the four target state candidates within the reachable area R0 are selected, but it is not necessary to select all of them, and an arbitrary number can be selected.

[0099] The reachable boundary R0B can be represented by, for example, a discrete predicted trajectory of a virtual moving body, wherein the discrete predicted trajectory of the virtual moving body is obtained in the following manner: an arbitrary acceleration, an arbitrary maximum steering angular velocity, i.e., a limiting value in the moving body 1 is repeatedly input into a simple moving body dynamics model such as the following formula (7) from an arbitrary initial state, such as an arbitrary speed, an arbitrary starting position, an arbitrary azimuth, etc. Figure 6 A conceptual diagram showing this expression is shown.

[0100] [Formula 7]

[0101]

[0102] In formula (7), t is a time variable, dt is a sampling time in the control cycle, umax is a maximum steering angular velocity input, and it is assumed that there are two modes of rotation directions in the left and right directions. α is an arbitrary acceleration input, δ is the steering angle, L is the wheelbase, and β is the sideslip angle of the moving body. Here, the maximum steering angular velocity input is defined by, for example, the maximum value of the rotation angle (deg / sec) of the steering wheel of the car per second, and is set to a value that takes safety into consideration.

[0103] exist Figure 6 In the example of FIG. 1 , it is assumed that the starting position is set to the starting position of the toll booth section where the dividing line BL disappears, and the virtual mobile body VMV moves with the initial state quantity (x t0 ,y t0 ,θ t0 , v t0 ) exists at the starting position STP. Furthermore, any input value is repeatedly provided to the formula (7) that specifies the action of the virtual moving body at each time, thereby obtaining the following Figure 6The discrete predicted trajectories of the virtual mobile body VMV at each time are shown. Here, the arbitrary input value means a value that is different enough to obtain a predicted trajectory that defines the reachable area R0.

[0104] In the first embodiment, the discrete predicted trajectory obtained by repeatedly inputting the upper limit value of the input value, that is, the predicted trajectory facing Figure 6 The tracks represented by the multiple virtual mobile bodies VMV on the left and the discrete predicted tracks obtained by repeatedly inputting the lower limit of the input value, i.e., the tracks facing Figure 6 The trajectory represented by the multiple virtual mobile bodies VMV on the right is used as the reachable boundary.

[0105] Here, the upper limit of the input value refers to, for example, the maximum steering angular velocity input umax or the maximum acceleration input. Depending on the situation, the maximum steering angular velocity input umax and the acceleration α are fixed to 0 and input, or the steering angular velocity u is fixed and the acceleration α is input as a constant acceleration. In addition, the lower limit of the input value refers to, for example, the minimum steering angular velocity input or the minimum acceleration input.

[0106] The upper limit and lower limit of the input value are not limited to values ​​related to the performance limit of the vehicle 1 , and may be values ​​that take riding comfort into consideration or values ​​arbitrarily determined by the user.

[0107] As an example of a value that takes riding experience into consideration, acceleration, jerk, rotation speed, and rotation acceleration can be set to be less than a predetermined value. In addition, at least one of them can be set to be less than a predetermined value. The predetermined value can be set to a fixed value set in advance or a value that can be adjusted by the user.

[0108] By setting the upper limit and the lower limit of the input value to values ​​that take the riding experience into consideration, the possibility that any target state can be reached with a good riding experience increases.

[0109] The dynamics model of the virtual moving body is not limited to equation (7), and other models such as a two-wheel model that approximates four wheels to two wheels and a dynamics model of each moving body as a target can be used.

[0110] In addition, the reachable boundary R0B is not limited to the trajectory predicted by the dynamic model, but can also be set using a spline curve, a clothoid curve, an n-degree polynomial curve, a straight line, etc. arbitrarily set by the user.

[0111] The reachable area R0 is a static area calculated based on a fixed starting position, but may be a dynamic area calculated repeatedly each time the moving body moves, with the current state of the moving body 1 moving all the time as the starting position.

[0112] <Processing in Modification Example of Target State Candidate Generating Unit>

[0113] Explained Figure 1 The target state candidate generating unit 210 shown in FIG. 1 generates a target state newly determined based on at least one drivable state obtained from the surrounding environment information acquiring unit 110 and the current state of the mobile body 1 obtained from the own state acquiring unit 120 as a target state candidate, but it can also be as follows Figure 7 The structure shown.

[0114] like Figure 7 As shown, the target state candidate generation unit 210 can be configured to include: a virtual track generation unit 211 that generates a virtual track for traveling or reaching each drivable state; a virtual track evaluation unit 212 that evaluates the virtual track based on the dynamic constraints of the mobile body 1; and a target state candidate calculation unit 213 that selects a drivable state that can be reached using the virtual track without breaking the constraints as a target state candidate. The following describes a method for generating target state candidates using this configuration.

[0115] Figure 8 2 is a conceptual diagram for explaining the process of generating a virtual track to a drivable state in the virtual track generating unit 211. Figure 8 In , when i is set as the suffix, trji represents a virtual track up to the i-th drivable state.

[0116] Here, the initial state IS is set to the starting position of the toll booth section where the dividing line BL disappears, and the state quantity (x 0 ,y 0 ,θ 0 , v 0 ).

[0117] Regarding the virtual track, for example, the polynomial expression of formula (8) can derive the coefficients by solving the simultaneous equations of the initial state and the boundary conditions in each drivable state represented by the following formulas (9) to (14).

[0118] [Formula 8]

[0119] trj i =f i (x) = ∑c j i x j …(8)

[0120] In formula (8), j is the degree of the polynomial, and c is the coefficient of each degree. In the first embodiment, the degree is 5.

[0121] [Formula 9]

[0122] f(x0 )=y 0 …(9)

[0123] [Formula 10]

[0124] f(x pi )=y pi …(10)

[0125] [Formula 11]

[0126] f′(x 0 )=y 0 '…(11)

[0127] [Formula 12]

[0128] f′(x pi )=y pi '…(12)

[0129] [Formula 13]

[0130] f″(x 0 )=y 0 ″…(13)

[0131] [Formula 14]

[0132] f″(x pi )=y pi ″…(14)

[0133] Here, f′(x) represents the first-order differential of f(x), f″(x) represents the second-order differential of f(x), formula (9) represents the boundary conditions related to the position in the initial state, formula (10) represents the boundary conditions related to the position in each drivable state, formula (11) represents the boundary conditions related to the inclination in the initial state, formula (12) represents the boundary conditions related to the inclination in each drivable state, formula (13) represents the boundary conditions related to the curvature in the initial state, and formula (14) represents the boundary conditions related to the curvature in each drivable state.

[0134] By placing the target state candidates on a virtual trajectory where the moving body 1 can reach the target state, the possibility of reaching any target state increases.

[0135] In this embodiment 1, the virtual track generated by the virtual track generation unit 211 is generated by a polynomial two-point boundary value problem, but can also be generated by other methods such as model-based methods such as the potential energy method, model predictive control, sampling methods such as RRT (Rapidly Exploring Random Tree), and graph search algorithms such as the A* search algorithm.

[0136] In the virtual track evaluation unit 212, the virtual track generated by the virtual track generation unit 211 is dynamically evaluated. For example, the virtual track is evaluated for the possibility of reaching, such as a track having a portion with a curvature greater than the minimum turning radius of the mobile body 1 is set as an unreachable track because it exceeds the dynamic constraint. In addition, whether the generated track crosses or hits a static obstacle such as an outer wall obtained by the surrounding environment information acquisition unit 110 can be added to the evaluation items.

[0137] In addition, when the dynamic model of the mobile body using formula (7) is used to predict the trajectory of the virtual mobile body, the trajectory of the virtual mobile body that cannot be tracked by the mobile body 1 and the trajectory of the virtual mobile body that exceeds the constraints related to the dynamics of the mobile body 1 such as the maximum steering angle velocity or steering angle can be evaluated as unreachable trajectories.

[0138] The target state candidate calculation unit 213 uses a travelable state that can be reached by the virtual track evaluated as reachable as a target state candidate. Fig. 9 A conceptual diagram showing this process.

[0139] exist Fig. 9 In the virtual track generation unit 211, the virtual track trj generated 1 ~trj 7 In the example, a track evaluated as unreachable by the virtual track evaluation unit 212 is marked with an ×, and a track evaluated as reachable is marked with a 0, and all target state candidates become state quantities that can be traveled on reachable virtual tracks.

[0140] exist Fig. 9 In the example, there are 4 target state candidates, which constitute the target state candidate set. Fig. 9 In the example of , all four target state candidates that can be reached by traveling on a virtual track are selected, but it is not necessary to select all of them, and an arbitrary number can be selected. The arbitrary number may include, for example, a maximum number set by the user, a minimum number, and the total number that can be reached by the virtual track.

[0141] Regarding the target state candidates, the virtual track evaluation unit 212 can score multiple virtual tracks that reach each target state candidate according to a predetermined evaluation index, and use the track travel location where the score assigned to each virtual track is greater than or less than a predetermined threshold as the target state candidate. Regarding the evaluation index, for example, the straightness, length, maximum value of curvature, magnitude of curvature change rate, number of steering wheel switching, etc. of the virtual track can be cited.

[0142] In the above evaluation index, if it is straight-line performance, the following score distribution can be imagined: if it is a track that moves roughly straight, it is 10 points, if it is a track with no straight-line part, it is 0 points, etc. Regarding length, curvature, etc., points are also distributed according to safety, and their total score is compared with a predetermined threshold value as a target state candidate.

[0143] By scoring a plurality of virtual trajectories leading to each target state candidate, it is possible to change the score distribution according to the preference of the designer and the user, thereby changing the target state candidate.

[0144] As an evaluation method in the virtual track evaluation unit 212, a method of geometrically evaluating a virtual track can be cited. For example, a method is also considered in which the maximum curvature Kmax of each track is calculated from the virtual track, and the score of each track is calculated based on the curvature value so that the score of each track becomes 1 / Kmax1, 1 / Kmax2, ... 1 / Kmaxi, and the evaluation is performed. Here, the suffix i corresponds to the suffix of each track.

[0145] In addition, if it is the length of the track, the following method is also considered: when the length of the track is set to L, the score of each track is calculated based on the length value so that it becomes 1 / L1, 1 / L2, 1 / L3, ... 1 / Li, and the evaluation is performed. In addition, the reason for setting it as 1 / maximum curvature and 1 / length in the above is that among these geometric elements, smaller values ​​are generally better from the perspective of safety.

[0146] Furthermore, the above-mentioned score based on the maximum curvature and the score based on the length may be accumulated and the total score may be compared with a predetermined threshold value, and if there are geometric elements that should be given more importance, weighting may be performed.

[0147] In addition, it is also possible to make the following approximate score distribution more simply: 7 points are allocated when the maximum curvature value of the track is in the range of aa to bb, 6 points are allocated when it is in the range of bb to cc, and 5 points are allocated when it is in the range of cc to dd. <bb<cc<dd。

[0148] Here, return to Figure 3 After calculating the target state candidate in step S103, the local target state generation unit 220 calculates a local target state that can reach any target state candidate at an intermediate point when moving to the periphery of the target state candidate set (step S104). This is a process based on the concept that any target state candidate can be reached as long as it starts from the intermediate point. Fig.10 A conceptual diagram showing this process.

[0149] <Processing in Local Target State Generation Unit>

[0150] Fig.10 The diagram schematically shows a state in which the mobile body 1 attempts to enter a travel lane in front of the toll booth FS, and shows a state in which a local target state LTS is calculated at an intermediate point when the mobile body 1 moves to the periphery of the target state candidate set.

[0151] exist Fig.10 In the case where i is set as the suffix, S ci Each target state candidate is represented and defined by the following formula (15), but the state variable can also be a variable other than the variable in formula (15).

[0152] [Formula 15]

[0153]

[0154] In addition, Fig.10 In (x l ,y l ,θ l , v l ) are the elements of the local target state LTS, and the local target state S that summarizes the elements l It is defined by the following formula (16).

[0155] [Formula 16]

[0156]

[0157] exist Fig.10 In the example of , there are 3 target state candidates, which constitute the target state candidate set. By starting from the local target state LTS, any target state candidate can be reached.

[0158] exist Fig.10 In the example, the local target state LTS is set to a position that can reach any target state candidate, but it can also be Fig.11 As shown in FIG. 1 , the local target state LTS is set to an arbitrary position on a virtual trajectory that reaches the target state candidate located at the center of the target state candidate set.

[0159] exist Fig.11 In the process, the starting position of the toll booth section where the dividing line BL disappears is set as the initial state IS, and the state quantity (x 0 ,y 0 ,θ 0 , v 0 ).exist Fig.11 In, trj 2 represents the virtual trajectory up to the second target state candidate, l s is the distance that can be detected by autonomous sensors. Fig.11 As shown, let any position on the virtual track be the distance l before the target state candidate set. s The position of the target state candidate set is set as the line segment parallel to the y direction and including the position, and the position and the virtual track trj are 2 The intersection position is set as the local target state LTS.

[0160] In addition, any position on the virtual track can be set to be a distance l from the previous position. s A position closer to the position of Fig.12 The distance l from the target state candidate set is shown s Subtract any distance l a location.

[0161] exist Fig.11 The reason for setting the local target state LTS in such a way that the target state candidate located in the center of the target state candidate set is as the destination is that when the target state candidate in the center cannot be driven, when the track is corrected to other target state candidates, there are target state candidates on both adjacent sides, and the number of target state candidates that can correct the track increases.

[0162] In addition, by making the local target state the distance l that the autonomous sensor can detect in front of the target state candidate set s The position of the vehicle can obtain dynamic information in real time, such as dynamic obstacles and the color of the signal of the ETC gate. When such dynamic information can be detected, the possibility of changing the selected target state becomes higher. Therefore, for example, if a vehicle approaches the target gate, a faulty vehicle is detected and the gate is changed, and the situation where the target state has to be changed increases.

[0163] On the other hand, Fig.11 In the case of s Become and Fig.10 In the example of calculating the local target state LTS at an intermediate point when moving to the vicinity of the target state candidate set, a shorter distance is used than in the case where the local target state LTS is calculated, so the local target state LTS can be made closer to the position of the target state candidate set. After reaching the local target state LTS, the final target state can be avoided from becoming unable to drive.

[0164] By setting the local target state LTS in such a way that the central target state candidate in the target state candidate set is used as the destination, when the track is corrected to other target state candidates if the central target state candidate cannot be driven, target state candidates exist on both adjacent sides, making it easy to correct the track.

[0165] exist Fig.10 as well as Fig.11 In the example, the local target state LTS is 1 point, but it can also be Fig.13 As shown, the local target state is the region.

[0166] Fig.13 The diagram schematically shows a state in which the moving object 1 intends to enter the driving lane in front of the toll station FS. s is the distance that can be detected by the autonomous sensor. Fig.13 In the example, when i is set as the tail, A ci is the candidate state S that can be reached ci The reachable area, A s is the sensor detectable area that can detect the candidate set of target states using autonomous sensors, A l is the local target area. Local target area A l Defined as reachable area A ci and the sensor detectable area A of the target state candidate s Overlapping area.

[0167] Constitutes the reachable area A ci The reachable boundary of can be expressed, for example, by using the discrete predicted trajectory of the virtual moving body obtained as follows: assuming that the virtual moving body moves in the opposite direction to the direction through which the moving body 1 can pass when the initial position is used as each target state candidate, and repeatedly inputting arbitrary acceleration, arbitrary maximum steering angle velocity, that is, the limit value of the moving body 1 into a simple dynamic model of the moving body such as formula (7), thereby obtaining the discrete predicted trajectory of the virtual moving body. This is the regulation Fig.13 The left and right boundaries of each reachable area.

[0168] In each reachable area, the target state candidate can be reached no matter where the vehicle starts. Furthermore, as long as the vehicle is in the local target area A l Within, any target state candidate in the target state candidate set can be reached no matter where it is started.

[0169] The dynamics model of the virtual moving body is not limited to equation (7), and other models such as a two-wheel model that approximates four wheels to two wheels and a dynamics model of each moving body as a target can be used.

[0170] In addition, the reachable boundary is not limited to the trajectory predicted by the dynamic model, and can also be set using a spline curve, a clothoid curve, an n-degree polynomial curve, a straight line, etc. arbitrarily set by the user.

[0171] The upper limit and lower limit of the input value input to the dynamics model to obtain the reachable boundary are not limited to values ​​related to the performance limit of the vehicle 1 , but may be values ​​considering the riding experience or values ​​arbitrarily determined by the user.

[0172] exist Fig.10 as well as Fig.11 In the local target state LTS is 1 point, Fig.13 The target area is local, but it can also be Fig.14 As shown, the local target state is represented by the probability distribution of each target state candidate that can be reached.

[0173] Fig.14 The state in which the moving body 1 intends to enter the driving lane in front of the toll booth FS is schematically shown, and the probability distribution f of the local target state is p (s i ), the depth of the hatching represents the height of the probability, the densest area has the highest probability of reaching the target state candidate, and by starting from this area, the possibility of reaching any target state candidate becomes the highest.

[0174] <Weighting of target state candidates>

[0175] The local target state generation unit 220 can weight each target state candidate calculated by the target state candidate generation unit 210 according to a predetermined evaluation index, and calculate the local target state according to the weight. Fig.15 A conceptual diagram showing this process.

[0176] Fig.15 The state in which the moving body 1 intends to enter the driving lane in front of the toll booth FS is schematically shown. When i is set as the tail mark, w i is the weight given to the target state candidate. Regarding the weight, w 1 The heaviest, w 3 The lightest, becoming w 1 >w 2 >w 3 .

[0177] For the target state candidate S c1 The weight assigned is the heaviest, so the local target state LTS is close to the target state candidate S c1 The state quantity, that is, the state candidate S c1 The position of the target state candidate S is reached when starting from the local target state LTS. c1 The possibility of becoming higher.

[0178] Thus, by changing the weight according to the preference of the designer and the user, a local target state can be set.

[0179] Here, regarding the predetermined evaluation indicators for each target state candidate that become weighted indicators, for example, the weight is increased for candidates whose distance to the current moving body 1 is closer, and the weight is increased for candidates whose lateral width of the gate to be passed is wider, and the weight is reduced for candidates whose path through the gate is a curve, etc., and evaluation can be performed according to geometric conditions.

[0180] In addition, evaluation can be performed based on the congestion of the gate through which the target state candidate passes, whether the gate through which the target state candidate passes is an ETC gate or a general gate that requires a temporary stop, and on which side of the toll station the gate through which the target state candidate passes is located.

[0181] Furthermore, the difficulty of passing through the gate can be evaluated based on the complexity of the road shape in front of the gate that the target state candidate passes through or the passing speed required to pass through the gate.

[0182] <Weighting of Virtual Tracks>

[0183] The local target state generation unit 220 can also generate Figure 7 The virtual trajectory generation unit 211 in the target state candidate generation unit 210 shown in the figure weights the trajectory until reaching each target state candidate, and uses the weight of the virtual trajectory to set the local target state. Fig.16 A conceptual diagram showing this process.

[0184] Fig.16 The schematic diagram shows a virtual track starting from the initial state IS, where the starting position of the toll booth section where the dividing line BL disappears is set as the initial state IS. s is the distance that can be detected by the autonomous sensor. Fig.16 In (x l ,y l ,θ l , v l ) are the elements of the local target state LTS. When i is set as the tail, trj i represents the virtual trajectory up to the i-th target state candidate, w i It is a virtual track trj i The weight assigned, (x li ,y li ) is the position where the virtual track and the detectable location of the target state candidate set intersect. It is called an arbitrary point on the virtual track.

[0185] About TRJ on Virtual Track i The weight assigned, w 1 The heaviest, w 3 The lightest, becoming w 1 >w 2>w 3 . For virtual track trj 1 The weight assigned is the heaviest, so the local target state LTS is close to the target state candidate S c1 The state quantity, that is, the state candidate S c1 When starting from the local target state LTS, the target state candidate S is reached. c1 The possibility of becoming higher.

[0186] In this way, by changing the weight according to the preference of the designer and the user, a local target state can be set.

[0187] Regarding the local target state LTS, for example, by taking the weighted average of any point and weight on the virtual track using formula (17) and formula (18), the weight of each virtual track can be reflected. Fig.16 In the case of the example, the suffix b at the end becomes 3.

[0188] The x-direction position x of the local target state LTS can be obtained using the following formula (17): l .

[0189] [Formula 17]

[0190]

[0191] The y-direction position y of the local target state LTS can be obtained using the following formula (18): l .

[0192] [Formula 18]

[0193]

[0194] For example, the weight is set based on the distance from the moving body 1, the amount of lateral movement required to move to each target state candidate, the congestion of the gate through which each target state candidate passes, the selection rate of the target state of other vehicles, the distance and position relationship with the target state in a wider area (described in Embodiment 2), etc. In addition, as shown in Formula (17) and Formula (18), the weight for the x direction and the weight for the y direction are set to the same w i , but it can also be different.

[0195] Here, return to Figure 3 After the local target state is calculated in step S104, the post-processing unit 20 determines whether the moving body is close to the local target state (step S105). For example, if the Euclidean distance between the moving body 1 and the local target state is less than a threshold value, or if the moving body 1 reaches the local target state, the moving body 1 will be judged as being close to the local target state (step S105). Fig.13Local target area A l If the moving object 1 is within the local target state, it is determined that the moving object 1 is approaching the local target state.

[0196] If it is determined in step S105 that the moving body 1 is not close to the local target state ("No"), the process proceeds to step S106, where the local target state is used as the target value, and the trajectory generation unit 310 ( Figure 1 ) generates the target trajectory of the moving body 1. Fig.17 A conceptual diagram showing this process.

[0197] Fig.17 The state in which the moving body 1 intends to enter the driving lane in front of the toll booth FS is schematically shown, and the target trajectory trj of the moving body 1 from the position of the moving body 1 to the local target state LTS generated by the trajectory generation unit 310 is shown. t .

[0198] On the other hand, when it is determined in step S105 that the moving object is approaching the local target state (in the case of "Yes"), the process proceeds to step S106, and the trajectory generation unit 310 generates a trajectory up to one target state among the target state candidates as a target trajectory. Fig.18 A conceptual diagram showing this process.

[0199] exist Fig.18 , it is shown that after the moving body 1 approaches the initial target state candidate, one target state is selected from three target state candidates, and the target trajectory trj to the selected target state is generated by the trajectory generation unit 310. t status.

[0200] In this case, as for the method of selecting one target state, for example, there can be mentioned a method of weighting the target state candidates according to a predetermined index and selecting a target state candidate with a large weight.

[0201] Here, as predetermined indicators, for example, there can be cited the distance to the mobile body 1, the amount of lateral movement required to move to each target state candidate, the congestion of the gate through which each target state candidate passes, the selection rate of the target state of other vehicles, the distance to a wider area target state (described in Implementation 2) and the positional relationship, etc.

[0202] In addition, if Fig.18 As shown, when there is a broken-down vehicle BV near one target state candidate, or when there is a target state candidate that cannot pass because the signal of the toll gate GT is red, a method of removing the corresponding target state candidate can be given.

[0203] In this way, depending on whether the judgment result of step S105 is "yes" or "no", the trajectory generated in the next step S106 is different. That is, when the judgment of step S105 is "yes", the trajectory generation unit 310 generates a trajectory to the target state, and when the judgment of step S105 is "no", the trajectory generation unit 310 generates a trajectory to the local target state. After generating an arbitrary trajectory in step S106, a series of processes are terminated. In addition, based on the generated trajectory, in the control amount calculation unit 320 of the motion control unit 300, the target control value for the moving body 1 for traveling along the target trajectory is calculated, and the actuator control unit 330 causes the actuator to operate in such a way that the moving body 1 tracks the target control value, but since these actions can use well-known technologies, the description is omitted.

[0204] Here, when the trajectory generating unit 310 generates a trajectory to a local target state, after the mobile body 1 approaches the local target state to a certain extent, if the local target state is suddenly changed, the mobile body 1 needs to make a sudden trajectory correction, and sometimes the trajectory that can reach the changed local target state cannot be found, so such a sudden change of the local target state can be limited. Specifically, the change of the target state amount can be limited to the amount that the change of the state amount per unit time becomes more than a predetermined amount.

[0205] By limiting the rapid change of the local target state, it is possible to prevent the moving body 1 from making a rapid trajectory correction and from being unable to find a trajectory that leads to the local target state after the change.

[0206] Fig.19 as well as Fig. 20 An example of such a sudden change in the local target state is shown. Fig.19 , the target trajectory trj of the moving body 1 from the current position of the moving body 1 to the initial local target state LTS is shown. t , and when it changes suddenly to the local target state LTSX, the initial target trajectory trj t It is impossible to reach the changed local target state LTSX. In this case, a new target track trj is generated. t , it takes time.

[0207] exist Fig. 20 In the figure, the situation where the initial local target state LTS exists near the toll station FS is shown. When it changes suddenly to the local target state LTSX, sometimes due to the presence of a faulty vehicle BV in front of the toll station gate GT, a track to reach the changed local target state LTSX cannot be found or the calculated time is not available.

[0208] Furthermore, it is also possible to Figure 3The processing period of steps S101 to S104 is restricted. Fig.21 Describe the restriction.

[0209] exist Fig.21 In the process of moving body 1 traveling in area RA, the starting position of the toll booth section where the dividing line BL disappears is set as the starting position STP, the front side of the starting position STP is set as area RA, and the side of the toll booth gate GT is set as area RB. Figure 3 In the processing of steps S101 to S104, after the moving body 1 passes through the starting position STP, the processing of steps S101 to S104 is not performed, and the values ​​of the target state candidate set and the local target state LTS calculated during the traveling in the area RA are used.

[0210] This prevents the local target state that is the target from always changing or the target value of the moving body 1 from vibrating after the moving body 1 passes the starting position STP, thereby preventing fluctuations in the generated trajectory, that is, fluctuations in the target value when generating the trajectory.

[0211] In addition, the starting position STP is not limited to the point where the dividing line BL disappears, but can also be set to a point a predetermined distance before the point where the dividing line BL disappears.

[0212] As described above, according to the passing point generating device 200 of embodiment 1, it is considered to correct the target passing point at the initial stage of determining the target passing point to determine multiple candidates for the target passing point. In addition, the local target state LTS that can reach any target passing point is set to the intermediate point, and the vehicle travels to the intermediate point as the destination. Therefore, when the target passing point changes before and during driving near the intermediate point, the vehicle can also respond to the change of the target passing point flexibly, and the passing probability of the target passing point can be increased. In addition, in the case where the target passing point is impassable, the action of changing the target passing point to another passing point can be implemented without performing a sharp track correction process.

[0213] <Implementation Method 2>

[0214] Fig. 22 is a block diagram showing the structure of a mobile body 1 equipped with a passing point generating device according to Embodiment 2 of the present disclosure. Fig. 22 About and use Figure 1 The same components as those of the moving body 1 described above are denoted by the same reference numerals, and duplicate descriptions thereof will be omitted.

[0215] Fig. 22 The illustrated moving object 1 includes a surrounding environment information acquisition unit 110 , a self-state acquisition unit 120 , and a wide-area environment information acquisition unit 130 in an autonomous sensor information acquisition unit 100 .

[0216] The wide-area environment information acquisition unit 130 acquires wide-area information of a wider range than the surrounding environment information acquisition unit 110. This information includes, for example, route information of lanes, a wider-area destination, a passing position, and a relay point, and is acquired from, for example, information pre-specified by a user, predetermined route information and position information of map information possessed by the mobile body 1, a high-precision locator, and an on-vehicle communication device.

[0217] Furthermore, the mobile body 1 includes a wide-area target state generation unit 230 , a target state candidate generation unit 210 , and a local target state generation unit 220 in the passage point generation device 200 .

[0218] The wide-area target state generation unit 230 generates a wide-area target state, i.e., a target state of a wider area than the drivable state obtained by the surrounding environment information acquisition unit 110 and the target state candidate set generated by the target state candidate generation unit 210 , based on information such as the wide-area destination obtained from the wide-area environment information acquisition unit 130 . Fig.23 A conceptual diagram showing this process.

[0219] Fig.23 The moving body 1 passes through the starting position STP and is approaching the toll station FS. A wide target state GTS exists on the other side of the toll station FS.

[0220] exist Fig.23 In (x g ,y g ,θ g , v g ) are the elements of the wide area target state GTS, (x g ,y g ) is the wide-area target position, v g is the wide-area target velocity, θ g Here, the state of the wide-area target only needs to include information on at least the position, and may not include information on the speed and the azimuth. Fig.23 The target route TR in is an example of information acquired by the wide-area environmental information acquisition unit 130. Fig.23 The wide-area target state GTS in becomes the initial position of the mobile body 1 required for traveling along the target route TR.

[0221] exist Fig.23 In the example of , the target state candidates are three drivable states selected from the drivable state set, and the ☆ mark is painted black as the target state candidate set.

[0222] The target state candidate generating unit 210 sets the starting position STP as the initial position, and generates a target state candidate from a set of drivable states in an overlapping area between a reachable area kinematically reachable by the moving body 1 from the initial position and a reachable area kinematically reachable by the moving body 1 to a wide-area target state GTS. Fig.24 A conceptual diagram showing this process.

[0223] exist Fig.24 In the figure, the reachable area R1 is the area that the moving body can reach kinematically from the starting position STP, and multiple reachable boundaries R1B are represented by the mark 0. The reachable boundaries R1B located in front and behind each other are connected to form reachable boundary lines, and the area sandwiched by two reachable boundary lines becomes the reachable area R1.

[0224] The reachable region R2 is a region where the mobile body can reach the wide-area target state GTS kinematically. A plurality of reachable boundaries R2B are indicated by zeros. The reachable boundaries R2B located in front and behind each other are connected to form a reachable boundary line, and the region sandwiched by the two reachable boundary lines becomes the reachable region R2. The target state candidate is a state quantity in the region where the reachable region R1 and the reachable region R2 overlap.

[0225] By deriving the reachable boundaries R1B and R2B from the discrete predicted trajectory of the virtual moving body obtained using the simple dynamic model of the moving body such as the formula (7) described in the first embodiment, the reachable regions R1 and R2 can be obtained.

[0226] <Processing in Modification Example of Target State Candidate Generating Unit>

[0227] In addition, it can be configured to replace Fig. 22 The target state candidate generation unit 210 shown in FIG. Figure 7 As shown in the target state candidate generating unit 210, it comprises: a virtual track generating unit 211, which generates a virtual track; a virtual track evaluating unit 212, which evaluates the virtual track based on the dynamic constraints of the mobile body 1; and a target state candidate calculating unit 213, which selects a drivable state that can be reached by using the virtual track without breaking the constraints as a target state candidate.

[0228] Fig.25 2 is a conceptual diagram for explaining the process of generating a virtual track to a drivable state in the virtual track generating unit 211. Fig.25 In the case where i is set as the tail, trj bi is the virtual track up to the ith drivable state, trj ai It is a virtual trajectory from the i-th drivable state to the wide-area target state GTS.

[0229] First, the virtual trajectory generation unit 211 generates a virtual trajectory from the starting position STP to the drivable state and a virtual trajectory from the drivable state to the wide-area target state GTS, respectively. Next, the virtual trajectory is dynamically evaluated in the virtual trajectory evaluation unit 212. Finally, a passing state that can be passed by a combination of two reachable trajectories (both the virtual trajectory from the starting position STP to the drivable state and the virtual trajectory from the drivable state to the wide-area target state are evaluated as reachable) is set as a target state candidate.

[0230] In addition, the virtual orbit can be expressed using a polynomial such as the formula (8) described in the first embodiment by solving the simultaneous equations of the boundary conditions represented by the formulas (9) to (14) and deriving the coefficients.

[0231] In addition, regarding the target state candidates, the virtual track evaluation unit 212 scores the multiple virtual tracks from the starting position STP to each target state candidate and the multiple virtual tracks from each target state candidate to the wide-area target state GTS according to a predetermined evaluation index. Moreover, regarding the virtual tracks to each target state candidate and the virtual tracks from each target state candidate to the wide-area target state GTS, the target state that can be driven using a combination of scores of both being above a threshold is set as the target state candidate. In addition, instead of scoring the virtual tracks, the weights shown in Implementation Example 1 may be assigned. In this case, weights are assigned to the multiple virtual tracks from the starting position STP to each target state candidate and the multiple virtual tracks from each target state candidate to the wide-area target state GTS.

[0232] As described above, according to the passing point generating device 200 of the second embodiment, the probability of traveling to the target state can be increased, and an action such as changing the target state when the target state cannot be traveled can be performed without abrupt track correction. Furthermore, after reaching the target state, when the wide-area target point ahead is taken as the destination, the target state takes the wide-area target state into consideration, so it is easy to reach the wide-area target point.

[0233] <Implementation Method 3>

[0234] Fig.26 1 is a block diagram showing the configuration of a control server 2 (control device) equipped with a passing point generation device and a mobile body 1 according to Embodiment 3 of the present disclosure. Fig.26 About and use Figure 1 The same components as those of the moving body 1 described above are denoted by the same reference numerals, and duplicate descriptions thereof will be omitted.

[0235] Fig.26The control server 2 shown includes a travel section information acquisition unit 400 and a passing point generation device 200 .

[0236] The travel section information acquisition unit 400 includes a section information acquisition unit 410. The section information acquisition unit 410 acquires information on the travelable space of the toll booth section, such as the position information of the outer wall that specifies the toll booth section, the objective state quantity of the mobile body 1, that is, the state quantity in the absolute coordinate system rather than the vehicle coordinate system, the objective state quantity of the moving obstacle, at least one travelable position of the mobile body 1 as a travel target, the travel speed, the travel azimuth, etc. Information on these is collected from the roadside sensor installed in the travel section.

[0237] The passing point generation device 200 includes a target state candidate generation unit 210 and a local target state generation unit 220 .

[0238] The target state candidate generation unit 210 generates a newly determined target state as a target state candidate for changing the target state based on the objective state quantity of the mobile body 1 obtained by the segment information acquisition unit 410, at least one drivable position of the mobile body 1 as a driving target, the driving speed, the driving azimuth, and other information.

[0239] The local target state generation unit 220 calculates one local target state reachable by any target state candidate at an intermediate point when moving to the periphery of at least one target state candidate obtained from the target state candidate generation unit 210 .

[0240] The state quantity of the local target state calculated by the local target state generating unit 220 is sent to the moving body 1 , and the trajectory generating unit 310 of the motion control unit 300 generates a trajectory consisting of the path and speed that the moving body 1 should travel until reaching the local target state.

[0241] Thus, in the third embodiment, the passing point generating device 200 is mounted on the control server 2 , and the local target state is transmitted from the control server 2 to the mobile body 1 through communication between the control server 2 and the mobile body 1 . Fig. 27 A conceptual diagram showing the system.

[0242] Fig. 27 The diagram schematically shows a state in which a moving body 1 entering a driving lane in front of a toll booth FS communicates with a control server 2 that controls the moving body 1. In addition, a plurality of roadside sensors RS are provided in the toll booth section, and each roadside sensor RS and the moving body 1 communicate with each other via the control server 2.

[0243] The travel section information acquisition unit 400 of the control server 2 acquires information such as objective state quantities of the moving body 1 and objective state quantities of the moving obstacle MOB in the toll booth section via a plurality of roadside sensors RS.

[0244] In this way, the control server 2 and the mobile body 1 communicate with each other, and the state quantity of the local target state calculated by the local target state generation unit 220 of the passing point generation device 200 in the control server 2 is provided to the mobile body 1 through the communication. Therefore, the local target state generation unit 220 generates the target state candidate and the local target state in consideration of the communication delay associated with the transmission from the control server 2 to the mobile body 1.

[0245] For example, regarding the use of Figure 5 Reachable area, use of Figure 8 The dynamics arrival possibility of the mobile body 1 required for the target state candidate such as the virtual trajectory described above is input into the dynamics model such as formula (7), not only the dynamics of the mobile body 1 but also the communication delay, thereby performing calculations taking the communication delay into consideration.

[0246] In addition, in the first embodiment, Fig.10 as well as Fig.13 The calculation of the local target state is related to the dynamical possibility of the mobile body, and the local target state is generated by considering not only the dynamics of the mobile body but also the communication delay.

[0247] For example, a method of setting the position of the local target state slightly closer to the front in consideration of a tracking delay caused by a communication delay until the target state candidate is reached may be mentioned.

[0248] As described above, the passage point generating device 200 of the third embodiment is installed in the control server 2. Therefore, compared with the case where the passage point generating device 200 is installed in the mobile body 1, the calculation processing is not concentrated in the mobile body 1, and the calculation load in the mobile body 1 is reduced.

[0249] Furthermore, the control server 2 can monitor a wider area than the moving body 1. Therefore, the passage point calculated by the passage point generating device 200 using the information recognized by the control server 2 is more likely to be a more suitable position than the passage point calculated by the moving body 1.

[0250] The control server 2 can instruct a plurality of moving objects about the target state, and the plurality of moving objects can automatically pass through the gate while coordinating with each other.

[0251] Furthermore, even in the case of regulation control in which communication delay occurs, any target state candidate can be reached, and the probability of passing through the toll gate GT is improved.

[0252] In addition, each component of the passing point generating device 200 according to the above-described first to third embodiments can be configured using a computer and implemented by executing a program on the computer. Fig.28 The processing circuit 50 shown implements the passing point generating device 200. In the processing circuit 50, a processor such as a CPU or a DSP (Digital Signal Processor) is applied to implement the functions of each part by executing a program stored in a storage device.

[0253] In addition, dedicated hardware can also be applied to the processing circuit 50. When the processing circuit 50 is dedicated hardware, the processing circuit 50 may be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a combination thereof.

[0254] Regarding the pass point generating device 200 , the functions of the components may be realized by independent processing circuits, or these functions may be realized by integrating them into one processing circuit.

[0255] In addition, Fig.29 , a hardware configuration is shown when the processing circuit 50 is constituted by a processor. In this case, the functions of each part of the location generating device 200 are realized by a combination of software or the like (software, firmware, or software and firmware). The software or the like is described as a program and stored in the memory 52. ​​The processor 51, which functions as the processing circuit 50, realizes the functions of each part by reading and executing the program stored in the memory 52 (storage device). That is, it can be said that the program causes the computer to execute the process and method of the operation of the constituent elements of the location generating device 200.

[0256] Here, the memory 52 can be, for example, RAM, ROM, flash memory, EPROM (Erasable Programmable Read Only Memory), EEPROM (electrically erasable programmable read only memory) and other non-volatile or volatile semiconductor memories, HDD (Hard Disk Drive), magnetic disk, floppy disk, optical disk, compact disk, mini disk, DVD (Digital Versatile Disc) and its drive device, etc., or all storage media used in the future.

[0257] In the above, a configuration is described in which the functions of the components of the location generation device 200 are implemented by either hardware or software. However, the present invention is not limited thereto, and a configuration may be used in which a part of the components of the location generation device 200 is implemented by dedicated hardware, and another part of the components is implemented by software. For example, the functions of a part of the components may be implemented by the processing circuit 50 as dedicated hardware, and the functions of another part of the components may be implemented by the processing circuit 50 as the processor 51 reading and executing a program stored in the memory 52.

[0258] As described above, the pass point generating device 200 can realize the above-mentioned functions by hardware, software, etc., or a combination thereof.

[0259] Although the present disclosure has been described in detail, the above description is by way of example only in all aspects and the present disclosure is not limited thereto. It should be understood that numerous modifications not shown in the examples can be envisioned without departing from the scope of the present disclosure.

[0260] In addition, the present disclosure can freely combine the various embodiments or appropriately modify or omit the various embodiments within the scope of the disclosure.

Claims

1. A passing point generating device, which generates a passing point that a moving object should reach, comprising: a target state candidate generating unit that calculates a plurality of target state candidates including at least a state quantity of a position based on the surrounding environment information of the mobile body and the state quantity of the mobile body; and The local target state generation unit calculates a local target state that is a local target state that can reach any target state candidate at an intermediate point around the plurality of target state candidates, and outputs the local target state as the passing point.

2. The passing point generating device according to claim 1, wherein: The target state candidate generating unit generates a set of state quantities close to each other so that adjacent target states are within a predetermined range so that the mobile body can reach the arbitrary target state candidate when reaching the local target state.

3. The passing point generating device according to claim 1 or 2, wherein: The target state candidate generating unit performs calculations so that any of the target state candidates are within a reachable area that is kinematically reachable by the moving body.

4. The passing point generating device according to any one of claims 1 to 3, wherein: The target state candidate generating unit performs calculations so that the arbitrary target state candidate can be reached by movement taking into account the riding feeling of the moving body.

5. The passing point generating device according to any one of claims 1 to 4, wherein: further comprising a wide-area target state generating unit that generates a target state of a wider area than each target state of the plurality of target state candidates and outputs the target state as the wide-area target state, The target state candidate generation unit performs calculations so that any of the target state candidates can reach the wide-area target state.

6. The passing point generating device according to claim 5, wherein: The plurality of target state candidates are target states for the mobile body to pass through a toll gate, The wide-area target state includes at least a state quantity of a position of the moving body as a destination after the moving body passes through the toll gate.

7. The passing point generating device according to any one of claims 1 to 6, wherein: The target state candidate generating unit comprises: A virtual track generation unit generates a plurality of virtual tracks passing through a plurality of target states; a virtual trajectory evaluation unit that evaluates the plurality of virtual trajectories based on dynamic constraints of the moving body; as well as The target state candidate calculation unit calculates the plurality of target states that can be reached using a virtual track that does not violate the constraint among the plurality of virtual tracks as the plurality of target state candidates.

8. The passing point generating device according to any one of claims 1 to 7, wherein: The local target state generation unit sets, as the local target state, an arbitrary position on a virtual track that reaches a target state candidate located most centrally among the target state candidates calculated by the target state candidate calculation unit, among the plurality of virtual tracks.

9. The passing point generating device according to any one of claims 1 to 8, wherein: The local target state generation unit restricts a change in the local target state immediately before the moving object reaches the local target state.

10. The passing point generating device according to any one of claims 1 to 9, wherein: The local target state generation unit weights the plurality of target state candidates according to a predetermined evaluation index, and calculates the local target state according to the weight value.

11. The passing point generating device according to claim 7 or 8, wherein: The target state candidate generating unit scores the plurality of virtual tracks according to a predetermined evaluation index, compares the scores respectively assigned to the virtual tracks with a threshold value, and sets locations traveled on the virtual tracks that are above the threshold value or below the threshold value as the plurality of target state candidates.

12. The passing point generating device according to claim 7 or 8, wherein: The local target state generation unit weights virtual tracks among the multiple virtual tracks that reach the multiple target state candidates calculated by the target state candidate calculation unit, and sets a state obtained by weighted averaging arbitrary positions on the virtual tracks and weights as the local target state.

13. The passing point generating device according to any one of claims 1 to 12, wherein: The passing point generating device is mounted on a control device that controls the moving body through communication with the moving body. The local target state generation unit calculates the local target state in a manner that the arbitrary target state candidate can be reached by taking into account a communication delay of a regulation control.

Images