Map division system, map update system, and map division method
The system optimizes autonomous driving by dividing maps to overlap partially and manage reference maps efficiently, addressing inefficiencies and stability issues in conventional systems, particularly in indoor environments.
Patent Information
- Application Number
- JP2023079181
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-12
- Publication Date
- 2024-11-22
AI Technical Summary
Conventional autonomous driving systems face inefficiencies due to the need for vehicles to stop at map boundaries for updates, leading to reduced driving efficiency and stability, especially in environments with limited GNSS coverage like indoor warehouses, and inadequate map division methods that do not account for sufficient environmental features for accurate self-position estimation.
A system that divides maps to partially overlap based on vehicle and environmental information, allowing for seamless transitions and reduced computational load by managing reference maps efficiently, using global and local route planning to ensure stable navigation.
This approach reduces calculation load on onboard controllers and ensures stable vehicle operation by minimizing unnecessary stops and decelerations, maintaining efficient navigation in diverse environments.
Smart Images

Figure 2024163497000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a map division system, a map update system, and a map division method. [Background technology]
[0002] Due to the labor shortage caused by the declining birthrate and aging population, and the expansion of the e-commerce market, there is a demand for labor saving and improvement of work efficiency in logistics warehouses and factories. Therefore, the introduction of autonomous driving systems that operate autonomous vehicles that can operate without a human driver is being promoted. In conventional autonomous driving systems, the range in which a vehicle can move is limited from the viewpoint of safety, and it was common that the working range of on-site workers and the moving range of the vehicle were different. On the other hand, with the development of autonomous driving technology in the robotics field, even if the movement lines of a vehicle and a person overlap, the vehicle itself can determine the need for avoidance or detours, and can select flexible actions according to the surrounding situation.
[0003] In order to realize wide-area autonomous driving, autonomous vehicles are required to drive in environments that change from moment to moment. For example, in warehouses and factories involved in transportation work, depending on the destination, vehicles are required to drive between shelves, in complicated passages, etc. In general, autonomous vehicles are equipped with on-board sensors that compare current measurement information with map information created in advance and detect the position of the vehicle in the created map coordinate system and obstacles that obstruct the vehicle's driving.
[0004] In addition, the vehicle-mounted controller, which performs various calculations related to autonomous driving, is composed of an information processing device such as a computer, and needs to hold reference map information, which is image information. Incidentally, in order to reduce the cost per vehicle, a control server is provided to manage a large number of vehicles. Compared to the control server, the vehicle-mounted controller has limited computing performance due to a central processing unit (CPU) and memory. If the vehicle side holds an entire map of the driving area (especially an entire map of a wide driving area) or an entire high-resolution map due to this computing performance restriction, there is a high possibility that problems such as memory pressure and increased computation load such as route planning processing to a target position will occur. Therefore, a technology has been proposed in which the entire map of the driving environment is divided into multiple areas, and the map referenced by the vehicle is updated sequentially according to the progress of the vehicle, such as the technology described in Patent Document 1.
[0005] Patent document 1 states that "the map update system divides a three-dimensional map that defines the robot's movable range in space into multiple areas representing specified ranges, generates a path for the robot on the three-dimensional map, and updates reference areas that need to be referenced on the robot's path in accordance with the progress of the robot's path." [Prior art documents] [Patent documents]
[0006] [Patent Document 1] Patent Publication No. 2021-196488 Summary of the Invention [Problem to be solved by the invention]
[0007] As described above, a technology has been proposed in which the map of the entire driving environment is divided into a plurality of areas, and the map referenced by the vehicle is updated sequentially according to the progress of the vehicle. However, the technology described in Patent Document 1 is premised on the premise that the vehicle stops once at the boundary between the divided maps to update the map. For this reason, there is a problem that the vehicle decelerates or stops near the boundary between the divided maps due to map switching at the timing of crossing the map or route planning for the next map, which reduces the driving efficiency of the vehicle.
[0008] In addition, the technology described in Patent Document 1 divides the whole area map into equal intervals or into rooms or work areas separated by walls, etc., so there is room for innovation in how to divide the whole area map. In the automotive field, the main function of vehicle self-location estimation, including general navigation, is basically based on the Global Navigation Satellite System (GNSS). Therefore, even if there are few features such as surrounding obstacles, the vehicle can run. On the other hand, it is difficult to use GNSS in indoor driving environments such as logistics warehouses and factories. For this reason, vehicle self-location estimation generally uses a matching technology between on-board sensor information such as LiDAR (Light Detection and Ranging) and cameras and a whole area map created in advance using the on-board sensor information. In addition, in order to realize stable autonomous driving in indoor driving environments such as logistics warehouses, sufficient features around the vehicle are required, such as features effective for matching, for example, unevenness of walls and markers installed in the environment. Therefore, it is insufficient to simply divide the whole area map into rooms or work areas.
[0009] The present invention has been made in consideration of the above-mentioned circumstances, and an object of the present invention is to reduce the calculation load on an in-vehicle controller and ensure the stability of vehicle driving when switching between split maps. [Means for solving the problem]
[0010] The present invention comprises a global route planning unit that plans a global route from a vehicle's travel start position to a target position by referring to a global map of a driving environment that has been created in advance; a map division unit that, when dividing the global map into a plurality of divided maps along the global route so as to include a portion of the global route, divides the divided maps so that adjacent divided maps partially overlap, based on vehicle information of the vehicle and feature information of the global map; and a map update unit that transmits the divided maps to the vehicle in accordance with the vehicle's location and updates the vehicle's reference map. Effect of the Invention
[0011] According to the present invention having the above configuration, it is possible to reduce the calculation load on the in-vehicle controller and ensure the stability of vehicle driving when switching between split maps. Problems, configurations and effects other than those described above will become apparent from the following description of the embodiments. [Brief description of the drawings]
[0012] [Figure 1] 1 is a diagram showing an example of a running environment according to a first embodiment of the present invention; [Diagram 2] 1 is a block diagram showing an example of a functional configuration of an autonomous driving system according to a first embodiment of the present invention. [Diagram 3] 1 is a diagram illustrating an example of the configuration of an autonomous driving vehicle according to a first embodiment of the present invention. [Figure 4] 1 is a block diagram showing the hardware configuration of a map division system and a map updating system according to a first embodiment of the present invention. [Diagram 5] 1 shows an overall map of a driving environment according to a first embodiment of the present invention. [Figure 6] FIG. 13 is a diagram showing an image of dividing an entire area map at equal intervals. [Figure 7] FIG. 2 is a diagram for explaining an obstacle occupancy rate calculation process in the map dividing system according to the first embodiment of the present invention. [Figure 8] FIG. 4 is a diagram showing the score of each area in the distance measurement range of a vehicle sensor, calculated in the map dividing system according to the first embodiment of the present invention. [Figure 9]FIG. 2 is a diagram for explaining the process of expanding a divided map of a current area in the map dividing system according to the first embodiment of the present invention. [Figure 10] FIG. 2 is a diagram for explaining a first pattern of map clipping due to an obstacle sparse area in the map dividing system according to the first embodiment of the present invention. [Figure 11] FIG. 11 is a diagram for explaining a second pattern of map clipping due to an obstacle sparse area in the map dividing system according to the first embodiment of the present invention. [Figure 12] FIG. 2 is a diagram for explaining map division in a direction perpendicular to the traveling direction of a vehicle in the map division system according to the first embodiment of the present invention. [Figure 13] FIG. 2 is a diagram for explaining switching of reference maps when a vehicle travels autonomously in the map updating system according to the first embodiment of the present invention. [Figure 14] 5 is a flowchart showing a procedure of an autonomous driving process in the autonomous driving system according to the first embodiment of the present invention. [Figure 15] 5 is a flowchart showing a procedure of a map division area calculation process in the map division system according to the first embodiment of the present invention. [Figure 16] 5 is a flowchart showing a procedure of a calculation process of an obstacle occupancy rate in the map dividing system according to the first embodiment of the present invention. [Figure 17] 5 is a flowchart showing a procedure of a map division process in the map division system according to the first embodiment of the present invention. [Figure 18] FIG. 11 is a block diagram showing an example of the functional configuration of an autonomous driving system according to a second embodiment of the present invention. [Figure 19] 10 is a flowchart showing a procedure of a map division area calculation process in a map division system according to a second embodiment of the present invention. [Figure 20] 10 is a flowchart showing a procedure for a calculation process of an obstacle occupancy rate in a map dividing system according to a second embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0013] Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings. In this specification and the drawings, components having substantially the same functions or configurations are designated by the same reference numerals, and redundant description will be omitted.
[0014] First Embodiment First, an example of a driving environment according to the first embodiment of the present invention will be described. FIG. 1 is a diagram showing an example of a driving environment according to this embodiment. A driving environment 100 is a driving environment in an example of a warehouse related to a conveyance business, and in FIG. 1, the "X" direction, which is the horizontal direction on the drawing, indicates the width direction of the warehouse, and the "Y" direction intersecting with the "X" direction indicates the length direction of the warehouse. Also, in FIG. 1, the same figures indicate the same things, and only one of them will be described with a reference symbol.
[0015] As shown in FIG. 1, the traveling environment 100 is an environment in which a shelf 111, a pallet 113, a vehicle charging unit 114, a conveyor 115 for transporting luggage from other areas, and the like are placed. A load 112 indicated by a gray square is luggage, and is placed on the shelf 111, the pallet 113, the conveyor 115, and the like. The traveling environment 100 is divided into two parts (two rooms) by a wall 116. In addition, in the traveling environment 100, there is a path surrounded by the shelf 111, the pallet 113, the conveyor 115, and the like. The transport work is performed, for example, by autonomous travel of a vehicle 101, which is a cart-type transport vehicle shown in FIG. 1. Note that, although FIG. 1 shows the traveling environment 100 having two rooms, the present invention is not limited thereto, and the traveling environment 100 may have three or more rooms, or may have only one room. Also, although FIG. 1 shows only one vehicle, it is possible for multiple vehicles to exist in the traveling environment 100.
[0016] Next, an example of the functional configuration of the autonomous driving system according to this embodiment will be described. Fig. 2 is a block diagram showing an example of the functional configuration of the autonomous driving system 300 according to this embodiment. As shown in Fig. 2, the autonomous driving system 300 includes a map distribution server 301, an in-vehicle controller 220 of the vehicle 101, and the vehicle 101. The map distribution server 301, the in-vehicle controller 220, and the vehicle 101 are connected so as to be able to communicate with each other in this order.
[0017] [Example of functional configuration of map division system and map update system] The map distribution server 301 is an example of a map division system and a map update system, and includes an entire area map management unit 302, a vehicle information management unit 303, a target position input unit 304, a global route planning unit 305, a map division unit 306, and a map distribution unit 307. The target position input unit 304, the global route planning unit 305, the map division unit 306, and the map distribution unit 307 are connected in this order. Furthermore, each of the entire area map management unit 302 and the vehicle information management unit 303 is connected to the global route planning unit 305. Furthermore, the vehicle information management unit 303 is also connected to the map division unit 306.
[0018] The whole area map management unit 302 manages a whole area map (see FIG. 5 described later) of the driving environment 100 (see FIG. 1). The whole area map is a grid map created by a SLAM (Simultaneous Localization and Mapping) method or the like using measurement information of an on-board sensor 211 described later of the vehicle 101 traveling throughout the driving environment 100 and movement information (described later) of the vehicle 101 measured by an encoder 209 described later.
[0019] The vehicle information management unit 303 manages (stores) vehicle information including dimensional values of the vehicle 101 (vehicle width, vehicle length, etc.), performance information of the on-board sensor 211 described below, and performance information of the on-board controller 220. The performance information of the on-board sensor 211 includes, for example, the maximum irradiation distance of the LiDAR, the horizontal resolution of the laser light, the distance measurement range, and the distance measurement error. Note that some LiDARs are not stable as the distance is measured farther away, so if the distance at which stable distance measurement is possible for surrounding objects in the driving environment is known through a prior experiment, the value is acquired. Here, the on-board sensor 211 is assumed to have the ability to acquire information (feature amount information) about obstacles in a 360-degree periphery of the vehicle itself.
[0020] The performance information of the in-vehicle controller 220 includes the CPU performance and memory capacity of the information processing device constituting the in-vehicle controller 220, and the maximum size (image size) of a map that can be referred to, which is experimentally calculated from a route plan and a following algorithm described later. Hereinafter, the maximum size of the map that can be referred to is referred to as the "maximum size of the reference map." Note that the unit of the maximum size of the reference map may be the capacity of the image size (megabytes (MB)), or may be the number of pixels if the configuration of the image (such as the configuration of a grayscale image with 0 to 255 gradations) is set in advance. In addition, although the description will be given with one vehicle as an example in this embodiment, if there are multiple vehicles traveling in the traveling environment 100, the vehicle information management unit 303 manages the vehicle information of each vehicle.
[0021] The target position input unit 304 inputs the entire area map from the entire area map management unit 302, and sets the coordinates in the entire area map of the destination of the task currently being performed by the vehicle 101 as the target position of the autonomous driving. Here, the target position of the autonomous driving may be directly input by an operator via an external device such as a mobile display, for example. Also, the target position of the autonomous driving may be automatically calculated by acquiring the destination of the task set in advance in the map distribution server 301. Also, the target position input unit 304 outputs the target position of the autonomous driving to the global route planning unit 305.
[0022] The global route planning unit 305 inputs the global map from the global map management unit 302, inputs the dimensional values of the vehicle 101 from the vehicle information management unit 303, and inputs the target position from the target position input unit 304. The global route planning unit 305 also refers to a global map of the driving environment created in advance to plan a global route (see FIG. 5 described later) from the driving start position of the vehicle 101 to the target position. The global route is expressed as a set of lattice coordinates (matrix and column numbers, etc.) on the global map, and the target position is reached by driving in the order of the coordinates in the set. The global route planning unit 305 also plans the target speed and maximum speed of the vehicle when passing through each lattice included in the global route, and outputs them together with the global route. Note that any route search method, such as Dijkstra's algorithm or A* (A star) algorithm, may be used as the route planning method. In addition, the global route planning unit 305 outputs the entire area map on which the global route has been planned to the map division unit 306, and transmits it to a local route planning unit 313 (described later) of the in-vehicle controller 220.
[0023] The map division unit 306 acquires vehicle information of the vehicle 101 from the vehicle information management unit 303, and acquires an entire area map on which a global route is planned from the global route planning unit 305. The map division unit 306 also divides the entire area map into a plurality of divided maps along the global route so as to include a part of the global route. When dividing the entire area map into a plurality of divided maps, the map division unit 306 performs a map divided area calculation process based on the vehicle information of the vehicle 101 and feature amount information of the entire area map, in which adjacent divided maps are divided so as to partially overlap each other. In this embodiment, the feature amount information of the entire area map is information on obstacles in the entire area map (grid map). The map division unit 306 also outputs the divided map (hereinafter referred to as a "divided map") to the map distribution unit 307. The map divided area calculation process in the map division unit 306 will be described in detail later with reference to Figs. 15 to 17.
[0024] The map distribution unit 307 inputs the divided map from the map division unit 306, and inputs the location of the vehicle 101 from a self-position estimation unit 312 (described later) of the in-vehicle controller 220. The map distribution unit 307 is also an example of a map update unit, and transmits the divided map to the vehicle 101 (a reference map management unit 311 (described later) of the in-vehicle controller 220) according to the location of the vehicle 101, i.e., the progress of the vehicle 101, and updates the reference map of the vehicle 101.
[0025] [Example of vehicle controller configuration] Next, the functional configuration of the in-vehicle controller 220 will be described. As shown in FIG. 3 described later, the in-vehicle controller 220 is an example of an information processing device that is arranged inside the vehicle 101 and controls the operation of each component of the vehicle 101 so as to make the vehicle 101 travel autonomously. As shown in FIG. 2, the in-vehicle controller 220 includes a reference map management unit 311, a self-position estimation unit 312, a local route planning unit 313, and a control command generation unit 314. The reference map management unit 311 is connected to each of the self-position estimation unit 312 and the local route planning unit 313. In addition, the self-position estimation unit 312 is connected to each of the local route planning unit 313 and the control command generation unit 314. In addition, the local route planning unit 313 is connected to the control command generation unit 314. The control command generation unit 314 is connected to the drive unit 208 of the vehicle 101. Furthermore, the reference map management unit 311 receives divided maps from the map distribution unit 307 of the map distribution server 301 , and outputs them to the self-position estimation unit 312 and the local route planning unit 313 .
[0026] The self-position estimation unit 312 inputs the divided map from the reference map management unit 311, and uses the input divided map as the reference map. The self-position estimation unit 312 also inputs point cloud information detected by the in-vehicle sensor 211 of the vehicle 101, performs map matching processing between the point cloud information and the reference map, and estimates the self-position and attitude. Map matching is a technique in which, when the point cloud information is regarded as a figure, the point cloud is translated, rotated, and enlarged / reduced to fit into a grid map (reference map). In the map matching processing, any method used for position estimation, such as the AMCL (Adaptive Monte Carlo Localization) method, may be used. The self-position estimation unit 312 also outputs the estimated self-position to the local route planning unit 313 and the control command generation unit 314. The self-position estimation unit 312 also transmits the estimated self-position to the map distribution unit 307 of the map distribution server 301.
[0027] The local route planning unit 313 receives a global route from the global route planning unit 305 of the map distribution server 301, inputs a divided map from the reference map management unit 311, and inputs the self-position from the self-position estimation unit 312. The local route planning unit 313 also refers to the divided map along the global route, sets a grid located between the self-position and the target position (for example, a grid on the global route 2 m ahead from the self-position) as a local target position, and plans a local route from the self-position to the local target position. The local route planning unit 313 also outputs the planned local target position and local route to the control command generation unit 314. Note that the local route is not a set of grid coordinates, but is formed by a set of finer coordinates or a function expressing a curve, and is a smooth route that takes into account the motion characteristics of the vehicle 101. Specifically, the local route planning unit 313 plans a local route using a method such as spline function interpolation or DWA (Dynamic Window Approach) that takes into account a motion model of the vehicle 101. The method for planning the local route is not limited to the above-mentioned method, and any method can be applied as long as it is a method capable of planning a route that is more detailed than the global route.
[0028] The control command generating unit 314 inputs the local target position and the local route from the local route planning unit 313. The control command generating unit 314 also inputs movement information of the vehicle 101 from the encoder 209, and inputs the vehicle's own position from the self-position estimating unit 312. The control command generating unit 314 also performs path tracking control to generate a drive command for the vehicle 101 using a path tracking algorithm so that the vehicle 101 passes through the local route from its own position toward the local target position. The control command generating unit 314 also outputs the generated drive command for the vehicle 101 to the drive unit 208 of the vehicle 101. For example, the Pure Pursuit method or the DWA method is used as the path tracking algorithm.
[0029] 2, only the driving unit 208, the encoder 209, and the on-board sensor 211 of the vehicle 101 are shown, and the illustration and description of the other components will be described in detail with reference to FIG. 3 described later. The driving unit 208 is connected to a control command generating unit 314 of the on-board controller 220, and drives the driving wheels 206 of the vehicle 101 according to a control command from the control command generating unit 314. The encoder 209 outputs measured movement information of the vehicle 101 (described later) to the self-position estimating unit 312 and the control command generating unit 310 of the on-board controller 220. The on-board sensor 211 outputs detected measurement information (described later) to the self-position estimating unit 312 of the on-board controller 220. The configuration of the vehicle 101 will be described in detail with reference to FIG. 3 described later.
[0030] [Example of autonomous vehicle configuration] Next, a configuration example of an autonomous vehicle (vehicle 101) will be described. FIG. 3 is a diagram showing a configuration example of an autonomous vehicle according to this embodiment. The vehicle 101 includes a vehicle frame 201 having a bumper 202, a transfer device 203 that can move up and down and left and right on the vehicle frame 201, and a loading member 204 that supports a load 112. The vehicle 101 also includes a driving wheel 206, a driven wheel 207, and an in-vehicle sensor 211 disposed on the vehicle frame 201. Inside the vehicle frame 201, a transfer motor 205 that drives the transfer device 203 and a drive unit 208 that drives the drive wheel 206 are provided. Inside the vehicle frame 201, an encoder 209 and an in-vehicle controller 220 are provided. The in-vehicle controller 220 has been described in FIG. 2, so a duplicate description will be omitted.
[0031] The drive unit 208 includes a driving motor (not shown), and converts a drive command for the vehicle 101 input from a control command generation unit 314 (described later) of the in-vehicle controller 220 into a predetermined current value and outputs it to the driving motor. In addition, the encoder 209 acquires the amount of rotation of the driving motor output at each time, and measures movement information including the travel distance and current speed of the vehicle, taking into account the wheel diameter of the drive wheels 206.
[0032] The on-vehicle sensor 211 is, for example, a LiDAR. The LiDAR is a sensor that measures an object present in an irradiation range and the distance to the object while changing the irradiation angle of the laser light. The LiDAR acquires point cloud information indicating the position of the object from information on the optical axis angle at the time of laser irradiation and information on the distance to the object. Hereinafter, the point cloud information of the obstacle detected by the on-vehicle sensor 211 and the distance to the obstacle are collectively referred to as "measurement information". In this embodiment, a 2D (Dimensions)-LiDAR that acquires two-dimensional information of only one layer in the direction perpendicular to the laser light is taken as an example of the on-vehicle sensor 211, but the present invention is not limited to this. The on-vehicle sensor 211 may be a 3D-LiDAR that irradiates multiple layers and acquires three-dimensional information. In addition, only one on-vehicle sensor 211 is installed on the vehicle 101 shown in FIG. 2, but the present invention is not limited to this, and multiple on-vehicle sensors 211 may be installed on the vehicle 101.
[0033] [Example of hardware configuration for map division system and map update system] Next, a description will be given of the hardware configuration of the map division system and map update system, i.e., the map distribution server 301, in the autonomous driving system 300. Fig. 4 is a block diagram showing an example of the hardware configuration of the map division system and map update system according to this embodiment. The map distribution server 301 performs various calculations related to the map division process and the distribution process of the divided maps.
[0034] 4, the map distribution server 301 includes a CPU 30a, a ROM (Read Only Memory) 30b, a RAM (Random Access Memory) 30c, a storage device 30d, and a communication interface 30e. The CPU 30a, the ROM 30b, the RAM 30c, the storage device 30d, and the communication interface 30e are connected via a bus 30f so as to be able to transmit and receive information data to and from each other.
[0035] The CPU 30a controls the operation of each component in the map distribution server 301. For example, the CPU 30a controls global route planning processing in the global route planning unit 305, map division area calculation processing in the map division unit 306, distribution processing of divided maps in the map distribution unit 307, etc. Note that a GPU (Graphics Processing Unit) may be used instead of the CPU 30a, or the CPU 30a and a GPU (Graphics Processing Unit) may be used in combination.
[0036] The ROM 30b is configured with a storage medium such as a non-volatile memory, and stores programs and data executed and referenced by the CPU 30a.
[0037] The RAM 30c is configured with a storage medium such as a volatile memory, and temporarily stores information (data) necessary for each process performed by the CPU 30a.
[0038] The storage device 30d is composed of a computer-readable non-transient recording medium storing a program executed by the CPU 30a, and is composed of a storage device such as an HDD (Hard Disk Drive). The storage device 30d stores programs for the CPU 30a to control each part, an OS (Operating System), a controller, and other programs and data. The storage device 30d also includes an overall area map management unit 302 and a vehicle information management unit 303 (see FIG. 2), and stores an overall area map, a general route, a divided map, vehicle information, and the like. Note that some of the programs and data stored in the storage device 30d may be stored in the ROM 30b. The computer-readable non-transient recording medium storing the program executed by the CPU 30a is not limited to an HDD, and may be, for example, a recording medium such as an SSD (Solid State Drive), a CD (Compact Disc)-ROM, or a DVD (Digital Versatile Disc)-ROM.
[0039] The communication interface 30e is composed of a NIC (Network Interface Card), a modem, etc., and establishes a connection with the vehicle controller 220, and transmits and receives information data to and from the vehicle controller 220 under the control of the CPU 30a.
[0040] Next, the whole area map of the driving environment 100 according to this embodiment will be described. Before the vehicle 101 starts autonomous driving, a whole area map of the whole driving environment is generated and registered in the whole area map management unit 302 of the map distribution server 301. In this embodiment, the resolution of the whole area map, that is, the size (cm / pixel) of each grid in real space is set to 5cm / pixel, taking into account the width of the whole driving environment. FIG. 5 is a diagram showing the whole area map P400 of the driving environment 100 shown in FIG. 1. In FIG. 5, the "X" and "Y" directions are the same as those directions shown in FIG. 1, so duplicated explanations will be omitted. Note that the resolution of the whole area map is not limited to 5cm / pixel, and may be other resolutions.
[0041] In FIG. 5, the gray grids represent obstacles. The hatched grids represent grids included in the global route. The driving start position G401 shown in FIG. 5 is the starting position of the autonomous driving of the vehicle 101 in the driving environment 100. The driving target position G402 is the target position of the vehicle 101. The global route R403 is the global route from the driving start position G401 to the driving target position G402 planned by the global route planning unit 305. The vehicle 101 can arrive at the driving target position G402 by driving along the global route R403 shown in FIG. 5. Note that in FIG. 5, the obstacles represented by the gray grids correspond to the actual size of the obstacles.
[0042] Next, problems that arise when dividing an entire area map at equal intervals will be described. Fig. 6 is a diagram showing an image of dividing an entire area map P400 at equal intervals. In Fig. 6, the "X" and "Y" directions are the same as those shown in Fig. 1, so repeated explanations will be omitted. Fig. 6A shows some divided maps obtained by dividing the entire area map at equal intervals. The divided maps Pa to Pc are each a partial area on the lower left side of the entire area map P400 divided at equal intervals. The divided maps Pa to Pc have the same size and are equal to or smaller than the maximum reference map size of the in-vehicle controller 220.
[0043] FIG. 6A shows a distance measurement range R601 of the on-board sensor 211 (LiDAR) centered on an arbitrary grid C602 included in a global route R403. Since the vehicle 101 travels along the global route R403, if there is a certain amount of information (gray grids) about obstacles included in the distance measurement range R601, the estimation accuracy of the self-position estimation unit 312 is stable. Therefore, unnecessary acceleration / deceleration and deviation from the route of the vehicle 101 do not occur, and the vehicle 101 can travel autonomously stably. When the vehicle 101 travels through the area of the divided map Pa and enters the area of the divided map Pb, the reference map is switched to the divided map Pb. Then, when the vehicle 101 travels through the area of the divided map Pb and enters the area of the divided map Pc, the reference map is switched to the divided map Pc.
[0044] FIG. 6B is a diagram showing a state after the vehicle 101 has entered the area of the divided map Pc. As shown in FIG. 6B, in the divided map Pc, there is no obstacle information within the distance measurement range R601. In the distance measurement range R601, there is no obstacle information (feature amount information) in the upper half, and there is no reference map in the lower half. In this case, since there is no feature amount information with which the self-position estimation unit 312 can estimate the self-position, the self-position estimation unit 312 estimates the self-position using only the movement information detected by the encoder 209, and there is a high possibility that an estimation error in the self-position occurs. As a result, there is a high possibility that the vehicle 101 will accelerate or decelerate unnecessarily, deviate from the route excessively, and the like will occur.
[0045] The present invention has been made to solve the above problem. The map division unit 306 of the map distribution server 301 (map division system) according to this embodiment subdivides the ranging range R601 of the on-board sensor 211 of the vehicle 101, calculates the obstacle occupancy rate in each of the subdivided child regions, and divides the entire map according to the obstacle occupancy rate. FIG. 7 is a diagram for explaining the subdivision of the forward ranging range and the rear ranging range in the map division system according to this embodiment. FIG. 7A is a diagram for explaining the forward ranging range and the rear ranging range of the on-board sensor 211. FIG. 7B is a diagram for explaining the subdivision of the ranging range of the on-board sensor 211. In addition, in FIG. 7A, the "X" and "Y" directions are the same as those directions shown in FIG. 1, so that a duplicated explanation will be omitted. In this embodiment, since the reference map is a lattice map, the map division unit 306 divides the map in parallel in the "X" direction and the "Y" direction when dividing the map.
[0046] 7A shows the current reference map of the vehicle 101, a global route R403 from the travel start position G401 to the target position, and a distance measurement range R601 of the on-board sensor 211. When the vehicle 101 travels along the global route R403, it travels by moving from the travel start position G401 to one grid at a time along the global route R403. As shown in FIG. 7A, the vehicle 101 first moves from the travel start position G401 to the grid adjacent to the right of the travel start position G401. At this time, the travel vector indicating the travel direction of the vehicle 101 is an arrow F701, which is in the "X" direction.
[0047] The map division unit 306 first subdivides the measurement range R601 into two semicircular regions, a forward measurement range R702 and a rear measurement range R703, along a direction ("Y" direction) perpendicular to the traveling direction (direction indicated by the arrow F701) of the vehicle 101. Next, the map division unit 306 further subdivides each of the forward measurement range R702 and the rear measurement range R703. Note that since the subdivision of each of the forward measurement range R702 and the rear measurement range R703 is similar, the following description will be given taking the forward measurement range R702 as an example, and a description of the rear measurement range R703 will be omitted.
[0048] FIG. 7B shows the forward ranging range R702. In order to clearly show each divided child region, FIG. 7B does not show the gray grid indicating the obstacle, that is, the grid occupied by the obstacle. The map division unit 306 first divides the forward ranging range R702 into regions s1, s2, and s3. The region s1 is a semicircular region with a radius of 10 m. The region s2 is a sector-shaped ring region between a semicircle with a radius of 10 m and a semicircle with a radius of 20 m. The region s3 is a sector-shaped ring region between a semicircle with a radius of 20 m and a semicircle with a radius of 30 m. The method of dividing the forward ranging range R702 is not limited to the above method, and any division method may be applied. The number of divided regions also changes according to different division methods.
[0049] The map division unit 306 then further divides the regions s1, s2, and s3 into smaller regions by any central angle a1, a2 to an, for example, the central angles a1, a2 to an, as shown in Fig. 7B. Note that Fig. 7B shows an example in which the central angles are set as a1=a2...a6=30°, but the present invention is not limited to this, and the central angle can be changed to any angle, and the central angles a1, a2 to an may each be a different angle.
[0050] The map division unit 306 then calculates the number of obstacle occupancy lattices (not shown) in each subdivided child region (small regions surrounded by solid lines in FIG. 7B). The number of obstacle occupancy lattices included in each child region is, for example, "0", "1", "2", "3", etc. in each child region as shown in FIG. 7B. The map division unit 306 then assigns a weight value to each of the regions s1, s2, and s3, and calculates a score value S indicating the obstacle occupancy rate of the forward ranging range R702 using formula (1). The score value of the rear ranging range R703 is calculated using the same calculation method as that for the forward ranging range R702.
[0051]
number
[0052] In formula (1), W1, W2, and W3 are preset weight values for the region s1, region s2, and region s3, respectively. Here, the closer to the vehicle 101, the larger the weight value is set. For example, W1, W2, and W3 are set to "10", "5", and "1", respectively. Note that the setting of the weight value is not limited to the above-mentioned method, and may be arbitrarily changed as necessary. Also, in formula (1), s1(a), s2(a), and s3(a) are the number of occupied grids of the obstacle in the child region of each region, respectively, and n is the number of child regions included in each region (region s1, region s2, and region s3) ("6" in the example shown in FIG. 7B). Note that the number n of child regions included in each region may be a different value depending on the subdivision method.
[0053] Fig. 8 is a diagram showing scores of the forward ranging range and the rear ranging range corresponding to the traveling direction of the vehicle 101. Fig. 8A is a diagram showing an image of the forward ranging range and the rear ranging range corresponding to the traveling direction of the vehicle 101 on the whole area map. In Fig. 8A, the "X" and "Y" directions are the same as those directions shown in Fig. 1, so a duplicated explanation will be omitted. Fig. 8B is a diagram showing scores of the forward ranging range and the rear ranging range.
[0054] In Fig. 8A, an area AF801 represents a forward ranging range corresponding to the traveling direction of the vehicle 101, and an area AR802 represents a rear ranging range corresponding to the traveling direction of the vehicle 101. Also, "ID" shown in Fig. 8A is a number indicating the coordinate order of each lattice in the set of lattice coordinates of the global route R403. For example, as shown in Fig. 8A, the ID of the lattice at the travel start position is "1", and the ID of the 22nd lattice along the global route R403 is "22".
[0055] Fig. 8B shows the score values of the area AF801 and the area AR802 corresponding to the ID (1 to N) of each grid on the global route R403. Here, N is the grid number of the target position. Each score value shown in Fig. 8B is calculated by the above-mentioned formula (1). Note that in Fig. 8, the ID of each grid on the global route R403 is a number indicating the order of each grid, but the present invention is not limited to this, and for example, each grid may be a coordinate on the global map.
[0056] Next, the map cut-out in the map division unit 306 will be described. First, the map division unit 306 expands the divided map of the current area to the maximum size of the reference map of the in-vehicle controller 220 (see FIG. 9 described later). Next, the map division unit 306 detects an area with little obstacle information within the distance measurement range of the in-vehicle sensor 211 of the vehicle 101, that is, an area without (valid) feature amount information that can estimate the vehicle's location. Here, an area without feature amount information that can estimate the vehicle's location is detected based on whether or not the score indicating the obstacle occupancy rate is less than a predetermined value. Hereinafter, an area with a score less than a predetermined value is referred to as an "obstacle sparse area". In this embodiment, the predetermined value is "10". The predetermined value can be changed to another value.
[0057] Here, when an obstacle sparse area is detected, the case where the measurement range R601 of the on-board sensor 211 reaches the boundary in the "Y" direction of the maximum size of the reference map is determined as the first map cutout pattern. When an obstacle sparse area is detected, the measurement range of the on-board sensor 211 has a sufficient margin to the boundary in the "Y" direction of the divided map of the current area is determined as the second map cutout pattern. In the first map cutout pattern, the map division unit 306 sets the current position of the vehicle 101 as the cutout position of the divided map of the current area, and cuts out the divided map of the current area. In the first map cutout pattern, the map division unit 306 calculates the overlapping area between the divided map of the next area and the divided map of the current area based on the score of the measurement range of the on-board sensor 211, and determines the map cutout position in the opposite direction of the "Y" direction of the divided map of the next area (see FIG. 10 described later). In the second map cropping pattern, the measurement range of the onboard sensor 211 is sufficiently large to the boundary in the "Y" direction of the divided map of the current area, so even if an obstacle-sparse area is detected, the map division unit 306 does not perform map cropping (see Figure 11 described below).
[0058] Fig. 9 is a diagram for explaining the process of expanding a divided map of a current area in the map division system according to this embodiment. Fig. 9A is a diagram showing an image in which the ranging range R601 is superimposed when the vehicle 101 travels along a general route. Fig. 9B is a diagram for explaining the calculation of the overlapping area of adjacent divided maps. In Fig. 9, the "X" and "Y" directions are the same as those directions shown in Fig. 1, so a duplicated explanation will be omitted.
[0059] 9A shows a rectangular area R900 surrounding each of the overlapping distance measurement ranges R601. The map division unit 306 divides the rectangular area R900 into two areas by the side length L X and L Y Either of the above is set to the maximum reference map size (L X_max , L Y_max ), cut out the current rectangular area R900 as a segmented map. Here, L X and L Yare the side lengths of the rectangular area along the "X" and "Y" directions, respectively. X_max , L Y_max are the side lengths along the "X" and "Y" directions of the maximum size of the reference map, respectively.
[0060] Furthermore, the map division unit 306 cuts out a part of the area that overlaps the divided map of the current area and the divided map of the next area so that it is included in the divided map of the next area, depending on the processing time required to switch from the divided map of the current area to the divided map of the next area. Fig. 9B shows a part of the divided map of the current area (rectangular area R900), a part of the divided map of the next area R901, and an overlapping area R902 (hatched area) of the two divided maps.
[0061] The map division unit 306 calculates the side length L of the overlapping area R902 in the traveling direction (the "Y" direction) of the vehicle 101 by the following formula (2). mer1 Calculate.
[0062]
number
[0063] v in Equation (2) max is the maximum speed at which the vehicle 101 can travel. m is the processing time required to switch the reference map to the divided map of the next area, that is, the processing time required to change the memory area to be referenced. s is the calculation period for the driving control of the vehicle 101.
[0064] The map division unit 306 determines a cut-out position in the reverse "Y" direction for the divided map of the next area so that the calculated overlapping area R902 (hatched portion) is included in the divided map of the next area. This process solves the problem of there being no reference map in the reverse direction of the traveling direction of the vehicle 101 when switching from the divided map of the current area to the divided map of the next area (see FIG. 6B).
[0065] Fig. 10 is a diagram for explaining a first pattern of map clipping due to an obstacle sparse area in the map division system according to the present embodiment. Fig. 10A is a diagram for explaining the map clipping position of the first pattern. Fig. 10B is a diagram for explaining the calculation of the overlapping area shown in Fig. 10A. In Fig. 10, the "X" and "Y" directions are the same as those directions shown in Fig. 1, so repeated explanations will be omitted.
[0066] 10A shows the vehicle 101, the on-board sensor 211 of the vehicle 101, an area AF801 which is a forward ranging range corresponding to the traveling direction of the vehicle 101, and an area AF802 which is a rear ranging range. Here, it is assumed that the current traveling direction of the vehicle 101 is the "Y" direction, and that an area AR802 (rear ranging range) includes an obstacle-sparse area.
[0067] The target position L1001 represents the position of the side of the grid in the traveling direction, which is the current position of the vehicle 101. The position L1002 represents the boundary in the traveling direction of the distance measurement range R601 of the on-board sensor 211. Therefore, the distance between the target position L1001 and the position L1002 is the radius L of the distance measurement range R601. lidar / 2. Distance L y,i represents the distance from the target position L1001 to the boundary of the current divided map on the opposite side of the proceeding direction, that is, the length along the "Y" direction of the current divided map.
[0068] The map division unit 306 divides the map into 3 parts by a radius L lidar / 2 and distance L y,i The sum of these values is the maximum size of the reference map L Y_max If it is determined that the target position L1001 has been reached, the target position L1001 is cut out as the boundary of the current divided map in the traveling direction. In other words, the map from the target position L1001 to the boundary of the divided map of the current area in the opposite direction to the traveling direction is cut out as the current divided map.
[0069] In addition, the map division unit 306 calculates the overlapping area (hatched portion) between the divided map of the next area and the divided map of the current area, that is, the overlapping distance L shown in FIG. 10A, based on the score value of the area AR802 including the obstacle sparse area. mer2The map division unit 306, for example, as shown in FIG. 10B, divides the area AR802 into smaller areas, calculates the score of each divided area, and calculates the overlap distance L based on the calculated score. mer2 The subdivided areas s1, s2, and s3 shown in FIG. 10B are the same as those described in FIG. 7B, and a duplicated description will be omitted. In the example shown in FIG. 10B, the map division unit 306 does not further subdivide the areas s1, s2, and s3 by the central angle, but further subdivides the areas s1, s2, and s3 by a predetermined distance from the target position L1001. That is, the map division unit 306 subdivides the areas s1, s2, and s3 in parallel in the "X" direction. This is because, when dividing the reference map, which is a grid map, it is divided in parallel in the "X" direction and the "Y" direction.
[0070] As shown in FIG. 10B, the map division unit 306 subdivides the areas s1, s2, and s3 into child areas (small areas surrounded by solid lines) in parallel with the "X" direction at positions where the predetermined distance from the target position L1001 is 5 m, 10 m, 15 m, and 20 m, respectively. Hereinafter, the part of the area AR802 from the target position L1001 to 5 m from the target position L1001 is referred to as "area A1," and the part of the area AR802 from 5 m to 10 m is referred to as "area A2." In addition, the part of the area AR802 from 10 m to 15 m is referred to as "area A3," the part of the area AR802 from 15 m to 20 m is referred to as "area A4," and the part of the area AR802 beyond 15 m is referred to as "area A5." Then, the map division unit 306 calculates the number of grids occupied by the obstacle in each child region in the same manner as described with reference to FIG. 7B, and calculates the score value of each of the regions A1 to A5 using the formula (1).
[0071] Then, the map division unit 306 performs a calculation to accumulate the score values of the areas A1 to A5 in order, and when the accumulated score value reaches or exceeds a predetermined threshold, the distance to the target position L1001 of the boundary line between the area corresponding to the last accumulated score value and the next area is calculated as the overlap distance L mer2 In addition, the map division unit 306 divides the map into two parts by the overlap distance L mer2The map division unit 306 determines the cutting position in the reverse direction of the "Y" direction of the divided map of the next area so that the map range leaving the "Y" direction is included in the divided map of the next area. mer1 Also calculate L mer1 and L mer2 and the larger one may be determined as the overlap distance.
[0072] Fig. 11 is a diagram for explaining a second pattern of map clipping due to an obstacle-sparse area in the map division system according to the present embodiment. In Fig. 11, the "X" and "Y" directions are the same as those shown in Fig. 1, so duplicated explanations will be omitted.
[0073] The second pattern shown in FIG. 11 is different from the first pattern shown in FIG. 10A in that the radius L lidar / 2 and distance L y,i The sum of these values is the maximum size of the reference map L Y_max In this example, since the divided map of the current area has a margin for the maximum size of the reference map, even if an obstacle-sparse area is included, the map division unit 306 does not cut the map until the maximum size of the reference map is reached.
[0074] In the above explanation, the maximum size of the reference map is L Y_max That is, the process of determining the cut-out position of the divided map in the "Y" direction (travel direction) based on the maximum size of the reference map in the "Y" direction has been described. Next, the process of determining the cut-out position of the divided map in the direction perpendicular to the traveling direction of the vehicle 101 ("X" direction) will be described. Fig. 12 is a diagram for explaining map division in the direction perpendicular to the traveling direction of the vehicle 101 in the map division system according to this embodiment.
[0075] In Figure 12, the length of the divided map in the "Y" direction is L Y_max After determining the cut position in the "Y" direction of the divided map, the map division unit 306 determines the length L XBased on this, it is expanded equally in the positive and negative directions of the "X" direction, and the expanded L X is the length L in the "X" direction of the maximum size of the reference map X_max In the example shown in FIG. 12, when the divided map is expanded in the negative direction of the "X" direction, the boundary of the whole area map is reached, so after the boundary of the whole area map is reached, the divided map is cut only in the positive direction of the "X" direction. X Expanding L X L X_max The map division unit 306 stores the coordinate information of the rectangular area R1201 obtained by expanding the divided map in the positive and negative directions of the "X" direction as area information of the divided map.
[0076] Next, switching (updating) of the reference map when the vehicle 101 runs autonomously, and the reference map held by the reference map management unit 311 of the in-vehicle controller 220 will be described. Fig. 13 is a diagram for explaining switching of the reference map when the vehicle runs autonomously in the map division system (map update system) according to this embodiment. Fig. 13A is a diagram showing an image of switching of the reference map when the vehicle 101 runs autonomously. Fig. 13B is a diagram showing an image of the reference map held by the reference map management unit 311 of the in-vehicle controller 220.
[0077] FIG. 13A shows a global route R403 from a travel start position G401 of the vehicle 101 to a travel target position G402. FIG. 13A also shows divided maps (a) to (e) along the global route R403. When the vehicle 101 travels along the global route R403, the map distribution unit 307 of the map distribution server 301 first transmits the divided maps (a) and (b) as reference maps to the reference map management unit 311 of the in-vehicle controller 220. When the vehicle 101 enters the range of the divided map (b), the map distribution unit 307 switches the reference map to the divided maps (b) and (c) and transmits them to the reference map management unit 311. The map distribution unit 307 executes a process of transmitting the divided maps as the reference maps to the reference map management unit 311 until the vehicle 101 reaches the travel target position G402.
[0078] The reference map management unit 311 always holds two types of reference maps (divided maps). Specifically, as shown in FIG. 13B, the reference map management unit 311 holds a reference map M131 of the current area (e.g., divided map (a) and divided map (b)) and a reference map M132 of the next area (e.g., divided map (b) and divided map (c)). When the vehicle 101 is in the area of divided map (a), the reference map management unit 311 outputs the reference map M131 of the current area to the local route planning unit 313. When the vehicle 101 enters the area next to the area of divided map (a), the reference map management unit 311 outputs the reference map M132 of the next area to the local route planning unit 313.
[0079] [Autonomous driving processing procedure] Next, a description will be given of the procedure of the autonomous driving process in the autonomous driving system 300. Fig. 14 is a flowchart showing the procedure of the autonomous driving process in the autonomous driving system 300 according to this embodiment. The process described below starts when the vehicle 101 starts autonomous driving.
[0080] First, the self-position estimation unit 312 of the in-vehicle controller 220 of the vehicle 101 acquires the self-position, that is, the travel start position (step S11). In this process, the self-position estimation unit 312 transmits the self-position to the map distribution unit 307 of the map distribution server 301. Note that the map distribution unit 307 of the map distribution server 301 may store the previous travel end position of the vehicle 101 as the next travel start position.
[0081] Next, the target position input unit 304 of the map distribution server 301 sets a target position for autonomous driving (step S12). In this process, for example, a database in which coordinate values indicating the target position of the vehicle 101 are set in advance is imported, and the target position registered in the database is set. Note that other setting methods may be applied to set the target position. In this process, the target position input unit 304 also outputs the set target position to the global route planning unit 305.
[0082] Next, the global route planning unit 305 of the map distribution server 301 plans and generates a global route (see FIG. 5) on the entire area map based on the travel start position of the vehicle 101, the target position, and the vehicle dimension values input from the vehicle information management unit 303 (step S13). In addition, in this process, the global route planning unit 305 outputs the entire area map on which the global route is planned to the map division unit 306 and the local route planning unit 313 of the in-vehicle controller 220.
[0083] Next, the map division unit 306 of the map distribution server 301 acquires the performance information of the in-vehicle sensor from the vehicle information management unit 303 (step S14).
[0084] Next, the map division unit 306 acquires the maximum size of the reference map including the performance information of the in-vehicle controller 220 from the vehicle information management unit 303 (step S15).
[0085] Next, the map division unit 306 performs a map division area calculation process using the whole area map for which the global route is planned, the performance information of the in-vehicle sensor, and the maximum size of the reference map (step S16). The procedure of the map division area calculation process will be described in detail later with reference to Figs. 15 to 17.
[0086] Next, the map division unit 306 records area information of the divided map, which is a result of the map divided area calculation process (step S17).
[0087] Next, the map distribution unit 307 of the map distribution server 301 performs an initial map distribution to the in-vehicle controller 220 of the vehicle 101 (step S18). In this process, the map distribution unit 307 transmits a reference map of the current area (see reference map M131 of the current area in FIG. 13 ) including a divided map that includes the travel start position of the vehicle 101 to the reference map management unit 311 of the in-vehicle controller 220.
[0088] Next, the self-location estimation unit 312 of the in-vehicle controller 220 inputs the reference map of the current area from the reference map management unit 311 and estimates the self-location (step S19). In this process, the self-location estimation unit 312 outputs the estimated self-location to the local route planning unit 313 and transmits it to the map distribution unit 307 of the map distribution server 301.
[0089] Next, the map distribution unit 307 transmits the reference map to the reference map management unit 311 according to the location of the vehicle 101 (the above-mentioned self-location), and the reference map management unit 311 updates the reference map (see FIG. 13B) held therein (step S20). Also, in this process, when the self-location estimation unit 312 detects that the estimated self-location has entered the area of the next divided map (see divided map (b) in FIG. 13), it updates the reference map to the reference map of the next area.
[0090] Next, the local route planning unit 313 of the in-vehicle controller 220 sets a local target position using the self-position input from the self-position estimation unit 312 and a reference map including a part of the global route (step S21).
[0091] Next, the local path planning unit 313 plans and generates a local path from the self-position to the local target position (step S22).
[0092] Next, the control command generating unit 314 of the in-vehicle controller 220 performs path following control (step S23). In this process, the control command generating unit 314 generates a driving command for the vehicle 101 by a path following algorithm so as to pass through the local path from the vehicle's own position toward the local target position. The control command generating unit 314 also outputs the generated driving command for the vehicle 101 to the driving unit 208 of the vehicle 101. Then, the driving unit 208 drives the vehicle 101 so as to travel according to the driving command.
[0093] Next, the self-position estimation unit 312 judges whether the target position has been reached (step S24). In this process, the self-position estimation unit 312 compares the self-position with the target position, and if the distance between the two points is equal to or less than a predetermined threshold, for example, 0.1 m, it judges that the target position has been reached, and step S24 is judged as YES. The self-position estimation unit 312 also compares the self-position with the target position, and if the distance between the two points exceeds a predetermined threshold, it judges that the target position has not been reached, and step S24 is judged as NO. Note that the predetermined threshold set for the distance between the current position and the target position is not limited to 0.1 m, and may be changed to an appropriate value depending on the actual situation.
[0094] In the process of step S24, if the self-position estimation unit 312 determines that the vehicle has not reached the target position (NO determination in step S24), the process returns to the process of step S19, and the processes of steps S19 to S24 are repeatedly executed.
[0095] On the other hand, in the process of step S24, if the self-position estimation unit 312 determines that the vehicle has reached the target position (YES determination in step S24), the autonomous driving process ends with the driving completed state.
[0096] [Map division area calculation process procedure] Next, a procedure for map division area calculation processing in the map division unit 306 of the map distribution server 301 will be described. Fig. 15 is a flowchart showing the procedure for map division area calculation processing in the map division system according to this embodiment. The map division area calculation processing is called and executed as a subroutine in step S16 of the autonomous driving processing shown in Fig. 14.
[0097] First, the map division unit 306 of the map distribution server 301 performs a calculation process of the obstacle occupancy rate (step S70). Note that the procedure of the calculation process of the obstacle occupancy rate will be described in detail later with reference to FIG.
[0098] Next, the map division unit 306 performs map division processing based on the result of the obstacle occupancy calculation processing (step S80). After the processing of step S80, the map division area calculation processing ends. The procedure of the map division processing will be described in detail later with reference to FIG. 17.
[0099] 16 is a flowchart showing the procedure of the obstacle occupancy calculation process in the map division system according to this embodiment. The obstacle occupancy calculation process is called and executed as a subroutine in step S70 of the map division area calculation process shown in FIG.
[0100] First, the map division unit 306 acquires and updates the target position for which the obstacle occupancy rate is to be calculated (step S701). In this process, the initial position of the target position is the vehicle's travel start position, and the map division unit 306 updates the target position by moving one grid at a time from the travel start position along the global route.
[0101] Next, the map division unit 306 acquires a vector (see arrow F701 in FIG. 7A) indicating the traveling direction of the vehicle (step S702). If the subject position has reached the target position, the vector of the previous grid is used.
[0102] Next, the map division unit 306 acquires the forward ranging area and the rear ranging area (see FIGS. 7 and 8) which are the calculation target areas of the obstacle occupancy rate corresponding to the vector indicating the traveling direction of the vehicle (step S703).
[0103] Next, the map division unit 306 calculates a score indicating the obstacle occupancy rate of each of the forward ranging area and the rear ranging area (step S704).
[0104] Next, the map division unit 306 determines whether the subject position has reached the target position (step S705).
[0105] In the process of step S705, if the map division unit 306 determines that the subject position has not reached the target position (NO determination in step S705), the process returns to the process of step S701, and the processes of steps S701 to S705 are repeatedly executed.
[0106] On the other hand, in the process of step S705, if the map division unit 306 determines that the subject position has reached the target position (YES determination in step S705), the calculation process of the obstacle occupancy rate ends.
[0107] 17 is a flowchart showing the procedure of the map division process in the map division system according to this embodiment. The map division process is called and executed as a subroutine in step S80 of the map division area calculation process shown in FIG.
[0108] First, the map division unit 306 acquires and updates a target position (step S801). This process is similar to the process of step S701 described with reference to FIG. 16, and therefore a duplicated description will be omitted.
[0109] Next, the map division unit 306 expands the divided map (see FIG. 9) based on the global route size and the distance measurement range of the vehicle-mounted sensor (step S802).
[0110] Next, the map division unit 306 determines whether or not the current divided map has reached the maximum size of the reference map of the in-vehicle controller 220 (step S803).
[0111] In the process of step S803, if the map division unit 306 determines that the current divided map has not reached the maximum reference map size (NO determination in step S803), the process returns to the process of step S801, and the processes of steps S801 to S803 are repeatedly executed.
[0112] On the other hand, in the process of step S803, if the map division unit 306 determines that the current divided map has reached the maximum reference map size (YES determination in step S803), it calculates the cut-out position of the divided map and saves the area information of the divided map (step S804). In this process, the map division unit 306 executes map cut-out in the first pattern described in Fig. 10 or the second pattern described in Fig. 11 according to the ranging range of the in-vehicle sensor 211 when the obstacle sparse area is detected.
[0113] Next, the map division unit 306 determines whether the subject position has reached the target position (step S805).
[0114] In the process of step S805, if the map division unit 306 determines that the subject position has not reached the target position (NO determination in step S805), the process returns to the process of step S801, and the processes of steps S801 to S805 are repeatedly executed.
[0115] On the other hand, in the process of step S805, if the map division unit 306 determines that the subject position has reached the target position (YES determination in step S805), the map division process ends.
[0116] In the above description, the map division area calculation process is performed when the global route planning unit 305 plans one global route, but the present invention is not limited to this. For example, when the global route planning unit 305 plans multiple global routes, the map division unit 306 performs map division area calculation process along each of the multiple global routes to divide the entire map.
[0117] [effect] As described above, the map division system (map update system) according to the present embodiment divides the entire map based on the maximum size of the reference map of the in-vehicle controller 220 and the obstacle occupancy rate within the measurement range of the in-vehicle sensor 211. In addition, the map division system (map update system) transmits only the divided map of the current area and the divided map of the next area as reference maps to the in-vehicle controller 220 of the vehicle 101 according to the traveling position of the vehicle 101. The in-vehicle controller 220 stores only the divided map of the current area and the divided map of the next area as reference maps. Therefore, the map division system (map update system) according to the present embodiment keeps the reference map stored in the in-vehicle controller to a size equal to or smaller than the maximum size of the map that can be referred to, and the vehicle does not decelerate or stop near the boundary of the divided map, does not accelerate or decelerate excessively due to entering an area with few obstacles, does not deviate excessively from the route, and so on. Therefore, the map division system (map update system) according to the present embodiment reduces the calculation load of the in-vehicle controller and ensures the stability of vehicle travel when switching the divided map.
[0118] <Second embodiment> In the above first embodiment, a process (step S804 in FIG. 17, see FIG. 10) was described in which, when an obstacle sparsely populated area is detected within the rear ranging range of the on-board sensor 211 of the vehicle 101, a cut-out position of the divided map is determined based on a cost indicating an obstacle occupancy rate. However, in this process, for example, when a wide-area obstacle sparsely populated area occurs, that is, when an obstacle sparsely populated area occurs continuously over a wide area, it may not be possible to determine a cut-out position of the divided map. A map division system according to a second embodiment of the present invention has been made to solve the above problem.
[0119] A map division system according to a second embodiment of the present invention will be described below. In this embodiment, when an obstacle sparse area is included together with the forward ranging range and the rear ranging range of the on-board sensor 211 of the vehicle 101, it is treated as a wide-area obstacle sparse area occurring. In other words, in the wide-area obstacle sparse area, there is no information (feature amount information) about obstacles that can estimate the vehicle's location within the ranging range of the on-board sensor 211.
[0120] FIG. 18 is a block diagram showing an example of a functional configuration of an autonomous driving system 400 including a map division system (map update system) according to this embodiment. The in-vehicle controller 220 and the vehicle 101 of the autonomous driving system 400 are similar to the in-vehicle controller 220 and the vehicle 101 of the autonomous driving system 300 shown in FIG. 2, and therefore a duplicated description will be omitted. The map distribution server 401 (map division system and map update system) shown in FIG. 18 further includes a stable driving impossibility determination unit 402 in addition to the components of the map distribution server 301 shown in FIG. 2. The input side of the stable driving impossibility determination unit 402 is connected to the map division unit 306, and the output side is connected to the global route planning unit 305. Note that the components and functions of the map distribution server 401 other than the stable driving impossibility determination unit 402 are similar to the components and functions of the map distribution server 401 shown in FIG. 2, and therefore a duplicated description will be omitted.
[0121] The stable driving impossibility determination unit 402 detects that there is no feature amount information capable of estimating the vehicle's location within the distance measurement range of the vehicle-mounted sensor 211 of the vehicle 101, and outputs a stable driving countermeasure to the global route planning unit 305. Specific processing will be described with reference to Figs. 19 and 20.
[0122] [Map division area calculation processing] Here, the procedure of the map division area calculation process in the map division system according to this embodiment will be described. FIG. 19 is a flowchart showing the procedure of the map division area calculation process in the map division system according to this embodiment. As can be seen by comparing FIG. 19 with FIG. 15, in the map division area calculation process in the map division system according to this embodiment, a process of determining whether or not a wide-area obstacle sparse area has occurred (step S72) and outputting a stable driving measure (step S73) are added between the obstacle occupancy rate calculation process and the map division process. The obstacle occupancy rate calculation process (step S71) will be described in detail later in FIG. 20. Also, the map division calculation process (step S80) is the same as the process described in FIG. 15, so a duplicated description will be omitted.
[0123] In the process of step S72, if the stable driving impossibility determination unit 402 receives the occurrence location of a wide-area obstacle-sparsely populated area from the map division unit 306, it determines that a wide-area obstacle-sparsely populated area has occurred, and the process of step S72 is judged as YES. Also, if the stable driving impossibility determination unit 402 does not receive the occurrence location of a wide-area obstacle-sparsely populated area from the map division unit 306, it determines that a wide-area obstacle-sparsely populated area has not occurred, and the process of step S72 is judged as NO.
[0124] In the process of step S72, if the stable traveling impossibility determination unit 402 determines that a wide-area depopulated obstacle area has not occurred (NO determination in step S72), the process of step S80 is executed.
[0125] On the other hand, in the process of step S72, if the stable driving impossibility determination unit 402 determines that a wide-area sparsely populated obstacle area has occurred (YES determination in step S72), it outputs stable driving measures to the global route planning unit 305 (step S73). In the process of step S73, for example, low-speed driving measures, global route change measures, and obstacle installation measures are used as the stable driving measures.
[0126] [Measures for low speed driving] The low-speed travel countermeasure is a countermeasure in which the vehicle 101 changes the travel speed on the global route to a low speed. When a wide-area sparsely populated obstacle area occurs, the self-position estimation unit 312 estimates the self-position only from the wheel rotation count (odometry) acquired from the encoder 209, since there is no (valid) feature amount information with which the self-position can be estimated. In general, self-position estimation using odometry tends to have larger errors as the vehicle travels faster. The low-speed travel countermeasure changes the travel speed on the global route so that the vehicle 101 travels at a low speed (for example, 0.3 m / s) from a few meters before the occurrence of the wide-area sparsely populated obstacle area.
[0127] This measure requires the least number of steps, but it causes a problem of reduced work efficiency (transport efficiency) of the vehicle. In addition, for example, when wide-area sparsely populated obstacle areas occur continuously, the accuracy of the self-location estimation by the self-location estimation unit 312 deteriorates even if the vehicle travels at a low speed. For this reason, this measure should be considered as a temporary measure or a candidate measure in consideration of practical operation.
[0128] [Overall route change measures] The global route change measure is a measure to change the global route where a wide-area obstacle sparse area occurs to another global route. In this measure, the global route planning unit 305 plans and generates multiple global route candidates, taking into account the movement of other vehicles and other workers. When a wide-area obstacle sparse area occurs, the stable driving impossibility determination unit 402 outputs the global route change measure to the global route planning unit 305. Then, the global route planning unit 305 selects a global route candidate (detour route) that can avoid the occurrence of the wide-area obstacle sparse area from the multiple global route candidates, and changes the current global route. In addition, the global route planning unit 305 transmits the changed global route to the local route planning unit 313 of the in-vehicle controller 220.
[0129] This measure uses a detour route, which causes a problem of reduced vehicle work efficiency (transport efficiency). For this reason, like the slow-speed driving measure, this measure should be considered a temporary measure or a candidate measure.
[0130] [Measures to prevent obstacles] The obstacle installation countermeasure is a countermeasure to resolve the obstacle sparsely populated area by installing an obstacle near the occurrence point of the wide-area obstacle sparsely populated area. In this countermeasure, when a wide-area obstacle sparsely populated area occurs, the stable driving impossibility determination unit 402 outputs information on the occurrence point of the wide-area obstacle sparsely populated area and a notice to install an obstacle to the global route planning unit 305. Then, the global route planning unit 305 displays a notice to install an obstacle at the occurrence point of the wide-area obstacle sparsely populated area on a display device of the map distribution server 401. Note that the global route planning unit 305 may transmit a notice to install an obstacle at the occurrence point of the wide-area obstacle sparsely populated area to a terminal device held by a worker at the site (driving environment). The manager of the map distribution server 401 or a worker at the site checks the notification to install the obstacle and installs the obstacle, thereby resolving the problem caused by the obstacle sparsely populated area.
[0131] In addition, the obstacles are, for example, new shelves 111 and reflectors that allow LiDAR to measure distances stably even at long distances. After the obstacles are installed, it is necessary to recreate the map of the entire area and replan the global route. This measure requires more man-hours than the low-speed driving measure and the global route change measure, but it is a measure that can solve the root cause of the wide-area sparsely populated area.
[0132] In addition, the stable driving impossibility judgment unit 402 may refer to map information and global routes near the occurrence location of the wide-area sparsely populated obstacle area to judge whether or not there is continuous occurrence of wide-area sparsely populated obstacle areas, whether or not the transport efficiency is excessively reduced due to the global route candidate (detour route), etc., and select and execute one of the above three countermeasures.
[0133] Next, a procedure for calculating an obstacle occupancy rate in the map division system according to this embodiment will be described. Fig. 20 is a flowchart showing a procedure for calculating an obstacle occupancy rate in the map division system according to this embodiment. The processes other than steps S711 to S712 shown in Fig. 20 are similar to those shown in Fig. 16, so duplicated explanations will be omitted.
[0134] In the process of step S711, the map division unit 306 judges whether or not a wide-area obstacle-sparsely populated area has occurred. In this process, when the map division unit 306 judges that an obstacle-sparsely populated area is included together with the forward ranging area and the rear ranging area corresponding to the target position based on the score indicating the obstacle occupancy rate (see FIG. 8), it judges that a wide-area obstacle-sparsely populated area has occurred, and the process of step S711 is judged as YES. In addition, when the map division unit 306 judges that an obstacle-sparsely populated area is not included together with the forward ranging area and the rear ranging area corresponding to the target position, it judges that a wide-area obstacle-sparsely populated area has not occurred, and the process of step S711 is judged as NO.
[0135] In the process of step S711, if the map division unit 306 determines that a wide-area sparsely populated obstacle area has not occurred (NO determination in step S711), the process of step S705 is executed.
[0136] On the other hand, in the process of step S711, if the map division unit 306 determines that a wide-area obstacle-sparsely populated area has occurred (YES determination in step S711), it stores the location where the wide-area obstacle-sparsely populated area has occurred (step S712). Also, in this process, the map division unit 306 outputs the location where the wide-area obstacle-sparsely populated area has occurred to the stable driving impossibility determination unit 40. Note that the location where the wide-area obstacle-sparsely populated area has occurred includes the coordinates of the target position (grid) when it is determined that the wide-area obstacle-sparsely populated area has occurred and the ranging range of the in-vehicle sensor 211.
[0137] [effect] As described above, the map division system according to this embodiment detects the occurrence of a wide-area sparsely populated obstacle area and takes measures to ensure stable driving of the vehicle. Therefore, the map division system according to this embodiment has the same effect as the map division system according to the first embodiment, and can detect the occurrence of a wide-area sparsely populated obstacle area and prevent situations where stable driving is difficult.
[0138] It should be noted that the present invention is not limited to the above-described embodiments, and various other applications and modifications are possible without departing from the gist of the present invention as set forth in the claims. For example, the above-described embodiments have described in detail and specifically the configurations of the autonomous driving system and the map division system (map update system) in order to easily explain the present invention, and are not necessarily limited to those including all of the configurations described. In addition, it is possible to replace part of the configurations of the embodiments described here with the configurations of other embodiments, and it is also possible to add the configurations of other embodiments to the configurations of one embodiment. In addition, it is also possible to add, delete, or replace part of the configurations of each embodiment with other configurations. In addition, the control lines and information lines shown are those that are considered necessary for the explanation, and not all control lines and information lines in the product are necessarily shown. In reality, it can be considered that almost all components are connected to each other.
[0139] In each of the above embodiments, the feature amount information of the whole area map is information about obstacles in the whole area map, and the map division unit 306 calculates a cost indicating the occupancy rate of the obstacle based on the information about the obstacle, and performs the map division area calculation process. However, the present invention is not limited to this, and the feature amount information in the whole area map may include information about the feature amount on the map of the driving environment created by other methods. For example, the whole area map may be a heat map (represented by a grid map) created using the radio wave intensity of wireless communication. In this case, the feature amount information of the whole area map is information about the radio wave intensity on the heat map. Also, for example, when a grid map created using measurement information of LiDAR and a heat map created using the radio wave intensity of wireless communication are used as the whole area map, the feature amount information of the whole area map includes information about obstacles and information about radio wave intensity. That is, the feature amount information of the whole area map includes at least one of the multiple types of feature amount information held by the whole area map.
[0140] In the above embodiments, the map division unit 306 divides the entire area map by determining the overlapping area of adjacent divided maps according to the target speed and maximum speed calculated by the global route planning unit 305. However, the present invention is not limited to this. For example, when a speed limit or entry prohibition is set for a specific area due to traffic regulations, the map division unit 306 divides the entire area map based on the preset traffic rules even for the same vehicle and the same global route. Also, for example, when a target speed for the vehicle to travel is preset to a predetermined speed by an operator, the map division unit 306 calculates the overlapping area between the divided map of the next area and the divided map of the current area according to the preset target speed, and divides the entire area map. [Explanation of symbols]
[0141] 101...vehicle, 208...drive unit, 209...encoder, 211...vehicle sensor, 220, 220...vehicle controller, 300, 400...autonomous driving system, 301, 401...map distribution server, 302...global map management unit, 303...vehicle information management unit, 304...target position input unit, 305...global route planning unit, 306...map division unit, 307...map distribution unit, 311...reference map management unit, 312...self-position estimation unit, 313...local route planning unit, 314...control command generation unit, 402...stable driving impossibility determination unit
Claims
1. a global route planning unit that plans a global route from a vehicle travel start position to a target position by referring to a global map of a travel environment that has been created in advance; a map division unit that divides the entire map into a plurality of divided maps along the entire route so as to include a part of the entire route, based on vehicle information of the vehicle and feature amount information of the entire map, so that adjacent divided maps partially overlap each other; a map update unit that transmits the divided map to the vehicle in accordance with the location of the vehicle and updates a reference map of the vehicle. Map division system.
2. The vehicle information includes performance information of an in-vehicle sensor and performance information of an in-vehicle controller of the vehicle. The map division system according to claim 1 .
3. The feature amount information of the entire area map includes at least one of a plurality of types of feature amount information held by the entire area map. The map division system according to claim 2 .
4. A stable driving impossibility determination unit that detects that the feature amount information that can estimate the location of the vehicle does not exist within the distance measurement range of the vehicle-mounted sensor. The map division system according to claim 2 .
5. When the stable driving impossibility determination unit detects that the feature amount information from which the vehicle's location can be estimated does not exist within the distance measurement range of the in-vehicle sensor, the stable driving impossibility determination unit outputs information about a location where it was detected that the feature amount information from which the vehicle's location can be estimated does not exist and a notification that an obstacle will be installed. The map division system according to claim 4.
6. When the global route planning unit plans a plurality of the global routes, the map division unit divides the global map along each of the plurality of the global routes. The map division system according to claim 1 .
7. The map division unit divides the entire area map based on a preset traffic rule. The map dividing system according to any one of claims 1 to 6.
8. a global route planning unit that plans a global route from a vehicle travel start position to a target position by referring to a global map of a travel environment that has been created in advance; a map division unit that divides the entire map into a plurality of divided maps along the entire route so as to include a part of the entire route, based on vehicle information of the vehicle and feature amount information of the entire map, so that adjacent divided maps partially overlap each other; a map update unit that transmits the divided map to the vehicle in accordance with the location of the vehicle and updates a reference map of the vehicle. Map update system.
9. A step of planning a global route from a travel start position of the vehicle to a target position by referring to a global map of a travel environment created in advance; a step of dividing the entire area map into a plurality of divided maps along the entire area route so as to include a part of the entire area route, based on vehicle information of the vehicle and feature amount information of the entire area map, so that adjacent divided maps partially overlap each other; transmitting the divided map to the vehicle according to the location of the vehicle, and updating a reference map of the vehicle. Map division method.
Citation Information
Patent Citations
Map update system and map update program
JP2021196488A
