Electronic control device and parking support method

Abstract

An electronic control device for supporting parking of a vehicle in a parking lot includes: an information acquisition unit for acquiring detection information in which a sensor mounted on the vehicle detects the periphery of the vehicle; a parking section extraction unit for extracting regions of a plurality of parking sections in the periphery of the vehicle; a parking section occlusion space estimation unit for estimating occlusion space information including the height direction in the region of the parking section on the basis of the region of the parking section and the detection information; and a parking section vacancy likelihood estimation unit for calculating a vacancy likelihood indicating a possibility that a vehicle is not present in the parking section on the basis of the detection information and the occlusion space information in the region of the parking section.

Description

Electronic control device and parking support method Incorporation by reference

[0001] This application claims the priority of Japanese Patent Application No. 2025-005558, which was filed on January 15, 2025 (the year 2025 of Reiwa), and incorporates the content thereof by reference into this application.

[0002] The present invention relates to an electronic control device.

[0003] In recent years, for the purpose of realizing a parking support system that searches for available parking spaces in a parking lot and provides information to the driver, or automatically parks based on the search results, a technique for estimating the availability of parking spaces hidden behind obstacles has been proposed.

[0004] As the background art in this technical field, there is the following prior art. In Patent Document 1 (International Publication No. 2023 / 7785), depending on the ratio of the occlusion area in the parking section area, it is determined whether it is a parkable area or an area with a possibility of being available, and different identification display processes are performed according to the determination result . It has a parking area analysis unit that analyzes whether a vehicle can park or not in units of section areas by analyzing a camera captured image, and a display control unit that generates parking availability identification graphic data in units of section areas based on the analysis result and superimposes and displays it on the camera captured image. The parking area analysis unit calculates the ratio of the occlusion area to the total area of the section area for a section area where no parked vehicle is detected in the camera captured image, and determines whether the section area is a parkable area or an area with a possibility of being available according to the value of the calculated ratio. The display control unit describes an information processing device that superimposes and displays different graphic data in each area.

[0005] In the invention described in Patent Document 1, the occlusion region, which is an area that cannot be seen in the camera's captured image, is identified two-dimensionally in a bird's-eye view, and the value obtained by subtracting the area ratio of the occlusion region in the parking space from 1 (i.e., the area ratio of the visible area) is defined as the likelihood of the parking space being available. However, when a person tries to estimate the availability of a parking space that is hidden from view by an obstacle, they often make a three-dimensional judgment, including the height direction, not just the visible area. For example, if another vehicle is parked in a parking space that is hidden in the shadow of a low-profile vehicle such as a sports car, the likelihood of the parking space being available can be determined by whether the upper part of the other vehicle is visible behind the parked vehicle. Conversely, it is difficult to determine the likelihood of the parking space being available in a parking space that is hidden in the shadow of a tall vehicle such as a minivan. When making a judgment based on the two-dimensional degree of occlusion of the parking space as described in Patent Document 1, it is not possible to distinguish this difference in the height direction, and even if a parking space is clearly available to a person, it is difficult to determine whether the parking space is available without approaching it and checking the space two-dimensionally. Furthermore, in the case of parking spaces arranged in two rows, if all the parking spaces in the front row are occupied, the area of ​​the parking spaces in the back row is completely obscured in a two-dimensional view. Therefore, the method described in Patent Document 1 cannot determine whether there are any available parking spaces.

[0006] Thus, the method described in Patent Document 1 has problems such as delays in determining the availability of parking spaces hidden behind obstacles, and limitations on the range of parking spaces that can be determined.

[0007] A representative example of the invention disclosed in this application is as follows: an electronic control device for assisting in the parking of a vehicle in a parking lot, comprising: an information acquisition unit that acquires detection information obtained when a sensor mounted on the vehicle detects the area around the vehicle; a parking space extraction unit that extracts the areas of a plurality of parking spaces around the vehicle; a parking space occlusion space estimation unit that estimates occlusion space information including the height direction in the area of ​​the parking space based on the area of ​​the parking space and the detection information; and a parking space occupancy likelihood estimation unit that calculates the occupancy likelihood indicating the possibility that no vehicle is present in the parking space based on the detection information and the occlusion space information in the area of ​​the parking space.

[0008] According to one aspect of the present invention, it is possible to determine the availability of parking spaces hidden behind obstacles. Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

[0009] This is a functional block diagram showing the configuration of a vehicle system including an electronic control device according to an embodiment of the present invention. This is a diagram showing an example of the installation of an external sensor group on a vehicle according to this embodiment. This is a diagram showing an example of parking map data from the parking map data group of this embodiment. This is a diagram showing the relationship between the parking map data group and the parking space data group according to this embodiment. This is a diagram showing the relationship between the parking map data group and the parking space data group according to this embodiment. This is a diagram showing an example of the occlusion space data group according to this embodiment. This is a diagram showing an example of the occlusion space data group according to this embodiment. This is a diagram showing an example of the occlusion space data group according to this embodiment. This is a diagram showing an example of the parking space space occupancy model data group according to this embodiment. This is a diagram showing an example of the data for the parking space space occupancy model according to this embodiment. This is a diagram showing an example of the data for the parking space space occupancy model according to this embodiment. This is a diagram showing an example of the parking space availability likelihood data group according to this embodiment. This is a functional block diagram of the electronic control device according to this embodiment. This is a diagram showing an example of a scene used to explain the operation of the electronic control device according to this embodiment. This is a diagram explaining the process executed by the occlusion space estimation unit using the flowchart of this embodiment. This is a diagram explaining the generation of occlusion space data according to this embodiment. This is a diagram explaining the generation of occlusion space data according to this embodiment. This is a diagram explaining the generation of occlusion space data according to this embodiment. This is a diagram explaining the generation of occlusion space data according to this embodiment. This is a diagram illustrating the generation of occlusion space data in this embodiment. This is a flowchart of the process performed by the parking space occupancy model determination unit in this embodiment. This is a flowchart of the process performed by the parking space availability likelihood estimation unit in this embodiment. This is a diagram illustrating the process of step S1409 in this embodiment. This is a flowchart of the process performed by the driving plan unit in this embodiment. This is a diagram showing an example of the state of the vehicle and the parking space availability likelihood data group in this embodiment. This is a diagram showing an example of screen display information generated in this embodiment.

[0010] <Embodiment 1> Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[0011] (System Configuration) Figure 1 is a functional block diagram showing the configuration of a vehicle system 1 including an electronic control device 3 according to an embodiment of the present invention.

[0012] Vehicle system 1 is mounted on vehicle 2. Vehicle system 1 understands the situation of obstacles such as parking spaces and surrounding vehicles around vehicle 2, and provides appropriate driving assistance and driving control, especially for parking in parking lots. As shown in Figure 1, vehicle system 1 consists of an electronic control unit 3, an external sensor group 4, a vehicle sensor group 5, a map information management device 6, an actuator group 7, an HMI device group 8, and an external communication device 9. The electronic control unit 3, the external sensor group 4, the vehicle sensor group 5, the map information management device 6, the actuator group 7, the HMI device group 8, and the external communication device 9 are connected by an in-vehicle network N. Hereafter, vehicle 2 may be referred to as "our vehicle" 2 to distinguish it from other vehicles.

[0013] The electronic control unit 3 is an ECU (Electronic Control Unit). Based on various input information provided by the external sensor group 4, the vehicle sensor group 5, the map information management device 6, etc., the electronic control unit 3 generates driving plan information for the vehicle 2's parking assistance system and outputs it to the actuator group 7 and the HMI device group 8. The electronic control unit 3 has a processing unit 10, a storage unit 30, and a communication unit 40.

[0014] The processing unit 10 is configured to include, for example, a central processing unit, which is a CPU (Central Processing Unit). However, it may also be configured to include a GPU (Graphics Processing Unit), FPGA (Field-Programmable Gate Array), ASIC (Application Specific Integrated Circuit), etc., in addition to the CPU, or it may be configured to include any one of them.

[0015] The processing unit 10 has the following functions: an information acquisition unit 11, a vehicle position and attitude estimation unit 12, a parking space extraction unit 13, an occlusion space estimation unit 14, a parking space occupancy model determination unit 15, a parking space availability likelihood estimation unit 16, a driving plan unit 17, and an information output unit 18. The processing unit 10 realizes these functions by executing a predetermined operation program stored in the storage unit 30.

[0016] The information acquisition unit 11 acquires various information from other devices connected to the electronic control unit 3 via the in-vehicle network N and stores it in the storage unit 30. For example, it stores data such as the external sensor data group 31, which includes spatial recognition information and object recognition information from the external sensor group 4; the vehicle sensor data group 32, which relates to the movement and state of the vehicle 2 detected by the vehicle sensor group 5, etc.; and the parking lot map data group 33, which relates to the roads the vehicle 2 travels on and is distributed by the map information management device 6.

[0017] The vehicle position and attitude estimation unit 12 determines the position and attitude of the vehicle 2 in the parking lot map data group 33 based on information such as the external sensor data group 31, the vehicle sensor data group 32, and the parking lot map data group 33 (see Figures 3 and 4) acquired by the information acquisition unit 11 and stored in the storage unit 30.

[0018] The parking space extraction unit 13 extracts parking spaces around the vehicle by referring to the parking map data group 33 based on the position and orientation identified by the vehicle position and orientation estimation unit 12, and stores the extracted parking spaces as a parking space data group 34 in the storage unit 30.

[0019] The occlusion space estimation unit 14 estimates the spatial areas that have not been confirmed by the external sensor group 4 in the parking spaces included in the parking space data group 34, based on the spatial recognition information of the external sensor group 4 included in the external sensor data group 31, and stores the estimated spatial areas as occlusion space data group 35 (see Figures 5B and 5C) in the storage unit 30.

[0020] The parking space occupancy model determination unit 15 generates a parking space occupancy model data group 36, described later, as needed, and stores the generated parking space occupancy model data group 36 in the storage unit 30. The parking space occupancy model data group 36 may be pre-recorded in the storage unit 30 at the time of product shipment, and the parking space occupancy model determination unit 15 may not exist if there is no need to change the parking space occupancy model. The parking space occupancy model determination unit 15 is executed when the parking space occupancy model data group 36 is changed according to the region of the vehicle. For example, the electronic control device 3 acquires a parking space occupancy model pre-generated by the center system 100 via the external communication device 9, and the parking space occupancy model determination unit 15 sets the parking space occupancy model. Alternatively, for example, the parking space occupancy model may be generated sequentially based on recognition information of other vehicles acquired from the external sensor group 4 of the vehicle 2.

[0021] The parking space vacancy likelihood estimation unit 16 calculates the vacancy likelihood of parking spaces included in the parking space data group 34. For parking spaces whose vacancy is uncertain due to obstruction, the unit generates a virtual space occupancy model based on the parking space space occupancy model data group 36 and estimates the vacancy likelihood of that parking space by comparing it with the corresponding occlusion space data group 35. The estimated vacancy likelihood is stored in the storage unit 30 as a parking space vacancy likelihood data group 37 (see Figure 8) in a format associated with the parking space data group 34.

[0022] The driving plan unit 17 determines a recommended target parking space for vehicle 2 to park based on the parking space data group 34, parking space availability likelihood data group 37, etc., and generates recommended driving route information and driving control plan information for driving the recommended driving route. In manual driving mode, the driving plan unit 17 formats the recommended driving route into information to be presented to the occupant on the HMI device group 8, and in automatic driving mode, it formats the information to be presented to the occupant as well as information for controlling the actuator group 7. The driving plan unit 17 then stores the formatted information as driving plan data group 38 in the storage unit 30.

[0023] The information output unit 18 outputs various information to other devices connected to the electronic control unit 3 via the in-vehicle network N. For example, the information output unit 18 outputs control information included in the driving plan data group 38 determined by the driving plan unit 17 to the actuator group 7 to control the driving of the vehicle 2. Also, for example, the information output unit 18 outputs information about parking spaces that are likely to be available, including their availability likelihood, included in the parking space availability likelihood data group 37, as well as recommended route information included in the driving plan data group 38, to the HMI device group 8 to present to the occupants and support their parking actions.

[0024] The storage unit 30 is configured to include, for example, storage devices such as HDDs (Hard Disk Drives), flash memory, and ROMs (Read Only Memory), as well as non-volatile storage media such as RAMs. The storage unit 30 stores programs processed by the processing unit 10 and data sets necessary for such processing. It may also be used as the main memory when the processing unit 10 executes a program, temporarily storing data necessary for program calculations. In this embodiment, the storage unit 30 stores information for realizing the functions of the electronic control device 3, such as an external sensor data group 31, a vehicle sensor data group 32, a parking lot map data group 33, a parking space data group 34, an occlusion space data group 35, a parking space occupancy model data group 36, a parking space availability likelihood data group 37, and a driving plan data group 38.

[0025] The external sensor data group 31 is a collection of data relating to spatial recognition information and object recognition information from the external sensor group 4. Spatial recognition information is information that represents the characteristics of the space around the vehicle 2 based on observation information observed by the external sensor group 4, such as three-dimensional point cloud information observed by LiDAR or multi-cameras. Alternatively, a voxel map that integrates point cloud information chronologically to improve the reliability of detection information or an occupancy grid map that includes height information may also be used. Object recognition information is information about environmental elements understood from the observation information of the external sensor group 4, such as vehicles, parking spaces, and pedestrians. Object recognition information includes, for example, the type, position, speed, and orientation of the object in question.

[0026] The vehicle sensor data group 32 is a collection of data relating to the movement and state of the vehicle 2, and includes vehicle information detected by the vehicle sensor group 5 and other devices and acquired by the information acquisition unit 11. The vehicle information includes, for example, information such as the position of the vehicle 2, driving speed, steering angle, accelerator operation amount, and brake operation amount.

[0027] The parking map data group 33 is a collection of data about roads and parking spaces within the parking lot surrounding the vehicle 2, acquired from the map information management device 6, etc. The parking space data group 34 is a collection of data about parking spaces around the vehicle 2. The occlusion space data group 35 is a collection of data about spatial areas in the parking space around the vehicle 2 that are obscured by obstacles and cannot be detected by the external sensor group 4. The parking space space occupancy model data group 36 is a collection of data that models the probability of each location in the spatial area of ​​the parking space being occupied by a parked vehicle, assuming the presence of a parked vehicle. The parking space vacancy likelihood data group 37 is a collection of data about the vacancy likelihood of each parking space around the vehicle 2. The driving plan data group 38 is a collection of data about recommended driving route information generated by the driving plan unit 17 and driving control plan information for driving along that driving route.

[0028] The communication unit 40 is configured to include, for example, a network card compliant with communication standards such as IEEE 802.3 or CAN (Controller Area Network), and transmits and receives data with other devices of the vehicle system 1 according to various protocols.

[0029] In this embodiment, the communication unit 40 and the processing unit 10 are described separately, but some of the processing of the communication unit 40 may be executed within the processing unit 10. For example, hardware devices equivalent to those in communication processing may be provided in the communication unit 40, while other device drivers and communication protocol processing may be provided in the processing unit 10.

[0030] The external sensor group 4 is a collection of devices capable of detecting the surrounding conditions of the vehicle 2 in three dimensions. The external sensor group 4 consists of, for example, sensor devices such as multi-cameras and LiDAR, and a processing device for spatial recognition and object recognition around the vehicle 2 using the observed information, and outputs the recognized spatial and object information to the network N.

[0031] The vehicle sensor group 5 is a collection of devices that detect various states of the vehicle 2. Each vehicle sensor detects, for example, the position and attitude of the vehicle 2, its speed, steering angle, accelerator pedal operation, brake pedal operation, etc., and outputs these to the in-vehicle network N. The position and attitude of the vehicle 2 is, for example, a positioning result based on radio waves received from satellites of the Global Navigation Satellite System (GNSS), and is information regarding its position and orientation in a global coordinate system.

[0032] The map information management device 6 is a device that manages and provides digital map information of the area around the vehicle 2, and has digital road map data and parking map data for a predetermined area including the area around the vehicle 2. Based on the location information of the vehicle 2 output from the vehicle sensor group 5, the map information management device 6 identifies the current position of the vehicle 2 on the map, that is, the road and lane the vehicle 2 is traveling on, and outputs the identified current position of the vehicle 2 and the map data of its surroundings to the in-vehicle network N. The map information management device 6 may dynamically acquire parking maps within a predetermined distance from the current position from the center system 100 via the external communication device 9, or it may acquire parking maps that were generated and saved when the vehicle 2 traveled in the past.

[0033] The actuator group 7 is a group of devices that control control elements such as steering, brakes, and accelerators that determine the movement of the vehicle. The actuator group 7 controls the movement of the vehicle based on operation information from the occupants, such as the steering wheel, brake pedal, and accelerator pedal, and control information output from the electronic control unit 3.

[0034] The HMI device group 8 is a group of devices for receiving information input from the occupant to the vehicle system 1 and for notifying the occupant of information from the vehicle system 1. The HMI device group 8 includes a display, speaker, vibrator, switch, etc.

[0035] The external communication device 9 is a communication module that communicates wirelessly with the outside of the vehicle system 1, and can communicate with, for example, a center system 100 that distributes data and provides services to the vehicle system 1, or with the Internet.

[0036] The center system 100 is a server that can communicate with the vehicle 2 via an external communication device 9, and collects data from the vehicle 2 and distributes the necessary data to the vehicle 2. For example, the center system 100 may distribute the latest information on digital road map data and parking lot map data to the map information management device 6, or distribute the requested parking space occupancy model for the parking lot to the vehicle 2.

[0037] Figure 2 shows an example of the installation of the external sensor group 4 on the vehicle 2.

[0038] In the example shown in Figure 2, the detection area 111-115 is formed by the external sensor group 4, enabling spatial recognition and object recognition of the entire surrounding area of ​​the vehicle 2. The external sensor group 4 may be implemented with a single sensor device, such as LiDAR, or with a combination of multiple sensor devices, such as a multi-camera system. In this embodiment, it is not necessarily required to recognize the entire surrounding area; it is sufficient to recognize a predetermined range around the vehicle 2.

[0039] Figure 3 shows an example of parking lot map data from the parking lot map data set 33.

[0040] The parking map data 200 stored in the parking map data group 33 visualizes and represents the constituent elements included in the data relating to a particular parking lot. The parking map data consists, for example, of road information relating to the roads used to travel through the parking lot and parking space information relating to the parking spaces located within the parking lot.

[0041] Route information can, for example, follow the structure of general digital map data. Specifically, it consists of "links" that indicate partial routes and "nodes" that are the connection points between links.

[0042] A link includes information such as identification information for uniquely identifying the link, a sequence of shape points representing the track shape formed by the link, a track direction indicating the possible traveling direction, identification information of the nodes at both ends, and a reference to the related parking section information. For example, link 213 includes information where the identification information is 213, the sequence of shape points is the position coordinates of nodes 223 and 224, the traveling direction is the direction 233 from node 223 to node 224, and the identification information of the nodes at both ends is 223 and 224. When the sequence of shape points is a straight line, it is represented by the position coordinates of two points of the nodes at both ends. However, when it is curved, it needs to include the position coordinates of sufficient shape interpolation points for representing the shape of the link. The related parking section information is information on the parking sections facing the partial track represented by the link. For example, the parking section information related to link 216 is the parking sections in the upper half row of the parking section group 201, and the parking section information related to link 215 is the parking sections in the lower half row of the parking section group 201 and the upper half row of the parking section group 202. By referring to the related parking section information, the partial track for parking in a specific parking section can be identified.

[0043] A node includes information such as identification information for uniquely identifying the node, the position coordinates of the node, and a list of identification information of the links connected to the node. For example, the identification information of the links connected to node 223 is 212, 213, and 215. By tracing the identification information of the nodes included in the link and the identification information of the links included in the node, the connection relationship of the partial track can be calculated.

[0044] The parking section information is configured to be able to calculate the position, orientation, width, and length of each parking section. For example, the position of the parking section is the position coordinates of the center point of the rectangle of the parking section, and the orientation of the parking section is the azimuth of the reference direction of the parking section (e.g., the direction of a vehicle parked forward). The width of the parking section is the length of the section corresponding to the lateral direction of the vehicle, and the length of the parking section is the length of the section corresponding to the longitudinal direction of the vehicle. Also, each parking section is assigned identification information for uniquely identifying the parking section.

[0045] Note that the position and orientation in the parking lot map data may be represented in a global coordinate system or a coordinate system that can be mutually converted with the global coordinate system.

[0046] FIGS. 4A and 4B are diagrams showing the relationship between the parking lot map data group 33 and the parking section data group 34.

[0047] The parking section information (parking lot map data 200) of the parking lot map data group 33 shown in FIG. 4A is expressed in a coordinate system corresponding to the global coordinate system. On the other hand, the parking section data group 34 shown in FIG. 4B is parking section information expressed in a relative coordinate system centered on the vehicle 2.

[0048] In FIG. 4A, the position and orientation of the vehicle 2 in the parking lot map data 200 shown in FIG. 3 are represented. The position and orientation of the vehicle 2 are specified by the host vehicle position and orientation estimation unit 12. When the position and orientation of the vehicle 2 are specified, the positional relationship with respect to each parking section centered on the vehicle 2 can be calculated. The parking section data group 34 shown in FIG. 4B is obtained by coordinate-converting the position and orientation information of the parking section information included in the parking lot map data group 33. The parking section information (FIG. 4A) of the parking lot map data group 33 and the parking section information (FIG. 4B) of the parking section data group 34 are associated with each other by the identification information of the parking section.

[0049] FIGS. 5A, 5B, and 5C are diagrams showing an example of the occlusion space data group 35.

[0050] FIG. 5A is a view from above of an occlusion space region 421 formed by being shielded by another vehicle 401 in the external sensor group 4 of the vehicle 2. The occlusion space data group 35 is data representing a space region that is shielded by an occlusion object and cannot be detected by the external sensor group 4 in a predetermined parking section. In other words, the occlusion space data is obtained by cutting out the occlusion space region formed by the occlusion object from the space region part of a predetermined parking section.

[0051] Figure 5B shows an example of a data representation of occlusion space data 431. The occlusion space data 431 can be represented, for example, by dividing the target parking space into a grid and storing the height of the occlusion space at that location in each grid cell using a grid map. Furthermore, the occlusion space data does not necessarily have to represent the height of occlusion in all parking spaces; it may also represent only the height of the portion passing through the center of the parking space in the front-to-back direction of a vehicle parked in the parking space (i.e., the area enclosed by the thick frame in Figure 5B, along the x' axis which indicates the depth direction of the parking space). The depth direction in a parking space is the direction away from the road facing the parking space.

[0052] Figure 5C is a visual representation of the occlusion space data 431 in the parking space 411. The occlusion space data 431 is obtained by cutting out the occlusion space region 421 along the parking space 411 and represents the height of the occlusion space (obscured by the occluder) in each grid cell.

[0053] Figure 6 shows an example of a parking space occupancy model data set 36.

[0054] The parking space occupancy model data group 36 is composed of, for example, a region ID 701, a parking lot ID 702, and model data 703.

[0055] The regional ID 701 is information for identifying a region, and may be used to identify a region at the level of a country or a region containing multiple countries, or it may be used to identify a region at the level of a city or town. The parking lot ID 702 is identification information for uniquely identifying a parking lot. The regional ID 701 and parking lot ID 702 are included in the map information managed by the map information management device 6.

[0056] Model data 703 is model data for parking space occupancy in the region and parking lot specified by region ID 701 and parking lot ID 702. Details of the parking space occupancy model data will be described later with reference to Figure 7. In the parking space occupancy model data group 36 shown in Figure 6, the name of the model is recorded, but the model data itself may be stored, or information for referencing the corresponding model data (for example, path information to the model data file) may be stored.

[0057] If the value "Any" is stored for either Region ID 701 or Parking ID 702, it means that no specific region or parking lot is designated. When referencing parking space occupancy model data corresponding to a given parking lot, data with a matching Parking ID 702 is prioritized. If no data with a matching Parking ID 702 is found, data with a matching Parking ID 702 and Region ID 701 is prioritized. If neither Parking ID 702 nor Region ID 701 matches, data with both Region ID 701 and Parking ID 702 being "Any" is referenced.

[0058] Figures 7A and 7B show an example of data for a parking space occupancy model.

[0059] The parking space occupancy model represents the probability distribution of occupancy (the probability that each point in the spatial region is occupied by a parked vehicle) in a parking space, assuming that a parked vehicle is present in the parking space. Figures 7A and 7B show one example of this representation, where the parking space 601 is represented by the occupancy probability distribution 611 viewed in the x'-axis direction and the occupancy probability distribution 612 viewed in the y'-axis direction. The occupancy probability distribution is represented, for example, by the boundary lines (contour model) of regions where the probability exceeds a predetermined value. In the occupancy probability distributions 611 and 612 in Figures 7A and 7B, the boundary lines for 95%, 80%, 50%, and 20% are represented, respectively. The region enclosed by the 95% boundary line means that, if a parked vehicle is present, there is a 95% or greater probability that the region is occupied by a parked vehicle.

[0060] Vehicles vary in size and shape, from small cars to large vehicles, and their parking methods also differ. However, even with variations in vehicles and parking, there are common tendencies in the occupancy of parking spaces. For example, the center of a vehicle is highly likely to be located in the center of a parking space, and the center of the vehicle is taller to ensure a larger passenger compartment. On the other hand, the front of the vehicle is often shorter. Even if the occlusion space height is the same at 1.8m, the probability of seeing the roof of a parked vehicle is higher in the center of the parking space, while the probability of not seeing the parked vehicle is higher in the front. These characteristics are modeled in parking space occupancy models.

[0061] Furthermore, it is desirable that the parking space occupancy model reflects regional characteristics such as the differences in the types of vehicles used and the way parking is done, depending on the region. For example, some regions prefer large vehicles, while others have a higher proportion of small vehicles. In regions where large vehicles are preferred, a model with wide boundary lines extending outwards, such as the occupancy probability distributions 621 and 622 in Figure 7B, is suitable. In regions where forward parking is more common than rear parking, a model inverted on the y' axis is suitable because the direction of the vehicles changes. When a region-specific model is constructed in this way, a model with the identification information of the target region set in region ID 701 is added to Figure 6.

[0062] Similarly, since some parking lots have vehicle restrictions, a parking space occupancy model that reflects these characteristics may be suitable. In that case, Figure 6 describes a model in which the identification information of the target parking lot is set in parking lot ID 702.

[0063] The occupancy probability distribution can be represented in various ways. For example, the spatial area of ​​the parking space can be divided into a three-dimensional grid, and the occupancy probability of each position can be stored in a three-dimensional grid map. Alternatively, the occupancy probability distributions 611 and 612 in Figure 7A can be represented using a two-dimensional grid map instead of a contour model. Furthermore, the model can be simplified to represent only the occupancy probability distribution along the x' axis (on the line where y'=0).

[0064] Thus, the parking space occupancy model is configured such that the likelihood of vacancy for a given height in the occlusion space is larger in the central region than in the outer region of the parking space. This allows for accurate reflection of the shape of the vehicle and enables more accurate estimation of the occlusion space than an occupancy model with uniform height.

[0065] Figure 8 shows an example of a group of parking space availability likelihood data 37 output by the parking space availability likelihood estimation unit 16.

[0066] The parking space availability likelihood data set 37 is data that numerically represents the probability that each parking space is available. For example, the parking space availability likelihood is expressed as a value between 0.0 and 1.0, where 0.0 means that the parking space is definitely not available (a parked vehicle is present or parking is prohibited), and 1.0 means that the parking space is definitely available. The parking space availability likelihood is managed in association with the parking space (preferably with the identification information assigned to the parking space). In Figure 8, it is expressed in association with the parking space information of the parking map data set 33, but if it is managed in association with the identification information, it can be easily associated with the parking space data set 34.

[0067] The operation of the vehicle system 1 will be explained with reference to Figures 9 to 19.

[0068] The electronic control unit 3 identifies the extent to which parking spaces around the vehicle 2 are visible spatially based on information acquired from the external sensor group 4, vehicle sensor group 5, map information management device 6, external communication device 9, etc., and estimates the probability (likelihood of availability) that the parking space is available. The electronic control unit 3 also generates driving plan information for controlling the vehicle 2's movement based on the estimated likelihood of availability of the parking space, and outputs the refined driving plan information to the actuator group 7. The electronic control unit 3 may also generate information to support the occupant's actions in parking the vehicle 2 and output it to the HMI device group 8. In this way, the vehicle system 1 assists in driving the vehicle 2 to park in the parking lot.

[0069] Figure 9 is a functional block diagram of the electronic control unit 3.

[0070] The electronic control unit 3 is configured such that the processing of the information acquisition unit 11, the vehicle position and attitude estimation unit 12, the parking space extraction unit 13, the occlusion space estimation unit 14, the parking space occupancy model determination unit 15, the parking space availability likelihood estimation unit 16, the driving plan unit 17, and the information output unit 18 in Figure 1 is executed in an appropriate order. The series of processes is repeatedly executed at predetermined time intervals (for example, every 100 milliseconds).

[0071] The information acquisition unit 11 acquires necessary information from other devices via the in-vehicle network N and stores the acquired information in the storage unit 30. For example, it acquires the external sensor data group 31 from the external sensor group 4, the vehicle sensor data group 32 from the vehicle sensor group 5, the parking lot map data group 33 from the map information management device 6, and the parking space occupancy model data group 36 from the external communication device 9, and passes them on to the subsequent processing unit.

[0072] The vehicle position and attitude estimation unit 12 estimates the position and attitude of vehicle 2 on the parking lot map based on data acquired from the external sensor data group 31, the vehicle sensor data group 32, and the parking lot map data group 33, and outputs the estimated position and attitude. For example, the position and attitude of vehicle 2 can be estimated by comparing the global coordinate system information included in the parking lot map data group 33 with the GNSS positioning information related to the vehicle's position and attitude included in the vehicle sensor data group 32. Alternatively, if the parking lot map data group 33 includes a three-dimensional point cloud map of the parking lot, the position and attitude of vehicle 2 may be estimated by comparing the three-dimensional point cloud information included in the external sensor data group 31 with the three-dimensional point cloud map. The three-dimensional point cloud map may be one that has been generated offline in advance and downloaded from the center system 100 to the map information management device 6 via the external communication device 9, or it may be one that has been generated using SLAM (Simultaneous Localization and Mapping) technology or the like when vehicle 2 has driven through the same parking lot, and stored in the map information management device 6 for use. Furthermore, the estimated position and orientation information should be output in association with the parking lot map data set 33.

[0073] The parking space extraction unit 13 calculates a group of parking space data 34 around the vehicle 2 based on the position and orientation information of the vehicle 2 included in the parking map data group 33 and the parking map data 200 shown in Figure 3, and outputs the calculated parking space data group 34 to the subsequent processing.

[0074] The occlusion space estimation unit 14 estimates the occlusion space within the parking space area based on the external sensor data group 31 and the parking space data group 34, and outputs it as occlusion space data group 35 to the parking space vacancy likelihood estimation unit 16.

[0075] The parking space occupancy model determination unit 15 determines the parking space occupancy model data to be used from the group of parking space occupancy model data 36 acquired from the center system 100 via the external communication device 9, based on information such as the parking space position and orientation information of the vehicle 2 identified by the vehicle position and orientation estimation unit 12.

[0076] The parking space vacancy likelihood estimation unit 16 estimates the vacancy likelihood of a parking space in the parking space data group 34 by matching the parking space space occupancy model data with the occlusion space data, and outputs the estimated vacancy likelihood in relation to the parking space data group 34.

[0077] The driving plan unit 17 generates driving plan information (control command values, etc.) for controlling the driving of the vehicle 2, or driving plan information (information on available parking spaces in the surrounding area, etc.) to be presented to the occupants, based on the vehicle sensor data group 32, parking space data group 34, parking space availability likelihood data group 37, etc., according to the driving mode of the vehicle 2. The driving plan unit 17 passes the generated driving plan information as a driving plan data group 38 to the information output unit 18 and instructs it to output to an appropriate external device.

[0078] The information output unit 18 outputs the driving plan information of the driving plan data group 38 to the actuator group 7 and the HMI device group 8 based on the content of the instructions from the driving plan unit 17.

[0079] Figure 10 shows an example of a scene used to explain the operation of the electronic control unit 3. It is assumed that the likelihood of vacancy in the areas of parking spaces 901 and 902 is estimated from the vehicle 2.

[0080] (Processing of the occlusion space estimation unit 14) The process performed by the occlusion space estimation unit 14 will be explained using the flowchart in Figure 11.

[0081] First, in steps S1001 and S1002, the occlusion space estimation unit 14 acquires spatial recognition information included in the external sensor data group 31 and the parking space data group 34.

[0082] Next, in step S1003, the occlusion space estimation unit 14 extracts parking spaces for which no detection data exists in the spatial recognition information as candidates for available parking spaces. The detection data is, for example, three-dimensional point cloud data, and it checks whether a detection point cloud of obstacles exists in the area of ​​each parking space in the parking space data group 34. If a detection point cloud exists in a parking space, it means that an obstacle exists in the area of ​​that parking space, and therefore it can be determined that the parking space is not available. On the other hand, parking spaces for which no detection point cloud exists may be available, but they may be obscured by another obstacle and the external sensor group 4 has not been able to detect the obstacle. For this reason, parking spaces for which no detection point cloud exists are extracted as "candidates" for available parking spaces and are used to estimate the likelihood of availability in subsequent processing.

[0083] Steps S1004 to S1010 are processes performed for each of the available parking space candidates extracted in step S1003.

[0084] In step S1004, one unprocessed vacant parking space candidate VPS is selected.

[0085] If no selection target exists (N in S1005), this process is terminated. On the other hand, if a selection target exists (Y in S1005), the process proceeds to S1006, and an external sensor S that includes the area of ​​the candidate vacant parking space VPS within its detectable range is identified. The detectable range is the area that each external sensor is specified to be detectable in terms of its specifications, as shown in 111 to 115 of Figure 2, and is predetermined as a design value based on the sensor's mounting position and observation range (field of view).

[0086] If there are no external sensors S that include the area of ​​the candidate vacant parking space VPS within their detectable range (N in S1007), the occlusion space data OS (VPS) of the selected candidate vacant parking space VPS is set to OS_MAX (S1011). OS_MAX is data that means that the entire target space of the parking space is an occlusion space (a space that cannot be detected by external sensors). If there are external sensors S (Y in S1007), detection data for the front and back sides of the area of ​​the candidate vacant parking space VPS is extracted from the spatial recognition information derived from the external sensors S (S1008), and occlusion space data OS (VPS) is generated based on the extracted detection data using the method described later (S1009). Then, regardless of whether there are external sensors S that include the area of ​​the candidate vacant parking space VPS within their detectable range, the occlusion space data OS (VPS) is stored in the storage unit 30 as part of the occlusion space data group 35, and the process returns to step S1004.

[0087] The details of the processing of the occlusion space estimation unit 14 will be explained with reference to Figures 12A, 12B, 12C, 13A, and 13B.

[0088] Figures 12, 12B, and 12C illustrate the generation of occlusion space data OS(902) for parking space 902.

[0089] Parking space 902 is partially obscured by another vehicle 931. The triangles and circles in Figure 12A indicate a portion of the detection data in the spatial recognition information. Since no detection data exists for parking space 902, it is extracted as a candidate for an empty parking space (S1003 in Figure 11). On the other hand, the parking spaces where other vehicles 931 and 932 are located are not considered as candidates for empty parking spaces because detection data exists for each of them.

[0090] Since parking space 902 is included in the detectable range 111 of Figure 2, an external sensor targeting the front (e.g., a front camera) is identified as the external sensor S (S1006). Detection data from the front side (circles) and detection data from the back side (triangles) of parking space 902 are extracted from the spatial recognition information of the external sensor S (S1008), and occlusion spatial data is generated based on the positional relationship of the two detection data (S1009).

[0091] As shown in Figure 5, occlusion space data represents the relationship between the position within the parking space area and the height of the spatial area not visible to the external sensor (occlusion space). Here, we will explain how to determine the height of the occlusion space. For example, to determine the height of the occlusion space at the position of the star in Figure 12A, a line 1101 is drawn radially from the installation position of the external sensor that observes the detectable range 111, passing through the position of the star, and detection data is extracted from the front and back sides of the parking space 902 that are near this line. Among the extracted detection data, detection point 1111, which is at the position with the highest elevation angle on the front side, and detection point 1112, which is at the position with the lowest elevation angle on the back side, are extracted. The area with a lower elevation angle than detection point 1111 is occluded by an obstacle on the front side, and the area with a higher elevation angle than detection point 1112 is visible because of an obstacle on the back side. For this reason, the boundary of the occlusion space is approximated by a line connecting detection point 1111 and detection point 1112.

[0092] Figure 12B is a view of the vehicle 2 from the side, and Figure 12C is a view from the installation position of the external sensor that observes the detectable range 111 of the vehicle 2. The height of the occlusion space at the star-marked position is set to the height 1121 where it intersects with the line connecting detection point 1111 and detection point 1112. Similarly, the height of the occlusion space at each reference position of the parking space 902 is calculated, and the occlusion space data OS(902) shown in Figure 5 is generated.

[0093] Depending on the environment of the parking space, there may be no detection points on the near or far side. For example, if there are no obstacles on the near side, there will be no detection points on the near side. In that case, it means that there are no obstructing obstacles, so the height of the occlusion space will be 0. Also, if there are no detection points on the far side, it is not possible to determine how far back is visible, so the above method cannot identify the occlusion space. In that case, it is best to calculate the height of the occlusion space at each reference position at the intersection of the line connecting the installation position of the relevant external sensor and the detection point on the near side, and the vertical line at each reference position. This method can also be used even if there are detection points on the far side. That is, the height of the occlusion space at a predetermined position within the parking space area is calculated based on the detection information that is closer to the vehicle than the area of ​​the parking space and has the largest elevation angle.

[0094] Figures 13A and 13B illustrate the generation of occlusion space data OS(901) for parking space 901.

[0095] Since parking space 901 is included in the detectable range 112 of Figure 2, an external sensor targeting the left side (e.g., a side camera) is identified as the external sensor S (S1006). As shown in Figure 13B, parking space 901 is obscured and not visible by other vehicles in the row in front. In Figure 13A, triangles indicate detection data points on the front side of parking space 901, circles indicate detection data on the back side of parking space 901, and squares indicate detection data outside the range of parking space 901. The occlusion space data OS(901) of parking space 901 can be obtained in the same manner as described with reference to Figure 12.

[0096] So far, we have explained an example using a single external sensor, but even when using multiple external sensors (for example, a front camera and a side camera), the occlusion space data OS can be obtained through a similar process.

[0097] (Processing by the Parking Space Occupancy Model Determination Unit 15) Figure 14 is a flowchart of the processing performed by the Parking Space Occupancy Model Determination Unit 15. The processing shown in Figure 14 is best performed at the timing of map information acquisition.

[0098] First, in S1501, the parking space occupancy model determination unit 15 identifies the area and parking lot in which the vehicle 2 is traveling (S1501). The area and parking lot in which the vehicle 2 is traveling are determined by the vehicle position and orientation estimation unit 12, which identifies the position and orientation of the vehicle 2 in the parking lot map data group 33. For example, the area and parking lot can be identified by including parking lot identification information (parking lot ID) and area identification information (area ID) as metadata for the relevant parking lot map data.

[0099] Next, the parking space occupancy model data set 36 is referenced to determine the parking space occupancy model for the identified area and parking lot (S1502). As shown in Figure 6, based on the area ID and parking lot ID, model data that matches the conditions of the area and parking lot is selected as the target for application.

[0100] Then, the determined parking space occupancy model is stored in the storage unit 30 (S1503), and this process is terminated.

[0101] (Processing by the Parking Space Availability Likelihood Estimation Unit 16) Figure 15 is a flowchart of the processing performed by the Parking Space Availability Likelihood Estimation Unit 16.

[0102] In steps S1401 to S1403, the parking space availability likelihood estimation unit 16 obtains the latest information on the parking space data group 34, the occlusion space data group 35, and the parking space space occupancy model data group 36 necessary for processing, and in steps S1404 to S1412, calculates the availability likelihood for each parking space.

[0103] When the parking space availability likelihood estimation unit 16 extracts one unprocessed parking space PS (Y in S1404 and S1405), it first obtains information on the previous availability likelihood L(t-1) of the parking space PS (S1406). The previous availability likelihood of the parking space PS is the previous calculation result of the availability likelihood by the parking space availability likelihood estimation unit 16 and can be obtained from the parking space availability likelihood data group 37.

[0104] Next, the parking space availability likelihood estimation unit 16 determines whether parking space PS is a candidate for an available parking space extracted in step S1003 of Figure 10 (S1407). If parking space PS is not a candidate for an available parking space (N in S1407), it means that there is an obstacle in the area of ​​parking space PS, so the parking space availability likelihood estimation unit 16 sets the estimated availability likelihood p based on the current observation information to 0.0 (S1412). On the other hand, if parking space PS is a candidate for an available parking space (Y in S1407), the process proceeds to S1408.

[0105] The parking space occupancy likelihood estimation unit 16 generates a parking space occupancy model in the area of ​​parking space PS based on the parking space occupancy model data set 36 (S1408). For example, since the parking position changes depending on the size of the parking space, it is good practice to linearly transform the basic parking space model to match the size of the parking space. Then, the generated parking space occupancy model and the occlusion space data of parking space PS are compared in three dimensions, and an estimated occupancy likelihood value p is calculated (S1409). The specific method for calculating the estimated occupancy likelihood value will be described later with reference to Figure 16.

[0106] Next, the value of the current vacancy likelihood L(t) is determined based on the estimated vacancy likelihood p calculated from the current observation information and the previous vacancy likelihood L(t-1) (S1410). Various methods can be considered for determining the value of vacancy likelihood L(t). For example, it may be obtained by weighting the previous value and the estimated value from the observation information, such as L(t) = w1 × p + w2 × L(t-1) (where w1 + w2 = 1). The weighting coefficients w1 and w2 may be fixed or dynamically changed considering the reliability of the estimated value based on the observation information. Furthermore, when the vacancy likelihood L(t) becomes close to 0.0 or 1.0 and becomes definitive, the vacancy likelihood L(t) may be set to 0.0 or 1.0 and the update of the vacancy likelihood based on the observation information may be stopped. By using current observation data and past data, misjudgments of the vacancy likelihood due to temporary unseen conditions can be suppressed.

[0107] The likelihood of parking space PS being available L(t), determined in step S1410, is stored in the parking space availability likelihood data group 37 (S1411), and processing for that parking space PS is completed. Then, the process returns to S1404 and is repeated until there are no more unprocessed parking spaces.

[0108] Referring to Figure 16, the process of step S1409 performed by the parking space availability likelihood estimation unit 16 will be explained. In Figure 16, the occlusion space data 1611 of parking space 902 in Figure 10 is shown in the upper left, and the parking space occupancy model data for this parking lot is shown in the lower left.

[0109] When the occlusion space data 1611 is compared with the parking space occupancy model data, the boundary 1622 of the occlusion space in the cross-section 1621 on the x' axis and the parking space occupancy model data are superimposed, as shown on the right. The parking space vacancy likelihood estimation unit 16 refers to the parking space occupancy model data and obtains the maximum value of the occupancy probability at the location through which the boundary 1622 of the occlusion space passes. In the example shown in Figure 16, the maximum value of the occupancy probability at the location through which the boundary 1622 passes is 80%. The boundary 1622 of the occlusion space is considered to be the boundary with the space detectable by the external sensor group 4. Assuming that a parked vehicle is present, this means that there is no parked vehicle in a location that is occupied with an 80% probability, so the probability of a parked vehicle being inside the occlusion space can be approximated as 20%. That is, the vacancy likelihood is calculated to be 80%.

[0110] In the example shown in Figure 16, the vacancy likelihood is calculated by referring to the distribution of heights in the occlusion space along the x' axis. However, the same process may be performed on the entire spatial region of the parking space 902 to calculate the vacancy likelihood. For example, if the occlusion space data 431 is represented by the grid map shown in Figure 5, the maximum value of the occupancy probability in the parking space occupancy model data for all cell locations may be obtained to calculate the vacancy likelihood.

[0111] As described above, the likelihood of a parking space being available is estimated by first estimating the occlusion space area in three dimensions, including the height direction, within the parking space, and then estimating the probability that a parked vehicle would be contained within this occlusion space area. Therefore, since it is possible to determine whether a parking space is available by considering the degree of occlusion in the height direction, it is possible to estimate the availability status even when it is difficult to determine availability using only two-dimensional occlusion.

[0112] (Processing by the Driving Planning Unit 17) Figure 17 is a flowchart of the processes performed by the Driving Planning Unit 17.

[0113] The driving plan unit 17 generates driving plan information to efficiently guide vehicle 2 to an available parking space based on the estimated likelihood of available parking spaces around vehicle 2. Methods of guidance include presenting a guidance route to a parking space with a high probability of being available when the occupant is manually driving, and controlling the vehicle's movement according to the guidance route when the vehicle system 1 is automatically driving. The driving plan unit 17 generates information for guiding vehicle 2 (information for presenting a guidance route and information for controlling the vehicle's movement).

[0114] The driving plan unit 17 acquires the likelihood data set 37 for available parking spaces and the parking lot map data set 33 (S1701, S1702).

[0115] Next, the driving plan unit 17 identifies the link e0 on which the vehicle 2 is traveling from the vehicle position and attitude information included in the parking map data group 33 (S1703). The link can be identified, for example, by a map matching method used in navigation systems.

[0116] The driving plan unit 17 then refers to the parking map data group 33 and extracts parking spaces that face the identified link e0 (i.e., accessible from the partial road of link e0) (S1704). The driving plan unit 17 then determines whether, among the extracted parking spaces, there is a parking space in front of the vehicle 2 with an availability likelihood value in the parking space availability likelihood data group 37 that is equal to or greater than a predetermined value. If a suitable parking space PS exists (Y in S1704), the suitable parking space PS is set as a target parking space candidate (S1705). A target parking space candidate is a candidate for where the vehicle 2 will park and will be the target point of the guidance route.

[0117] If there is no corresponding parking space PS, the probability of there being an available parking space on this section of the road is low, so the driving plan unit 17 calculates a driving route that guides the vehicle to a link where there is a high probability of there being an available parking space (S1712). In step S1712, the driving plan unit 17 refers to the parking map data group 33 and extracts the link information E = {e1, e2, ...} that is connected to link e0 ahead. Note that the link information that is connected ahead includes not only links that are directly connected to link e0, but also links that are connected without branching or merging. For example, links 213, 216, and 218 in Figure 3 are treated as separate links, but since there is no branching or merging of links between nodes 224 and 227, it is appropriate to treat them as a single continuous sequence of links.

[0118] Next, the driving plan unit 17 calculates the sum of the likelihood of available parking spaces facing each link included in the link information E (S1713), and selects the link with the largest sum as the next link to be driven after link e0 (S1714). Then, it virtually sets the vicinity of the end of the next link to be driven as a candidate target parking space (S1715). This allows the vehicle 2 to be guided to the route with the highest probability of having available parking spaces when there are several driving route options in the parking lot. In this example, the guidance route is searched using two links, the current link and the next link, but the guidance route may be searched using information from more multi-stage links, or the search range may be determined according to the distance of the links.

[0119] After S1705 or S1715, the driving plan unit 17 determines whether the driving mode of the vehicle system 1 is automatic driving mode or manual driving mode (S1706). If it is manual driving mode (N in S1706), it generates display data to show the guidance route on the screen of the HMI device group 8 for the occupant (S1710) and stores it in the storage unit 30 as driving plan data group 38 (S1711).

[0120] On the other hand, in automatic driving mode (Y in S1706), information for controlling the driving of vehicle 2 is generated and stored in the memory unit 30. First, the driving plan unit 17 acquires external sensor data group 31 and vehicle sensor data group 32 (S1707, S1708) to understand obstacles around vehicle 2 and the state of vehicle 2. Then, the driving plan unit 17 generates driving plan information to control the driving of vehicle 2 (S1709). The driving plan information includes, for example, the driving trajectory of vehicle 2 and control command values ​​for driving while following the driving trajectory.

[0121] The processing of the travel planning unit 17 will be explained with reference to Figures 18 and 19.

[0122] Figure 18 shows an example of the state of vehicle 2 and parking space availability likelihood data group 37 in the parking lot map data 200 shown in Figure 3.

[0123] Vehicle 2 is located on link 212, and since there are no parking spaces facing link 212, the system proceeds to step S1712 at the branch in step S1704. The links that connect in front of link 212 are a continuous sequence of links {link 213, link 216, link 218} (route direction 1701) and link 215 (route direction 1702). The driving planning unit 17 then calculates the sum of the vacancy likelihoods of the parking spaces facing each forward connecting link (S1713). The sum of the vacancy likelihoods of the links in route direction 1701 is 1.1, and the sum of the vacancy likelihoods of the links in route direction 1702 is 0.6, indicating that the sum of vacancy likelihoods is greater for the links in route direction 1701. Therefore, the driving plan unit 17 sets the link in the route direction 1701 as the next driving link (S1714), and sets node 226, which is the end of the link in the route direction 1701, as a virtual target parking space candidate (S1715). By guiding the vehicle to a road with a large sum of vacancy likelihood, it is possible to guide the vehicle to a parking space that is likely to be vacant.

[0124] Figure 19 shows an example of the screen display information generated in the example shown in Figure 18.

[0125] The system overlays the parking map data 200 with information on the planned guidance route 1801 and parking spaces 1811 that are likely to be available. By presenting this information to the occupants, the system can guide them to parking spaces that are likely to be available, thus assisting them in searching for available parking spaces.

[0126] As described above, this embodiment allows for the rapid and wide-ranging determination of the availability of parking spaces hidden behind obstacles. In other words, it is possible to estimate the likelihood of availability not only for parking spaces directly facing the vehicle 2's path, but also for a wide area of ​​parking spaces surrounding the vehicle 2. By utilizing the estimated likelihood of parking space availability, it is possible to select a route in the parking lot that has a high probability of having available parking spaces, present it to the occupants via the HMI device group 8, or control the vehicle 2's movement, thereby supporting the efficient search for available parking spaces during the vehicle 2's parking operation.

[0127] It should be noted that the present invention is not limited to the embodiments described above, but includes various modifications and equivalent configurations within the spirit of the attached claims. For example, the embodiments described above are described in detail for the purpose of clearly illustrating the present invention, and the present invention is not necessarily limited to having all the configurations described. Furthermore, some of the configurations of one embodiment may be replaced with those of another embodiment. Furthermore, configurations of other embodiments may be added to the configuration of one embodiment. Furthermore, some of the configurations of each embodiment may be added, deleted, or replaced with those of other embodiments.

[0128] For example, in the embodiment described above, the area of ​​the parking space around the vehicle 2 is calculated based on parking lot map data, but the area of ​​the parking space may also be estimated based on information detected by the external sensor group 4.

[0129] Furthermore, in the embodiments described above, it is assumed that each process performed by the electronic control unit 3 is executed in the same processing unit and storage unit, but it may also be executed in multiple different processing units and storage units. In that case, for example, processing software with a similar configuration may be installed in each of two or more storage units, and the processing may be divided and executed by each of the two or more processing units.

[0130] Furthermore, although each process performed by the electronic control unit 3 is realized by executing a predetermined program using a processor and RAM, it may be realized with proprietary hardware as needed. Also, in the above-described embodiment, the external sensor group, vehicle sensor group, actuator group, and HMI device group are described as separate devices, but any two or more of them may be combined as needed.

[0131] Furthermore, each of the aforementioned configurations, functions, processing units, and processing means may be implemented in hardware, for example, by designing them as integrated circuits, or they may be implemented in software by having a processor interpret and execute programs that realize each function.

[0132] Information such as programs, tables, and files that implement each function can be stored in memory, hard disks, SSDs (Solid State Drives), or recording media such as IC cards, SD cards, and DVDs.

[0133] Furthermore, the drawings show control lines and information lines that are considered necessary to explain the embodiments, and do not necessarily show all control lines and information lines included in actual products to which the present invention is applied. In practice, it can be assumed that almost all components are interconnected.

Claims

1. An electronic control device for assisting vehicle parking in a parking lot, comprising: an information acquisition unit that acquires detection information obtained when a sensor mounted on the vehicle detects the area around the vehicle; a parking space extraction unit that extracts the areas of a plurality of parking spaces around the vehicle; a parking space occlusion space estimation unit that estimates occlusion space information including the height direction in the area of ​​the parking space based on the area of ​​the parking space and the detection information; and a parking space occupancy likelihood estimation unit that calculates an occupancy likelihood indicating the possibility that no vehicle is present in the parking space based on the detection information and the occlusion space information in the area of ​​the parking space.

2. An electronic control device according to claim 1, wherein the parking space vacancy likelihood estimation unit calculates the vacancy likelihood based on the height distribution of the occlusion space information.

3. An electronic control device according to claim 2, characterized in that the height distribution of the occlusion space information is represented by the height of a region along the depth axis in the parking space.

4. An electronic control device according to claim 1, wherein the parking space vacancy likelihood estimation unit calculates the vacancy likelihood of the parking space by comparing the parking space occupancy model, which probabilistically represents the spatial area occupied by a parked vehicle assumed to be present in the parking space, with the occlusion space information.

5. An electronic control device according to claim 4, characterized in that the parking space occupancy model is configured such that the value of the likelihood of occupancy for a predetermined height of the occlusion space is larger in the central region than in the outer frame region of the parking space.

6. An electronic control device according to claim 4, wherein the parking space occupancy model is determined based on the area in which the vehicle is traveling.

7. An electronic control device according to claim 1, wherein the parking space occlusion space estimation unit calculates the height of the occlusion space at a predetermined position in the area of ​​the parking space based on detection information that is closer to the vehicle than the area of ​​the parking space and has the largest elevation angle.

8. An electronic control device according to claim 1, comprising: a parking lot road information acquisition unit that acquires information about the roads of the parking lot; and a driving plan unit that determines a recommended road route for the vehicle based on the road information, wherein the driving plan unit prioritizes selecting roads with a large sum of the calculated availability likelihoods.

9. A parking assistance method comprising an electronic control device assisting in parking a vehicle in a parking lot, wherein the electronic control device comprises a computing device that performs predetermined calculation processing and a storage device connected to the computing device, and the parking assistance method comprising: an information acquisition procedure in which the computing device acquires detection information in which a sensor mounted on the vehicle detects the area around the vehicle; a parking space extraction procedure in which the computing device extracts the areas of a plurality of parking spaces around the vehicle; a parking space occlusion space estimation procedure in which the computing device estimates occlusion space information including the height direction in the area of ​​the parking space based on the area of ​​the parking space and the detection information; and a parking space occupancy likelihood estimation procedure in which the computing device calculates and outputs an occupancy likelihood indicating the possibility that no vehicle is present in the parking space based on the detection information and the occlusion space information in the area of ​​the parking space.

Images