Vehicle positioning based on radar detection in a garage

Abstract

This document describes techniques and systems for vehicle localization based on radar detections in garages and other GNSS-denied environments. In some examples, a system includes a processor and a computer-readable storage medium including instructions that, when executed, cause the system to obtain structural data about a GNSS-denied environment and generate radar localization landmarks from the structural data. The radar localization landmarks include edges or corners of the GNSS-denied environment. The instructions also cause the processor to generate a polyline along or between the radar localization landmarks to generate a radar occupancy grid. The instructions further cause the processor to receive radar detections from one or more radar sensors and obtain a vehicle pose within the radar occupancy grid to localize the vehicle within the GNSS-denied environment. In this way, the system can provide highly accurate vehicle localization in GNSS-denied environments in a cost-effective manner.

Description

Background Technology

[0001] Vehicle localization is a technique that uses sensor data to locate a vehicle on a map (e.g., determine the vehicle's position on the map). Vehicle localization can be used to support autonomous vehicle operation (e.g., navigation, route planning, lane determination and centering, and curve execution in the absence of lane markings). To accurately locate a vehicle relative to its environment, vehicle localization involves obtaining the positions of stationary positioning objects (e.g., signs, poles, obstacles) from various sensors and navigation systems on the vehicle and associating these positions with known positions on a map. The vehicle's position or attitude can then be determined relative to these objects.

[0002] Some autonomous vehicles perform driving maneuvers that depend on positioning accuracy approaching sub-meter levels. Sub-meter accuracy can be achieved through navigation systems, including Global Navigation Satellite System (GNSS) receivers. However, when a vehicle is in a GNSS-denied environment (such as an indoor parking structure or garage), GNSS systems typically cannot receive positioning data. Summary of the Invention

[0003] This document describes techniques and systems for vehicle localization based on radar detection in parking garages and other GNSS-denied environments. In some examples, the system includes at least one processor and at least one computer-readable storage medium, the latter containing instructions that, when executed by the processor, cause the system to obtain structured data about the GNSS-denied environment and generate radar positioning landmarks for the GNSS-denied environment from the structured data. The radar positioning landmarks include edges or corners of the GNSS-denied environment, which may be a parking garage. The instructions also cause the processor to generate polylines along or between the radar positioning landmarks to generate a radar occupancy grid for the GNSS-denied environment. The instructions further cause the processor to receive radar detection from one or more radar sensors and obtain the vehicle attitude within the radar occupancy grid to locate the vehicle within the GNSS-denied environment. In this way, the system can provide highly accurate vehicle localization in GNSS-denied environments in a cost-effective manner.

[0004] This document also describes methods to be performed by the systems summarized above and other configurations of the systems described herein, as well as apparatus for performing these methods.

[0005] This invention provides a simplified concept related to vehicle location based on radar detection in a garage, which is further described in the detailed description and accompanying drawings. This invention is not intended to identify essential features of the claimed subject matter, nor is it intended to define the scope of the claimed subject matter. Attached Figure Description

[0006] The following figures illustrate a system and technology for achieving vehicle localization based on radar detection in a garage. The same numbers are used throughout the figures to refer to similar features and components: Figure 1 This is an example illustration of an environment in which vehicle positioning based on radar detection in a garage can be achieved according to the technology of this disclosure; Figure 2 This is an example illustration of a system that can be used to implement vehicle positioning based on radar detection in a garage, according to the technology disclosed herein; Figure 3 This is an example conceptual diagram illustrating the process of generating a radar reference map for locating vehicles in a garage, based on the technology disclosed herein. Figure 4 This is an example conceptual diagram of the process of generating a radar occupancy grid for locating vehicles in a garage. Figure 5-1 and Figure 5-2 This is a schematic diagram of example edges and lines extracted from satellite imagery used for locating vehicles in a parking garage; and Figure 6 An example method for vehicle localization based on radar detection in a parking garage is shown. Detailed Implementation Overview

[0007] As described above, vehicle positioning enables autonomous vehicle operation (e.g., navigation, route planning, lane determination and centering, and curve execution in the absence of lane markings). Autonomous vehicle operation can depend on driving operations with positioning accuracy approaching sub-meter levels. Sub-meter accuracy can be achieved through navigation systems, including Global Navigation Satellite System (GNSS) receivers. However, when a vehicle is in a GNSS-denied environment (e.g., an indoor parking structure or garage), GNSS systems typically cannot receive positioning data, and therefore autonomous vehicle operation may be unavailable in such environments.

[0008] For GNSS-denied environments, some systems can use Simultaneous Localization and Mapping (SLAM) techniques, which map features as a vehicle traverses an unknown environment. These SLAM systems can then use the mapped features to locate the vehicle. Such systems typically rely on vision, LiDAR, or radar systems and sophisticated feature extraction algorithms to locate the primary vehicle onto environmental features (e.g., parking garage walls).

[0009] Other systems can use absolute maps of GNSS-denied environments. Absolute maps anchor the environment to a global coordinate system and allow the primary vehicle to smoothly transition between open sky and indoor environments. However, such systems rely on expensive inertial measurement unit (IMU) systems that use vehicle mileage to maintain the vehicle's position within GNSS-denied environments. Furthermore, such absolute maps for GNSS-denied environments are not available for many GNSS-denied environments.

[0010] This document describes a method and system for vehicle localization based on radar detection in garages and other GNSS-denied environments. Radar localization begins with constructing a radar reference map (e.g., a radar occupancy grid) for the GNSS-denied environment. The radar reference map can be generated and updated using various techniques as described herein. Once the radar reference map is available, real-time localization can be achieved using inexpensive radar sensors. Using the techniques described in this document, data from radar sensors can be processed to identify stationary objects or landmarks near the vehicle. Comparing landmark data derived from the vehicle's onboard sensors and systems with the landmark data detailed in the radar reference map generates an accurate attitude of the vehicle in its environment. Highly accurate vehicle attitude can be obtained cost-effectively by using inexpensive radar systems. Example Environment

[0011] Figure 1 This is an example illustration of an environment 100 in which vehicle localization based on radar detection in a garage can be implemented according to the technology of this disclosure; specifically, a radar reference map can be generated, updated, or used in environment 100 to navigate vehicle 104 in a garage. In example environment 100, system 102 is in vehicle 104 (e.g., a main vehicle) traveling along road 106 in a parking garage or other GNSS-denied environment.

[0012] System 102 uses a radar system (not shown) to transmit radar signals (not shown). The radar system receives radar reflections 108 (radar detection) of the radar signals from object 110. In environment 100, radar reflection 108-1 corresponds to object 110-1 (e.g., a pillar), radar reflection 108-2 corresponds to object 110-2 (e.g., a vehicle), and radar reflection 108-3 corresponds to object 110-3 (e.g., a garage wall).

[0013] Radar reflector 108 can be used for navigation in environment 100 based on a radar reference map generated for the environment by system 102 or a remote system (not shown), such as a reference map. Figures 4 to 5-2 As discussed, radar reflection 108 can be further used in conjunction with existing radar reference maps to locate vehicles 104. Example System

[0014] Figure 2 This is an example illustration 200 of a system for implementing vehicle localization based on radar detection in a garage, according to the technology of this disclosure. Example illustration 200 includes a system 102 and a cloud system 202 for a vehicle 104. Although the vehicle 104 is shown as a car, it may include any vehicle (e.g., truck, bus, boat, airplane, etc.) without departing from the scope of this disclosure. System 102 and cloud system 202 may be connected via a communication link 204. One or both of system 102 and cloud system 202 may be used to perform the technology described herein.

[0015] like Figure 2 As shown, system 102 and cloud system 202 each include at least one processor 206 (e.g., processor 206-1 and processor 206-2, respectively), at least one computer-readable storage medium 208 (e.g., computer-readable storage media 208-1 and 208-2, respectively), radar positioning module 210 (e.g., radar positioning modules 210-1 and 210-2, respectively), and communication system 212 (e.g., communication systems 212-1 and 212-2, respectively). Communication system 212 facilitates the exchange of data via communication link 204, including parking garages (e.g., Figure 1 Radar reference maps for environments 100 or other GNSS-rejected environments.

[0016] System 102 further includes a navigation system 214 and a radar system 216. Navigation system 214 may include a geospatial positioning system (e.g., a Global Positioning System (GPS), Global Navigation Satellite System (GNSS or GLONASS) sensor), an inertial measurement system (e.g., a gyroscope or accelerometer), or other sensors (e.g., a magnetometer, a software positioning engine, wheel tick sensors, a lidar odometer, a visual odometer, a radar odometer, or other sensor odometers). Navigation system 214 may provide high-accuracy location data under ideal conditions (e.g., within one meter) or low-accuracy location data (e.g., within several meters) under non-ideal conditions (e.g., in a parking garage).

[0017] Radar system 216 includes radar hardware for transmitting and receiving radar signals (e.g., radar reflection 108). In some implementations, radar system 216 provides static detection to radar positioning module 210 (e.g., filtering can be performed within radar system 216).

[0018] Processor 206 (e.g., application processor, microprocessor, digital signal processor (DSP), or controller) is configured to execute computer-executable instructions 218 (e.g., instructions 218-1 and 218-2) stored in computer-readable storage medium 208 (e.g., non-transient storage devices such as hard disk drives, solid-state drives (SSDs), flash memory, read-only memory (ROM), erasable programmable ROM (EPROM), or electrically erasable programmable ROM (EEPROM)) to cause system 102 and cloud system 202 to perform the techniques described herein. Instructions 218 may be part of an operating system and / or one or more applications of system 102 and cloud system 202.

[0019] Instruction 218 instructs system 102 and cloud system 202 to operate on data 220 (e.g., create, receive, modify, delete, transmit, or display) data (e.g., 220-1 and 220-2). Data 220 may include application data, module data, sensor data, or input / output (I / O) data. Data 220 also includes a radar reference map of a GNSS-denied environment (e.g., a parking garage). Although shown as being within computer-readable storage medium 208, portions of data 220 may reside in random access memory (RAM) or cache of system 102 and cloud system 202 (not shown). Furthermore, instruction 218 and / or data 220 may be located remotely to system 102 and cloud system 202.

[0020] The radar positioning module 210 (or portions thereof) may be located within a computer-readable storage medium 208 or may be a separate component (e.g., executed in dedicated hardware communicating with the processor 206 and the computer-readable storage medium 208). For example, instructions 218 may cause the processor 206 to implement, or otherwise cause the system 102 or cloud system 202 to implement, the techniques described herein. Constructing radar reference maps

[0021] Figure 3 This is an example concept diagram 300 illustrating the process of generating a radar reference map for locating vehicles in a garage. Concept diagram 300 can be executed by system 102 and / or cloud system 202. Concept diagram 300 may include additional or fewer operations, or be executed in a different order.

[0022] At 302, radar detection 304 is received. Radar detection 304 includes stationary radar detection (e.g., detection of stationary objects from radar system 216, including signs, poles, obstacles, landmarks, buildings, overpasses, curbs, road-adjacent objects such as fences, trees, vegetation, leaves, or spatial statistical patterns) and corresponding coordinates (e.g., from navigation system 214) for each time and / or location. Radar detection 304 may include point clouds with corresponding uncertainties, and / or include various radar data or sensor measurements.

[0023] At position 306, radar occupancy grid 308 is obtained from system 102 or cloud system 202. The technology for determining radar occupancy grid 308 will revolve around... Figure 4 , Figure 5-1 as well as Figure 5-2 Further discussion. The radar occupancy grid 308 is a grid-based representation of a GNSS rejected environment (e.g., a parking garage). For example, the radar occupancy grid 308 can be a Bayesian, Dempster-Schafer, or other type of occupancy grid. Each cell of the radar occupancy grid 308 represents an independent portion of space, and each cell value of the radar occupancy grid 308 represents the probability (e.g., 0-100%) that the corresponding portion of the space is occupied. A probability of approximately 0% for a cell may indicate that the corresponding portion of the space is free or currently not permanently occupied, while a probability close to 100% may indicate that the corresponding portion of the space is occupied (e.g., occupied by a wall or pillar) and therefore not free space. The radar occupancy grid 308 can also be supplemented with information determined from radar detection 302.

[0024] At position 310, radar landmarks 312 are determined from radar occupancy grid 308. Radar landmark 312 is the center coordinate of the corresponding group of cells in radar occupancy grid 308 with a probability greater than a threshold. Radar landmark 312 includes clusters, outlines, or bounding boxes of cells in radar occupancy grid 308. Radar landmark 312 has weights based on one or more of the probability, classification, or cross-sectional values ​​of each radar landmark 312. Radar landmark 312 can be determined using binarization, clustering algorithms, or machine learning on radar occupancy grid 308. The determination of radar landmark 312 typically involves applying a threshold to radar occupancy grid 308 and removing any noise from radar occupancy grid 308. In other implementations, radar landmark 312 may already be determined or identified in the radar occupancy grid.

[0025] At 314, a radar reference map 316 is generated from radar landmarks 312 and / or radar occupancy grids 308. The radar reference map 316 may be a statistical reference map (e.g., Gaussian representation). The radar reference map 316 may include a set of Gaussian distributions 318 corresponding to the occupied areas. Each Gaussian distribution 318 (or a cell in the radar reference map 316) has associated location information (e.g., low-quality or high-quality location information depending on how the radar reference map 316 was generated). Each cell in the radar reference map 316 may have a single Gaussian distribution 318 or may be blank. Although not required, the radar reference map 316 has cells larger than the cells in the radar occupancy grids 308. The radar reference map 316 may be a standalone map or a layer in another map (e.g., a layer in a high-resolution map).

[0026] System 102 can use radar occupancy grid 308 and radar detection 304 to remove static objects (e.g., parked vehicles) from radar reference map 316. For example, system 102 can receive radar detection 304 from one or more radar sensor locations around vehicle 104. System 102 determines whether radar detection 304 is a static detection based on self-trajectory information and outputs any identified static detections.

[0027] The radar reference map 316 may contain metadata associated with the corresponding Gaussian distribution 318. For example, the metadata may contain information about the shape or size of clusters of Gaussian distributions 318. The metadata may also include object associations, such as certain Gaussian distributions 318 belonging to signs or fences. Location data may also be included in the metadata.

[0028] Figure 4 This is an example concept diagram 400 illustrating the process of generating a radar occupancy grid for vehicle localization in garages and other GNSS-rejected environments. Concept diagram 400 can represent the process performed by system 102 of vehicle 104 or cloud system 202. Figure 3 Operation 306 (determine radar occupancy grid). In other implementations, the output of concept diagram 400 can be obtained by system 102 from memory or cloud system 202. Figure 3 The radar occupancy grid 308 provides a reference map that system 102 can use to locate vehicle 104 in a world or map coordinate system. Concept map 400 can be performed by system 102 alone or in combination with cloud system 202. For the purposes of this discussion, this document describes system 102 as performing the operation of concept map 400.

[0029] System 102 begins with a series of processing operations to generate a set of edges and corners as radar-localized landmarks. At 404, system 102 performs edge and corner extraction on structural data 402 to generate a list of landmarks 406. Structural data 402 provides a two-dimensional footprint of the edges and corners of a GNSS-rejected environment. Because high-resolution maps are generally not available for parking garages and other GNSS-rejected environments, system 102 uses existing two-dimensional structural representations to obtain structural data 402 and generate the list of landmarks 406. For example, system 102 can use blueprints, design drawings, or satellite imagery to extract walls, pillars, edges, and corners of the GNSS-rejected environment. Image processing algorithms can be used to process the features included in structural data 402 to identify lines and corners.

[0030] If the structured data 402 is a satellite image, system 102 can perform operation 404 by binarizing the satellite image into a grayscale image and enhancing its contrast. System 102 can then apply the Prewitt approximation algorithm or another edge and corner extraction algorithm to extract edges and corners from the structured data 402 and generate a list of landmarks 406. As another example, system 102 can use a machine learning model to process color or grayscale images containing structured data 402.

[0031] At 408, system 102 converts the location of each object in the landmark list 406 to a world frame. For example, system 102 can then define the area where the parking garage is located in a global coordinate system or a map coordinate system. Then, system 102 can determine the coordinates of each landmark in the map coordinate system.

[0032] At 410, system 102 performs a series of additional processing operations to generate polylines along each edge of the landmark. A polyline is a geometric representation of the landmark relative to radar occupancy grid 308. For example, the position, orientation, and details (e.g., offset) of a wall or pillar can be used to generate polylines in radar occupancy grid 308 corresponding to the occupied space of the wall or pillar.

[0033] At 412, system 102 determines whether the distance between two landmarks is greater than a threshold. At 414, if the distance between landmarks is greater than the threshold, system 102 determines that the landmarks are not connected. At 416, if the distance between landmarks is not greater than the threshold, system 102 oversamples along the polyline to simulate the output of the radar occupancy grid. System 102 can also compare the length of the corresponding polyline with the grid size of radar occupancy grid 308. If the polyline is not longer than the grid cell of radar occupancy grid 308, the corresponding grid cell is marked as unoccupied. If the polyline is longer than the grid cell, system 102 can oversample to create an oversampled polyline. Oversampling involves adding more points along the corresponding polyline to simulate the radar occupancy grid output from radar detection.

[0034] At 418, system 102 can perform sensor offset adjustment to adjust the polyline or oversampled polyline. For example, if a known sensor perception offset exists for a particular type of sensor, system 102 can apply it to sampling along the polyline or oversampled polyline and generate a list of discrete points for the radar occupancy grid. Each cell of the radar occupancy grid represents the probability that the cell is occupied by a landmark or not. The radar occupancy grid can then be used for real-time positioning in garages or other GNSS-rejected environments, such as reference... Figure 3 As described.

[0035] Figure 5-1 and Figure 5-2 This is a schematic diagram of example edges and lines extracted from satellite imagery used for locating vehicles in a parking garage. Specifically, Figure 5-1 An example satellite image 500-1 shows a parking garage structure. Satellite image 500-1 provides a top view of the parking garage structure. The parking garage structure includes an outer wall 502 with an approximately rectangular shape. The parking garage structure also includes an inner wall 504 that provides ramps between parking levels of the parking garage structure and has an approximately U-shape.

[0036] Figure 5-2 Example landmark 500-2 shows a parking garage structure. System 102 or cloud system 202 can be used. Figure 4 Conceptual diagram 400 extracts lines, edges, and corners from satellite image 500-1 as landmarks 500-2. Then, a radar occupancy grid can be generated by assuming the top-level layout of the parking garage structure depicted in satellite image 500-1 represents each floor of the parking garage structure. If each floor of the parking garage structure has a different floor plan, a structural diagram or blueprint is used to construct the radar occupancy grid for each floor. Example Method

[0037] Figure 6An example method 600 for vehicle localization based on radar detection in a parking garage is illustrated. Method 600 is shown as a set of operations (or actions) performed, but is not limited to the order or combination of operations shown herein. Furthermore, any one or more operations can be repeated, combined, or recombined to provide other methods. References may be made in the sections discussed below. Figure 1 Environment 100 and Figure 1 The entities detailed in Figure 5 are referenced for illustrative purposes only. This technique is not limited to being performed by one or more entities. For example, the system could be... Figure 1 and Figure 2 System 102 or Figure 2 The cloud system 202.

[0038] At point 602, structural data about the GNSS-denied environment is obtained. For example, cloud system 202 can obtain structural data 402 about environment 100, which may include a parking garage. The structural data may include satellite imagery of the GNSS-denied environment. The structural data may also include blueprints or other design drawings of the GNSS-denied environment.

[0039] At point 604, radar location landmarks for a GNSS-rejected environment are generated from the structural data. These radar location landmarks include at least one edge or corner of the GNSS-rejected environment. For example, cloud system 202 can generate radar location landmarks (including the edges, corners, and pillars of a parking garage) for environment 100 from structural data 402.

[0040] The cloud system 202 can binarize satellite imagery of GNSS-rejected environments into grayscale images. Then, the cloud system 202 can apply extraction algorithms to the grayscale image to identify radar-located landmarks. Before performing the extraction process, the cloud system 202 can also enhance the contrast of the grayscale image, or perform pixel enhancement on at least some pixels of the grayscale image, to improve the extraction of radar-located landmarks. The cloud system 202 can also apply geofencing technology to the grayscale image to focus its processing on a specific area of ​​the grayscale image.

[0041] At position 606, a polyline is determined along or between radar location landmarks to generate a radar occupancy grid for the GNSS rejected environment. For example, cloud system 202 can generate a polyline along or between radar location landmarks. The polyline can be used to generate radar occupancy grid 308 for environment 100. Cloud system 202 can also convert each radar location landmark and radar occupancy grid 308 into a map coordinate system.

[0042] At 608, radar detections from one or more radar sensors of the vehicle can be received. For example, system 102 can receive radar detections 304 from one or more radar sensor locations of vehicle 104. System 102 can also generate an occupancy probability for any location in a GNSS denied environment. This probability can be based on the radar occupancy grid 308 and the radar detections 304.

[0043] At 610, the vehicle attitude within the radar occupancy grid can be determined based on radar detection to locate the vehicle in a GNSS-rejected environment. For example, system 102 can determine the vehicle attitude within radar occupancy grid 308 based on radar detection 304. In some implementations, system 102 can use radar detection 304 to remove temporary static objects (e.g., parked vehicles) from radar reference map 316. The vehicle attitude can be used to locate the vehicle within environment 100. System 102 can use the map coordinate system and radar occupancy grid 308 to manipulate vehicle 104 to enter a GNSS-rejected environment. System 102 can also use the vehicle attitude and radar occupancy grid 308 to manipulate the vehicle (e.g., parked vehicles) in a GNSS-rejected environment. Example

[0044] Example 1. A method comprising: obtaining structural data about a GNSS denied environment; generating radar positioning landmarks of the GNSS denied environment from the structural data, the radar positioning landmarks including at least one of the edges or corners of the GNSS denied environment; determining a polygonal line along or between the radar positioning landmarks to generate a radar occupancy grid of the GNSS denied environment; receiving radar detections from one or more radar sensors of a vehicle; and determining the attitude of the vehicle within the radar occupancy grid based on the radar detections to locate the vehicle within the GNSS denied environment.

[0045] Example 2. The method of Example 1, where the GNSS rejection environment is a parking structure.

[0046] Example 3. The method of Example 2, where the structural data includes satellite imagery of the parking structure.

[0047] Example 4. The method of Example 3, wherein generating radar-located landmarks for a GNSS-rejected environment includes: binarizing a satellite image of the GNSS-rejected environment into a grayscale image; and applying an extraction algorithm to the grayscale image to identify radar-located landmarks.

[0048] Example 5. The method of Example 4, wherein generating radar-located landmarks for a GNSS denied environment further includes: before applying the extraction algorithm to the grayscale image, performing at least one of the following operations: enhancing the contrast of the grayscale image; performing pixel enhancement on at least some pixels of the grayscale image; or applying a geofence to the grayscale image.

[0049] Example 6. Any of the foregoing examples, the method further comprising: converting each radar-located landmark and radar occupancy grid into a map coordinate system; and using the map coordinate system and radar occupancy grid to operate a vehicle to enter a GNSS-rejected environment.

[0050] Example 7. Any of the methods in the preceding examples, where the structured data includes blueprints or design drawings of the GNSS rejection environment.

[0051] Example 8. Any of the methods in the foregoing examples, the method further comprising: operating a vehicle in a GNSS denied environment using a radar occupancy grid and the vehicle's attitude.

[0052] Example 9. Any of the foregoing examples, the method further comprising: generating an occupancy probability for any location in a GNSS denied environment, the probability being based on a radar occupancy grid and radar detection.

[0053] Example 10. A system comprising: one or more processors configured to perform any of the methods described in the preceding examples.

[0054] Example 11. A non-transient computer-readable medium storing computer-executable instructions that, when executed by a processor, cause the processor to perform a method of any one of Examples 1 to 9. Conclusion

[0055] Although implementations of vehicle localization based on radar detection in garages and other GNSS-denied environments have been described in language specific to certain features and / or methods, the subject matter of the appended claims is not necessarily limited to the specific features or methods described. Rather, the specific features and methods are disclosed as example implementations of vehicles based on radar detection in garages. Furthermore, although various examples have been described above, each with specific features, it should be understood that a specific feature of an example need not be used exclusively with that example. Rather, any feature described above and / or depicted in the accompanying drawings may be combined with any of the examples, in addition to or in lieu of any features of the other features of these examples.

Claims

1. A method, the method comprising: The radar grid occupancy is determined by the following steps: Structural data about GNSS rejection environments are obtained by using existing two-dimensional construction structure representations; Edge and corner extraction is performed on the structured data to generate a list of radar location landmarks for the GNSS rejected environment, wherein the radar location landmarks include at least one of the edges or corners of the GNSS rejected environment; Convert the corresponding position of each object in the radar-located landmark list into a map coordinate system; Determine the polygonal lines along or between the radar location landmarks to generate the radar occupancy grid for the GNSS rejection environment. Each cell of the radar occupancy grid represents an independent portion of space, and each cell value of the radar occupancy grid represents the probability that the corresponding portion of space is occupied. The polygonal line is a geometric representation of the radar positioning landmark relative to the radar-occupied grid. Receives radar detection from one or more radar sensors of the vehicle; as well as The attitude of the vehicle within the radar-occupied grid is determined based on the radar detection, in order to locate the vehicle in the GNSS-rejected environment.

2. The method as described in claim 1, characterized in that, The GNSS rejection environment is a parking structure.

3. The method as described in claim 2, characterized in that, The structural data includes satellite images of the parking structure.

4. The method as described in claim 3, characterized in that, Performing edge and corner extraction on the structured data includes: Binarize the satellite image of the GNSS rejected environment into a grayscale image; and The extraction algorithm is applied to the grayscale image to identify the radar-located landmarks.

5. The method as described in claim 4, characterized in that, Performing edge and corner extraction on the structural data further includes: before applying the extraction algorithm to the grayscale image, performing at least one of the following operations: enhancing the contrast of the grayscale image; performing pixel enhancement on at least some pixels of the grayscale image; or applying a geofence to the grayscale image.

6. The method of claim 1, further comprising: The vehicle is operated using the map coordinate system and the radar occupancy grid to enter the GNSS denied environment.

7. The method as described in claim 1, characterized in that, The structural data includes blueprints or design drawings of the GNSS rejection environment.

8. The method of claim 1, further comprising: Operate the vehicle in the GNSS denied environment using the radar occupancy grid and the vehicle's attitude.

9. The method of claim 1, further comprising: Generate the occupancy probability of any location in the GNSS denied environment, the probability being based on the radar occupancy grid and the radar detection.

10. A system comprising: One or more processors, said one or more processors being configured to: The radar grid occupancy is determined by the following steps: Structural data about GNSS rejection environments are obtained by using existing two-dimensional construction structure representations; Edge and corner extraction is performed on the structured data to generate a list of radar location landmarks for the GNSS rejected environment, wherein the radar location landmarks include at least one of the edges or corners of the GNSS rejected environment; Convert the corresponding position of each object in the radar-located landmark list into a map coordinate system; Determine the polygonal lines along or between the radar location landmarks to generate the radar occupancy grid for the GNSS rejection environment. Each cell of the radar occupancy grid represents an independent portion of space, and each cell value of the radar occupancy grid represents the probability that the corresponding portion of space is occupied. The polygonal line is a geometric representation of the radar positioning landmark relative to the radar-occupied grid. Receives radar detection from one or more radar sensors of the vehicle; as well as The attitude of the vehicle within the radar-occupied grid is determined based on the radar detection, in order to locate the vehicle in the GNSS-rejected environment.

11. The system as claimed in claim 10, characterized in that, The GNSS rejection environment is a parking structure.

12. The system as claimed in claim 11, characterized in that, The structural data includes satellite images of the parking structure.

13. The system as described in claim 12, characterized in that, The one or more processors are configured to perform edge and corner extraction on the structured data in the following manner: Binarize the satellite image of the GNSS rejected environment into a grayscale image; and The extraction algorithm is applied to the grayscale image to identify the radar-located landmarks.

14. The system as described in claim 13, characterized in that, The one or more processors are configured to: generate the radar location landmark of the GNSS rejection environment by at least one of the following operations before applying the extraction algorithm to the grayscale image: enhancing the contrast of the grayscale image; Perform pixel enhancement on at least some pixels of the grayscale image; Alternatively, a geofence can be applied to the grayscale image.

15. The system as described in claim 10, characterized in that, The one or more processors are further configured to: The vehicle is operated using the map coordinate system and the radar occupancy grid to enter the GNSS denied environment.

16. The system as claimed in claim 10, characterized in that, The structural data includes blueprints or design drawings of the GNSS rejection environment.

17. The system as claimed in claim 10, characterized in that, The one or more processors are further configured to operate the vehicle in the GNSS denied environment using the radar occupancy grid and the vehicle's attitude.

18. The system as claimed in claim 10, characterized in that, The one or more processors are further configured to generate occupancy probabilities for any location in the GNSS denied environment, the probabilities being based on the radar occupancy grid and the radar detection.

19. A non-transitory computer-readable medium storing computer-executable instructions, which, when executed by a processor, cause the processor to: The radar grid occupancy is determined by the following steps: Structural data about GNSS rejection environments are obtained by using existing two-dimensional construction structure representations; Edge and corner extraction is performed on the structured data to generate a list of radar location landmarks for the GNSS rejected environment, wherein the radar location landmarks include at least one of the edges or corners of the GNSS rejected environment; Convert the corresponding position of each object in the radar-located landmark list into a map coordinate system; Determine the polygonal lines along or between the radar location landmarks to generate the radar occupancy grid for the GNSS rejection environment. Each cell of the radar occupancy grid represents an independent portion of space, and each cell value of the radar occupancy grid represents the probability that the corresponding portion of space is occupied. The polygonal line is a geometric representation of the radar positioning landmark relative to the radar-occupied grid. Receives radar detection from one or more radar sensors of the vehicle; as well as The attitude of the vehicle within the radar-occupied grid is determined based on the radar detection, in order to locate the vehicle in the GNSS-rejected environment.

20. The non-transient computer-readable medium as claimed in claim 19, characterized in that: The GNSS rejection environment is a parking structure; The structural data includes satellite images of the parking structure; and Performing edge and corner extraction on the structured data includes: The satellite image of the GNSS rejection environment is binarized into a grayscale image; as well as The extraction algorithm is applied to the grayscale image to identify the radar-located landmarks.

Images