Radar Reference Map Generation from HD Map Objects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Radar localization in autonomous vehicles faces challenges with incomplete or poor-quality radar reference maps, leading to increased driver takeovers, which can be unsafe and decrease driver satisfaction, as they require accurate and widespread mapping of stationary objects for reliable navigation.
Innovation Solution
The method involves generating a high-definition radar reference map by determining attributes of HD map objects and indicating occupancy cells in a radar occupancy grid, using a processor and computer-readable storage medium to create a robust and spatially efficient map, enabling accurate vehicle localization and reducing driver takeovers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If radar reference maps are used for vehicle localization, then autonomous navigation capability is improved, but map completeness and quality are insufficient leading to increased driver takeovers
Solution Approach 1:
The system performs preliminary mapping actions by generating a radar reference map through multiple vehicle runs before autonomous operation. The map is built in advance by collecting radar data from stationary objects (guardrails, signs, buildings) and processing them into a structured reference map that includes object locations, types, and radar cross-section values, ensuring reliable localization is available before needed
Solution Approach 2:
The system uses feedback from multiple vehicle runs to iteratively improve the radar reference map. By accumulating radar data from repeated traversals of the same environment and comparing new data with the existing map, the system refines object detections and updates the reference map to improve completeness and accuracy over time
2Reliability
If multiple vehicle runs are used to generate radar reference map, then map completeness is improved, but time consumption increases
Solution Approach 1:
The system performs preliminary mapping actions by generating a radar reference map through multiple vehicle runs before autonomous operation. The map is built in advance by collecting radar data from stationary objects (guardrails, signs, buildings) and processing them into a structured reference map that includes object locations, types, and radar cross-section values, ensuring reliable localization is available before needed
Solution Approach 2:
The system maintains continuous improvement of the radar reference map through ongoing data collection during regular vehicle operations. Each new vehicle run contributes additional radar data that is processed and integrated into the existing map, allowing the map to be refined continuously without requiring dedicated mapping expeditions
3Measurement precision
If radar reference map is updated frequently, then localization accuracy is improved, but processing complexity increases
Solution Approach 1:
The system extracts only the essential and relevant features from radar data for reference map generation. It focuses on detecting stationary objects with significant radar cross-sections (guardrails, signs, buildings) and extracts their locations, types, and radar cross-section values, while filtering out irrelevant transient objects to reduce processing complexity
Solution Approach 2:
The system applies different processing quality and detail levels to different regions of the reference map based on their importance. High-priority areas with critical localization objects receive more detailed processing and frequent updates, while less critical areas use simplified processing, optimizing the balance between accuracy and computational resources
Data Source
AI summary
Methods and systems are described that enable radar reference map generation. A high-definition (HD) map is received and one or more HD map objects within the HD map are determined. Attributes of the respective HD map objects are determined, and, for each HD map object, one or more occupancy cells of a radar occupancy grid are indicated as occupied space based on the attributes of the respective HD map object. By doing so, a radar reference map may be generated without a vehicle traversing through an area corresponding to the radar reference map.