Progressive Semantic Mapping for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating detailed, precise AV-quality maps for autonomous or semi-autonomous vehicle navigation are costly, time-intensive, and technically challenging, lacking scalable and efficient solutions.
Innovation Solution
A holistic pipeline that progressively builds higher quality maps by ingesting richer input data from various sources, including user devices and vehicles, assigning quality levels based on resolution, volume, recency, and verification metrics, and automatically scheduling resources for data collection and verification to upgrade map regions to AV-quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional methods are used to generate detailed AV-quality maps, then map precision and reliability are improved, but cost, time consumption, and technical complexity increase significantly
Solution Approach 1:
The patent segments the map generation process into multiple quality levels (e.g., basic navigation quality vs. AV-quality). Different regions of the map can have different quality levels, allowing the system to provide high precision where needed while reducing complexity and cost in other areas. This segmentation enables progressive enhancement of map quality without requiring complete re-generation of the entire map at maximum quality.
Solution Approach 2:
The patent applies partial action by generating map data at sufficient quality levels for specific purposes rather than always generating complete AV-quality maps. The system determines the appropriate quality level needed for each region based on operational requirements, avoiding unnecessary complexity and cost while maintaining adequate precision for autonomous vehicle operation.
2Manufacturing precision
If traditional methods are used to generate detailed AV-quality maps, then map precision and reliability are improved, but time consumption increases significantly
Solution Approach 1:
The patent implements preliminary action by pre-generating map data at multiple quality levels and pre-identifying regions that require AV-quality detail. This allows the system to quickly access and deploy appropriate map quality for autonomous vehicle operation without time-consuming on-demand generation, significantly reducing time consumption while maintaining precision where required.
Solution Approach 2:
The system generates map data at the minimum necessary quality level for each region rather than uniformly generating complete AV-quality maps everywhere. This partial action approach reduces overall time consumption while maintaining sufficient precision for safe autonomous vehicle operation in critical areas.
3Manufacturing precision
If traditional methods are used to generate detailed AV-quality maps, then map precision and reliability are improved, but capital investment increases significantly
Solution Approach 1:
The patent applies local quality by assigning different quality levels to different regions of the map based on their specific requirements. High-precision AV-quality map data is generated only for regions where autonomous vehicles operate, while other regions receive lower quality data. This local differentiation maintains necessary precision for safety while significantly reducing capital investment compared to generating complete high-quality maps of all areas.
Solution Approach 2:
The system invests capital partially by generating map data at the exact quality level needed for autonomous vehicle operation rather than over-investing in uniformly high-quality maps. This approach maintains sufficient precision for safe operation while reducing unnecessary capital expenditure on excessive quality levels.
Data Source
AI summary
Systems, methods, and non-transitory computer-readable media can determine map information defining a map, wherein the map comprises a plurality of regions. A quality level is assigned to each region of the plurality of regions based on map information available for that region. The quality level is associated with at least one of: a resolution metric, a volume metric, a recency metric, a verification metric, or an elegance metric associated with the map information available for that region. A first region of the plurality of regions is identified that is at risk of being downgraded to a lower quality level. Instructions are issued to one or more vehicles that cause the one or more vehicles to traverse the first region and capture sensor data within the first region.