Vehicle Location Using Sparse High-Accuracy Object Map Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous-driving vehicles face challenges in accurately determining their location due to unreliable Satellite Positioning System (SPS) signals, especially under weak signal conditions, and errors in dead reckoning caused by inaccurate sensor data.
Innovation Solution
The method involves obtaining first map data with high uncertainties and density, and second map data with lower uncertainties and lower density, to determine a location estimate based on specific objects and their angles or distances from the vehicle.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If SPS signals are used for location determination, then global position information can be obtained, but accuracy degrades significantly under weak signal conditions such as when line-of-sight is obstructed by tall buildings, mountains, or canyon walls
Solution Approach 1:
The patent introduces map data as an intermediary reference system. Instead of relying solely on direct SPS signals from satellites, the system uses pre-stored map data containing object locations and characteristics as a mediator to determine vehicle position. This allows the system to achieve accurate location determination even when SPS signals are obstructed, as the map data provides reference points that can be matched with sensor observations of the environment.
2Productivity
If dead reckoning is used to calculate updated position based on initial position and sensor data, then continuous position tracking is achieved, but error accumulates over long stretches of time resulting in significant position error
Solution Approach 1:
The patent implements a feedback mechanism where map data serves as a reference truth. The system continuously compares sensor-based position estimates against the known map data, and when discrepancies are detected (indicating error accumulation), the system uses the map reference to correct the position estimate. This feedback loop prevents unbounded error accumulation while maintaining continuous tracking capability.
Solution Approach 2:
The system performs preliminary action by pre-storing detailed map data containing object locations, types, and characteristics before the vehicle reaches the area. This advance preparation creates a reference framework that can be quickly consulted during operation to correct positioning errors without requiring real-time complex calculations or external signal sources.
3Quantity of substance
If high-density map data with high uncertainty is used, then more location points are available for reference, but the precision of individual location points is reduced
Solution Approach 1:
The patent merges multiple sources of information to compensate for individual uncertainties. Instead of relying on a single high-uncertainty location point, the system combines data from multiple map objects (different locations, types, and characteristics) along with sensor observations to determine vehicle position. This aggregation of multiple imperfect measurements produces a more accurate overall position estimate than any single measurement could provide.
Data Source
AI summary
A method of determining location of a vehicle includes: obtaining first map data including first identifiers of first objects and corresponding first locations and first uncertainties each of at least a first threshold distance; obtaining second map data including second identifiers of second objects and corresponding second locations and second uncertainties each of less than a second threshold distance, where the first threshold distance is at least twice the second threshold distance, and the first locations have a higher density than the second locations; and determining a location estimate, of the vehicle, based on at least a particular one of the second locations, corresponding to a particular one of the second objects, and at least one of an angle from the vehicle to the particular one of the second objects or a distance between the vehicle and the particular one of the second objects.