Road Sign Position Offset Correction Using Multi-Vehicle Observations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing navigation systems, especially those for autonomous and semi-autonomous vehicles, face inaccuracies in detecting road sign positions due to offsets between sensor data and actual sign locations, leading to potential collisions and navigation errors.
Innovation Solution
A method and system that compute positional offsets by obtaining road sign observations from multiple vehicles, including location, heading, and speed data, and derive a function to determine the accurate position of road signs using ground truth data and learned heading information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If road sign observations are captured by sensors installed in running vehicles, then road sign detection coverage is improved, but positional accuracy deteriorates due to offset between sensor data and actual sign location
Solution Approach 1:
The patent introduces an intermediary computational process that uses multiple sensor data sources (GPS, IMU, camera) and ground truth data to calculate and correct the offset between vehicle sensor observations and actual road sign positions. This intermediary calculation layer mediates between the raw sensor data and the final position determination, resolving the accuracy issue while maintaining broad detection coverage.
Solution Approach 2:
The system implements feedback by comparing observed road sign positions with ground truth data, calculating the offset error, and using this feedback to correct subsequent position estimates. The iterative refinement process continuously improves positional accuracy by learning from previous measurements and adjusting for systematic errors.
2Device complexity
If GPS logger information is used for road sign positioning, then system simplicity is improved, but positional accuracy deteriorates due to distance gap between GPS location and actual sign
Solution Approach 1:
The patent merges multiple data sources including GPS location information, IMU sensor data (heading, speed), camera observations, and ground truth data into a unified positioning system. By combining these complementary sources, the system achieves high positional accuracy without significantly increasing overall system complexity, as the integration is performed through a coordinated algorithmic framework.
Solution Approach 2:
The system dynamically adjusts positioning parameters by calculating offset corrections based on vehicle speed, heading, and observed sign characteristics. These parameter changes allow the system to adapt its positioning accuracy to different driving conditions and sign types, maintaining high precision while using relatively simple underlying sensors.
Data Source
AI summary
A method, system and computer program product for determining a positional offset associated with a location of a road sign are disclosed herein. The method comprises obtaining a first plurality of road sign observations of the road sign captured by a plurality of vehicles, wherein the first plurality of road sign observations comprise location data of the plurality of vehicles, heading data of the plurality of vehicles, and speed data of the plurality of vehicles. The method further comprises computing, by a processor, a plurality of longitudinal offsets between at least a second plurality of road sign observations of the first plurality of road sign observations and ground truth data associated with the road sign. Further, the method comprises deriving, by the processor, a function, based on the plurality of longitudinal offsets and the speed data in the at least second plurality of road sign observations, and determining, by the processor, the positional offset associated with the location of the road sign, based on a learned heading and the derived function, wherein the learned heading is based on the heading data of the plurality of vehicles.