Crowdsourced HD Map Alignment for Reliable Lane Feature Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for constructing high-definition maps from crowd-sourced data face challenges due to variability and reliability issues, particularly in updating maps for less-traveled road segments and regions that require accurate identification of features like lane lines and signs, which is labor-intensive and costly when relying on single data sources.
Innovation Solution
The method involves using semantic attributes to align and combine sensor data from multiple vehicles traversing different paths of a road segment, with confidence levels inversely proportional to distance, and refining features using a maximum likelihood estimator to generate a stable and accurate high-definition map, enabling autonomous vehicle control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If crowd sourced sensor data from multiple vehicles is used to construct high definition maps, then map coverage and update efficiency are improved, but data variability and reliability issues worsen
Solution Approach 1:
The patent combines sensor data from multiple vehicles traversing different paths along the same road segment. By merging data from multiple sources (first vehicle's first sensor data, second vehicle's second sensor data), the system achieves both improved map coverage and maintained reliability through data aggregation and cross-validation.
Solution Approach 2:
The system employs confidence scoring where confidence of sensor data is inversely proportional to distance from the vehicle path. This feedback mechanism allows the system to weight data appropriately, improving reliability by giving higher weight to more reliable data points while still incorporating broader coverage from multiple vehicles.
2Measurement precision
If sensor data confidence is inversely proportional to distance from path, then near-path features are identified with high confidence, but far-path features have low confidence and may be missed
Solution Approach 1:
The patent aligns and combines sensor data from multiple vehicles traveling along different paths. Features detected by one vehicle at a distance (low confidence) can be complemented by features detected by another vehicle closer to that location (high confidence), thereby reducing information loss while maintaining measurement precision through the aggregation of multiple perspectives.
3Manufacturing precision
If manual methods are used to update map data for less-traveled road segments, then map accuracy is maintained, but labor cost and time consumption increase
Solution Approach 1:
The system enables automatic map construction and updating by processing crowd sourced sensor data from multiple vehicles without requiring manual intervention. The automated alignment, feature identification, and map generation processes replace manual methods, reducing labor complexity while maintaining map accuracy through algorithmic processing of multi-vehicle data.
Data Source
AI summary
A method, apparatus and computer program product are provided for constructing a high definition map from crowd sourced data using semantic attributes to bootstrap map construction. Methods may include: receiving first sensor data from a first vehicle having traversed a first path along a first lane of a first road segment; identifying features from the first sensor data of the first road segment; receiving second sensor data from a second vehicle having traversed a second path along a second lane of the first road segment; identifying features from the second sensor data of the first road segment; aligning the identified features from the second sensor data with the identified features from the first sensor data of the first road segment; and combining the identified features from the first sensor data and the second sensor data based, at least in part, on the confidence of the respective sensor data.