Intersection Signal-Stop Line Pairing for Autonomous Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing map generation systems for autonomous driving struggle to accurately pair signal information and stop line information for all signals and stop lines in an intersection, requiring specific scene conditions and resulting in incomplete pairing.
Innovation Solution
A map generation apparatus that acquires probe data from multiple vehicles, identifies signal and stop line information using dedicated units, and pairs them using a method that identifies multiple pairing candidates across the intersection, employing both provisional and non-provisional pairing methods to ensure comprehensive and accurate pairing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If probe data is acquired from multiple vehicles to improve pairing completeness, then the coverage and reliability of map generation are improved, but the complexity of processing and matching signal-stop line pairs increases
Solution Approach 1:
The patent segments the pairing process into multiple stages: first identifying candidate pairs from probe data, then filtering and validating these candidates through specific criteria (spatial relationship, temporal coincidence, lane correspondence). This segmentation allows complex multi-vehicle data to be processed systematically in manageable steps, improving pairing completeness while controlling processing complexity through structured approach.
Solution Approach 2:
The patent performs preliminary identification of signal and stop line information from probe data before the actual pairing process. By pre-processing the data to extract and organize potential pairs, the system reduces the complexity of the main pairing operation and enables more reliable matching across multiple vehicles without overwhelming computational burden.
2Productivity
If traditional pairing methods are used that work for single vehicle, then processing is simpler, but pairing is incomplete for multi-vehicle scenarios
Solution Approach 1:
The patent creates a pairing system that is universal enough to handle both single-vehicle and multi-vehicle scenarios. The same core algorithms and criteria used for single-vehicle pairing are extended to multi-vehicle contexts, allowing the system to maintain high pairing efficiency while achieving complete pairing coverage across multiple vehicles through aggregated data processing.
Solution Approach 2:
The patent merges the pairing results from multiple vehicles into a comprehensive map representation. By combining the candidate pairs identified from different vehicles and resolving conflicts through validation criteria, the system achieves complete pairing coverage while maintaining the efficiency benefits of unified processing through integrated data structures and algorithms.
3Measurement precision
If comprehensive pairing methods are implemented to cover all signals and stop lines, then pairing accuracy is improved, but the computational time and resources required increase
Solution Approach 1:
The patent applies different quality criteria and validation methods to different types of signal-stop line pairs based on their local characteristics. By tailoring the pairing validation to specific local conditions (e.g., intersection types, lane configurations, visibility requirements), the system achieves high pairing accuracy for each local context while avoiding unnecessary computational overhead from applying overly strict criteria uniformly across all cases.
Data Source
AI summary
A server is provided with a probe data acquiring unit that acquires probe data from a plurality of vehicles; a signal information identifying unit that identifies, for a plurality of signals in an intersection, signal information related to respective signals based on the probe data; a stop line information identifying unit that identifies, for a plurality of stop lines in the intersection, stop line information related to respective stop lines based on the probe data; and a pairing unit that identifies, using a method for identifying a plurality of pairing candidates, a combination of the signal and the stop line as a pairing candidate among a plurality of signals in the intersection and a plurality of stop lines in the intersection, and pairs the signal information and the stop line information based on an identification result of the combination.