Map Data Alignment Using User-Identified Landmarks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inaccuracies in map data generated by vehicles can lead to incorrect device location determination, hindering operations and posing safety risks, particularly in environments with few computer-recognizable landmarks.
Innovation Solution
A hybrid approach combining automated landmark identification with user input to align sensor data, allowing additional landmarks to be identified, especially in landmark-sparse regions, thereby improving map data accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated landmark identification is used alone, then processing efficiency is improved, but measurement precision deteriorates in landmark-sparse regions
Solution Approach 1:
The patent combines automated landmark identification with user input mechanisms to create a hybrid system. The automated system processes sensor data to identify landmarks efficiently, while user input provides additional landmark positions and corrections, particularly in regions where automated identification fails due to sparse landmarks. This merging of automated and manual approaches resolves the contradiction by maintaining high processing efficiency while improving measurement precision through supplementary user data.
2Measurement precision
If user input is incorporated to identify additional landmarks, then map data accuracy is improved, but device complexity increases
Solution Approach 1:
The patent introduces a confidence metric as an intermediary mechanism that automatically determines when user input should be solicited. The system calculates confidence levels for automatically identified landmarks and only requests user input when confidence falls below a threshold or in landmark-sparse regions. This intermediary filtering mechanism improves map data accuracy by selectively incorporating user input while preventing excessive system complexity through automated decision-making about when human input is necessary.
3Quantity of substance
If multiple sensor data sets from different poses are aligned, then map data completeness is improved, but computational burden increases
Solution Approach 1:
The patent applies local quality optimization by focusing computational alignment efforts on specific regions where landmarks are sparse or confidence is low, rather than uniformly processing all sensor data. The system identifies problematic regions and directs user input requests and alignment computations specifically to those areas, improving map data completeness in critical regions while reducing overall computational burden by avoiding redundant processing in well-mapped areas.
Data Source
AI summary
Techniques are provided comprising receiving first sensor data associated with a first vehicle pose and second sensor data associated with a second vehicle pose. A user input is received via a user interface indicating a position associated with a feature represented in the first sensor data and the second sensor data. Based at least in part on the position associated with the feature, the first vehicle pose, and the second vehicle pose, an alignment between the first sensor data and the second sensor data is determined. Map data is determined based at least in part on the first sensor data, the second sensor data, and the alignment.