Incremental Vehicle Map Generation for Occluded Multi-Lane Roads
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing map generation systems for vehicle positioning face challenges in precision when lanes are hidden by other vehicles or when there are multiple lanes on a road, leading to impaired map information and potential safety issues.
Innovation Solution
A map generation apparatus equipped with sensors and a microprocessor that detects the exterior environment, recognizes surrounding features, generates maps, estimates vehicle position, determines map completion, and stores relevant information, allowing for incremental map generation and updating as the vehicle travels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If map generation is performed using captured images from a camera mounted on a vehicle, then the map can be created for vehicle positioning, but precision of map information is impaired when lanes are hidden by other vehicles or when there are multiple driving lanes on a road
Solution Approach 1:
The map generation process is segmented into multiple passes, with each pass focusing on specific driving lanes. The system divides the road into multiple lanes and generates map information for each lane separately, allowing complete mapping even when some lanes are occluded during any single pass.
Solution Approach 2:
The system performs preliminary map generation for visible lanes first, stores the incomplete map information, and then performs subsequent passes to fill in the occluded lanes. This preliminary action allows the system to progressively complete the map without requiring all lanes to be visible simultaneously.
2Measurement precision
If map generation is performed using captured images from a camera mounted on a vehicle, then the map can be created for vehicle positioning, but precision of map information is impaired when there are many driving lanes on a road that cannot be captured at one time
Solution Approach 1:
The wide road with multiple lanes is segmented into individual lane sections, each mapped separately. The system identifies and processes each lane independently across multiple capturing passes, ensuring complete coverage of all lanes even when the road width exceeds the camera's single-frame capture capability.
Solution Approach 2:
The map generation operates continuously across multiple vehicle passes. Each pass contributes additional lane information to the accumulating map data, maintaining continuous progress toward complete map coverage without requiring all lanes to be captured in a single continuous action.
3Loss of information
If the system stores and revisits incompletion sections to generate complete maps, then map completeness is improved, but the time and number of travels required for map generation increases
Solution Approach 1:
The system extracts and stores only the specific incompletion section information (which lanes are missing) rather than storing complete map data. This allows efficient comparison and targeted updates during subsequent passes, reducing the time required for map completion while maintaining completeness.
Solution Approach 2:
The system implements feedback by storing completion status of each lane section and using this information to guide subsequent map generation passes. The feedback mechanism identifies exactly which sections need updating, preventing redundant processing of already-complete areas and minimizing total generation time.
Data Source
AI summary
A map generation apparatus includes a microprocessor configured to perform: recognizing a surrounding environment based on detection data of a sensor; generating a map based on recognition information; estimating a position of the subject vehicle on the map; determining completion or incompletion of the map; and storing map information. The generating includes: generating the map of a driving section based on the recognition information; storing, in a memory, the map information corresponding to a completion section; and storing, in the memory, section information indicating an incompletion section together with position information of the subject vehicle. The generating further includes: when the subject vehicle travels next time on the incompletion section, generating a map of the incompletion section based on the recognition information; adding map information of the incompletion section to the map information; and rewriting the section information.