Lane Boundary Segments from Maps, Trajectories, and Landmarks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Maintaining an autonomous or semi-autonomous vehicle within a lane is challenging due to unclear lane markings and limited or unavailable high-definition map data, which can lead to safety issues.
Innovation Solution
A computing system generates, validates, and aligns candidate segments representing lane boundaries using low-resolution maps, historical vehicle data, and landmark data, ensuring accurate navigation even in the absence of high-definition maps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If only landmarks or standard maps are used to maintain ego-vehicle within a lane, then the system is simple to operate, but the reliability of lane boundary detection deteriorates when lane markings are unclear or absent
Solution Approach 1:
The patent combines multiple data sources including low-resolution maps, historical trajectory data, and landmark data to generate candidate segments representing lane boundaries. This merging of multiple information sources improves the reliability of lane boundary detection without requiring a single complex high-definition map system
Solution Approach 2:
The system uses a multi-functional approach where the same computing system processes and integrates different types of data (map data, historical trajectories, landmark information) to perform lane boundary detection, making the system adaptable to various conditions without requiring specialized hardware for each data type
2Measurement precision
If high-definition map data is used to provide precise lane information, then the manufacturing precision of map data improves, but the loss of time and resources to maintain and update the maps increases
Solution Approach 1:
The system pre-processes and stores historical trajectory data and landmark information in advance, so that when lane boundary detection is needed, the computing system can quickly integrate these pre-prepared data sources without requiring time-consuming real-time processing or frequent map updates
Solution Approach 2:
The system uses the ego-vehicle's own historical trajectory data to help determine lane boundaries, eliminating the need for external high-definition map updates. The vehicle essentially maps its own lane boundaries through accumulated operational data
3Reliability
If multiple data sources are integrated to generate candidate segments, then the reliability of lane boundary detection improves, but the computational complexity and data processing requirements increase
Solution Approach 1:
The patent segments the lane boundary detection problem into distinct components: generating candidate segments from low-resolution maps, generating additional candidate segments from historical trajectory data, validating these candidates against landmark data, and selecting the best matching segments. This segmentation makes the complex processing manageable and systematic
Data Source
AI summary
A system includes one or more processors that obtain data from one or more sources. The sources include sources any of low resolution map data, historical data of an ego-vehicle position, and landmark data. The processors generate candidate segments indicative of potential lane boundaries, the candidate segments include any of a first candidate segment generated from the low resolution map data, a second candidate segment generated from the historical data of the ego-vehicle position, and a third candidate segment generated from the landmark data. The processors validate the generated candidate segments, align the validated and generated one or more candidate segments, and construct a segment based on the aligned, validated, and generated one or more candidate segments, the segment defining a lane boundary. The segment is used to compute a navigation path for the ego-vehicle to operate the ego-vehicle within the lane boundary defined by the segment.