Road Edge Boundary Detection Using Prediction Scores in AV Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous vehicle mapping systems rely on manual detection of road edges, which is time-consuming, inaccurate, and resource-intensive, and fail to efficiently integrate new data or features, leading to incomplete and insufficient road edge boundary representations.
Innovation Solution
A computer-implemented method and system for automated road edge boundary detection using map data analysis, where prediction scores are generated based on sensor data and image parameters to identify and connect road edge elements into a polyline, improving the accuracy and efficiency of road edge identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual detection of road edges is used, then road edge boundaries can be identified, but the process is time-consuming and resource-intensive
Solution Approach 1:
The patent replaces manual mechanical detection with an automated computer vision system that uses image processing algorithms to detect road edge boundaries. The system processes images captured by sensors mounted on the vehicle, automatically identifying road edges through computational methods rather than human operators physically examining and marking boundaries.
Solution Approach 2:
The system enables the mapping process to serve itself by using the vehicle's own sensor data and onboard processing capabilities to automatically generate and update road edge boundary information. The vehicle's navigation system continuously captures images and processes them to maintain current road edge data without requiring external manual intervention.
2Measurement precision
If manual detection of road edges is used, then road edge boundaries can be identified, but the process is resource-intensive
Solution Approach 1:
The patent divides the road edge detection process into distinct segments: image capture by sensors, pre-processing to enhance relevant features, edge detection algorithms to identify boundary locations, and post-processing to refine the boundary representation. This segmentation allows each stage to be optimized independently, reducing overall computational resource requirements while maintaining detection accuracy.
3Loss of information
If existing mapping systems are used, then road edge information can be obtained, but the representation is incomplete and insufficient
Solution Approach 1:
The system continuously captures images and updates road edge boundary information as the vehicle moves along the roadway. Rather than relying on periodic manual updates or static maps, the system maintains continuous operation of sensors and processing algorithms, ensuring that road edge data remains current and complete by constantly adding new observations to the map.
4Productivity
If automated detection is implemented, then mapping efficiency is improved, but system complexity increases
Solution Approach 1:
The patent integrates road edge detection functionality into the vehicle's existing navigation and sensor system, allowing the same hardware components to serve multiple functions. The image processing system not only detects road edges but also contributes to general environmental perception, obstacle detection, and navigation, thereby amortizing the complexity across multiple useful functions rather than adding dedicated specialized equipment.
Data Source
AI summary
Systems, devices, products, apparatuses, and/or methods for generating a road edge boundary for an edge of a road in an AV map for controlling an autonomous vehicle on a roadway by obtaining map data associated with a map of a geographic location including a roadway associated with one or more locations of one or more vehicles in the roadway during one or more traversals of the roadway, determining one or more prediction scores based on the map data, including one or more predictions of whether the plurality of elements include road edge boundary locations, and generating in the map a road edge boundary based on the one or more prediction scores.