Road Boundary Estimation via Lane Marking Offset
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing road boundary detection systems face challenges in accurately estimating the position and geometry of road boundaries due to low contrast between drivable and non-drivable surfaces, irregularities, and adverse conditions like darkness, leading to noise and uncertainties that affect the precision of imminent road departure assessments and automatic vehicle control.
Innovation Solution
A boundary estimation system on-board a vehicle that monitors surroundings, detects lane markings and road boundaries, approximates their geometrical representations, and defines a fictive outer boundary by laterally shifting the lane marking representation based on relative offsets, enhancing the quality of road boundary information and improving risk assessment and control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensors such as cameras are used to detect road boundary, then the system can monitor surroundings and assess road departure risk, but the detection quality is inherently limited due to low contrast, irregularities, and adverse conditions leading to noise and uncertainties
Solution Approach 1:
The patent introduces lane marking detection as an intermediary element. Instead of directly detecting the road boundary which has low contrast and irregularities, the system first detects the lane marking (which has higher contrast and more regular geometry) and then uses the detected lane marking position combined with a predetermined offset to estimate the road boundary position. This intermediary approach transfers the detection task from a difficult target to an easier target while maintaining the ability to assess road departure risk.
Solution Approach 2:
The patent creates a fictive outer boundary by copying the geometrical representation of the lane marking and shifting it laterally by a predetermined offset. This copied and transformed representation serves as a surrogate for the actual road boundary, allowing the system to work with a cleaner, more reliable geometric model while still capturing the essential information needed for road departure assessment.
2Productivity
If the system uses detected road boundary positions directly, then it can assess imminent road departure risk, but the assessment precision is degraded by noise and uncertainties in the detected boundary information
Solution Approach 1:
The system creates a fictive outer boundary by copying the lane marking's geometrical representation and applying a lateral shift. This copied representation eliminates the noise and uncertainties present in direct road boundary detection while preserving the spatial relationship information needed for accurate road departure risk assessment.
Solution Approach 2:
The patent transforms the detection parameter from direct road boundary position to lane marking position plus offset. By changing the detection target to lane markings (which have better detectability) and then transforming the result through a predetermined offset parameter, the system achieves more precise measurements while maintaining the ability to assess road departure risk.
3Adaptability or versatility
If the system attempts to detect irregular road boundaries under adverse conditions, then it can provide comprehensive road boundary information, but the detection quality deteriorates due to darkness, road conditions, and contrast issues
Solution Approach 1:
The patent uses lane markings as an intermediary that is more visible and detectable under adverse conditions. Lane markings typically have higher contrast and more regular geometry than road boundaries, making them detectable in darkness and poor weather. The system detects this intermediary element and uses it to infer the road boundary, thereby maintaining adaptability while improving precision.
Solution Approach 2:
The system changes the detection parameter from road boundary characteristics (low contrast, irregular) to lane marking characteristics (high contrast, regular). This parameter change enables detection under adverse conditions while the predetermined offset transformation maintains the relationship to the actual road boundary, achieving both adaptability and precision.
Data Source
AI summary
A method performed by a boundary estimation system for estimating a boundary of a road on which a vehicle is positioned and which comprises at least a first lane marking. The system monitors surroundings of the vehicle, detects one or more positions of the at least first lane marking, and approximates a geometrical representation of the at least first lane marking based on one or more of the detected positions of the at least first lane marking. The system further detects one or more positions of a road boundary of the road, approximates a relative lateral offset between the geometrical representation of the at least first lane marking and the detected road boundary, and defines a fictive outer boundary of at least a section of the road based on laterally shifting at least a section of the geometrical representation of the at least first lane marking the relative lateral offset.