Lane Marking Estimation Using Roadside Objects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional driving support apparatuses struggle to accurately estimate lane markings when road width calculation parameters and data are inaccurate, leading to potential inaccuracies in detecting lane positions, especially when roadside objects pose a collision risk.
Innovation Solution
A driving support apparatus that includes a white line detection part, a roadside object detection part, and a white line estimation part to estimate lane markings based on detected roadside objects, allowing for proper output of driving support information to prevent lane departures, with features like virtual lane marking estimation and correction based on object categories and collision risks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If lane marking position is estimated based on road width calculation parameters and detected lane marking, then driving support information can be output when one lane marking is detected, but the position estimation becomes inaccurate when road width data or calculation parameters are not accurate
Solution Approach 1:
The patent introduces roadside objects (guardrails, curbs, sidewalks) as intermediary reference points to estimate lane marking positions. Instead of directly relying on potentially inaccurate road width parameters, the system detects roadside objects and calculates lane marking positions based on the spatial relationship between the detected lane marking and the roadside object, thereby improving estimation reliability and precision
Solution Approach 2:
The patent replaces the traditional method of estimating lane marking positions using road width calculation parameters with a vision-based method that detects roadside objects and uses their spatial relationships. This substitution of the estimation mechanism allows the system to achieve more accurate position estimation by leveraging visual features of roadside objects rather than relying on potentially inaccurate pre-stored road width data
2Ease of operation
If lane marking detection is used to prevent lane departure, then driving support can be provided, but inaccurate lane marking detection leads to improper driving support information
Solution Approach 1:
The patent uses roadside objects as intermediary reference points to improve the reliability of lane marking position estimation. By detecting roadside objects and using their spatial relationship with the detected lane marking, the system generates more reliable driving support information, ensuring that lane departure prevention guidance is based on accurate position data
Solution Approach 2:
The system continuously monitors the spatial relationship between detected lane markings and roadside objects, and uses this feedback to refine lane marking position estimates. This feedback mechanism ensures that driving support information remains reliable even when initial lane marking detections are uncertain or when road conditions change
Data Source
AI summary
A driving support ECU comprises: a roadside object detection section for detecting, in a case where a lane marking WR on at least one of the right side and left side is detected by a white line detection section and another lane marking WL is not detected, a roadside object GL on the other of the right side and left side; a white line estimation part for estimating a position of a lane marking VL1 on the other of the right side and the left side based on the detected roadside object GL; and an information outputting section for determining a departure from the lane in which the vehicle VC is running, based on the estimated position of the lane marking VL1 on the other of the right side and left side.