Virtual Lane Generation for LKAS Detection Failure
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Lane keeping assistance systems (LKAS) face challenges in accurately detecting lanes due to environmental factors like guardrails, snow, and rain, leading to nonrecognition or misrecognition, which can cause vehicle deviation and safety issues.
Innovation Solution
An apparatus and method for generating a virtual lane by judging lane detection using offset, heading angle, and lane accuracy, with a lane detection judging unit determining normal detection and a lane correction/estimation unit generating a virtual lane based on side lane information, enabling correction or estimation of undetected lanes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If lane detection is performed using camera-based LKAS, then lane keeping assistance can be provided, but nonrecognition or misrecognition of lanes occurs due to environmental factors such as guardrails, snow, and rain
Solution Approach 1:
The patent introduces an intermediary virtual lane generation module that creates virtual lane information based on vehicle motion characteristics when real lane detection fails. This intermediary system acts as a mediator between the camera-based detection system and the lane keeping control, providing reliable lane information even when environmental factors cause nonrecognition or misrecognition of actual lanes.
Solution Approach 2:
The system performs preliminary judgment of lane detection normality before relying on detected lane information. By pre-assessing whether the camera has correctly detected lanes and predicting vehicle motion characteristics in advance, the system can switch to virtual lane generation when detection is unreliable, preventing the harmful effects of environmental interference.
2Reliability
If virtual lane generation is implemented to correct undetected lanes, then lane recognition reliability improves, but device complexity increases
Solution Approach 1:
The patent segments the lane detection system into distinct functional modules: a lane detection judging unit that assesses detection normality, a virtual lane generation unit that creates virtual lanes when needed, and a lane keeping control unit that uses the generated information. This segmentation allows the complex virtual lane generation functionality to be added without overwhelming the overall system architecture, as each module has a specific, well-defined responsibility.
Solution Approach 2:
The system implements virtual lane generation only partially - specifically when lane detection is judged to be abnormal. Rather than always generating virtual lanes, the system selectively activates this functionality based on detection reliability assessment, reducing the overall computational burden and system complexity while maintaining high lane recognition reliability when needed.
3Measurement precision
If lane correction and estimation are performed using vehicle motion characteristics, then accuracy of undetected lane information improves, but calculation complexity increases
Solution Approach 1:
The patent changes the parameters used for lane information generation from direct camera-based lane detection to vehicle motion characteristics such as yaw rate, steering angle, and vehicle speed. By using these alternative parameters that are already available from vehicle sensors, the system achieves accurate lane position estimation without requiring complex image processing and lane detection algorithms.
Data Source
AI summary
Disclosed are an apparatus and a method for generating a virtual lane, which correct or estimate an undetected lane depending on detection of a lane at each side and a system for controlling lane keeping of a vehicle with the apparatus. The apparatus for generating a virtual lane according to the present invention includes: a lane detection judging unit configured to judge whether both lanes of a front road are normally detected based on a front image; a lane correction unit configured to generate a virtual lane by correcting one lane based on the other lane when it is judged that the one lane of both lanes is not normally detected; and a lane estimation unit configured to generate the virtual lane by estimating both lanes based on previously detected lane information when it is judged that either lane is not normally detected.