Lane-Keeping Steering Control Using Reliability-Guided Lane Prediction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Lane keeping assist systems (LKAS) face performance variability due to weather, illumination, and road conditions, leading to inaccurate lane detection and potential loss of lane keeping function, especially at intersections or road construction sites.

Innovation Solution

A lane keeping assist system that incorporates a camera, lane information generator, image storage, and a neural network learning device, which uses convolutional neural networks (CNNs) and long short-term memory (LSTM) to generate and predict lane information, and a steering controller that selects reliable lane information for steering based on image reliability, ensuring stable lane detection and vehicle control.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If camera-based lane detection is used, then the system can detect lane information, but detection accuracy deteriorates under adverse weather, illumination, or road conditions

Engineering Contradiction:
Improvelane detection accuracyVSAvoidlane detection reliability
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

A neural network learning device acts as an intermediary between the camera-based lane information generator and the steering controller. This intermediary processes and validates lane information, using stored historical image data to correct or supplement current frame analysis, thereby maintaining reliable lane detection even when camera-based detection accuracy deteriorates due to weather, illumination, or road conditions

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs preliminary actions by storing image information in advance in an image storage device. This pre-stored data serves as a reference base that can be recalled and utilized when current lane detection becomes unreliable, allowing the system to maintain accurate lane information without real-time camera dependency under adverse conditions

Inventive Principle:
Principle #10Preliminary action

2Speed

If the system relies solely on current camera images, then real-time response is maintained, but lane keeping function is lost when lane information becomes temporarily inaccurate

Engineering Contradiction:
Improveresponse speedVSAvoidlane keeping function stability
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

The system ensures continuity of useful action by continuously storing image information in the image storage device and continuously monitoring detection reliability. When reliability drops, the system seamlessly transitions to using pre-stored image information, ensuring uninterrupted and reliable lane keeping function without loss of stability, while maintaining real-time response through the steering controller's immediate processing of validated lane information

Inventive Principle:
Principle #20Continuity of useful action

3Measurement precision

If neural network learning is activated for all images, then lane detection accuracy improves, but processing time and computational load increase

Engineering Contradiction:
Improvelane detection accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system applies partial action by activating the neural network learning device only when necessary - specifically when the lane information generator determines that current image information has low reliability. This selective activation maintains high lane detection accuracy when needed while avoiding unnecessary processing time and computational load during normal operating conditions

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The system dynamically adjusts its processing approach based on real-time reliability assessment. The steering controller and lane information generator continuously evaluate image quality and detection confidence, switching between standard processing and enhanced neural network processing modes. This dynamic adaptation optimizes the balance between detection accuracy and processing time according to actual operating conditions

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3944141B1Lane keeping assist system of vehicle and lane keeping method using the same
Publication Date: 2024.10.16 HYUNDAI MOBIS CO LTD
  • EP3944141B1 patent drawingFigure 1
  • EP3944141B1 patent drawingFigure 2

AI summary

A lane keeping assist system includes: a camera configured to provide an image around a vehicle as image information; a lane information generator configured to generate image reliability information and first lane information, based on the image information; an image storage configured to store the image information for each predetermined time among predetermined times; a neural network learning device configured to generate second lane information based on the image reliability information and the image information stored for each predetermined time; and a steering controller configured to select either one of the first lane information and the second lane information as lane information, based on the image reliability information, and generate steering information based on the selected lane information.