Lane Change Detection Using Real Horizontal Distance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driver assistance systems for longitudinal guidance struggle to accurately detect vehicles cutting in or out, particularly in close proximity, due to uncertainties in lane assignment and reliance on virtual driving paths that do not accurately represent the actual lane.
Innovation Solution
The method involves detecting lane markings and image points from two-dimensional camera images to determine the real horizontal distance between vehicles and lane markings, using image analysis algorithms to assess whether a vehicle is cutting in or out, thereby utilizing actual lane markings for more precise detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a virtual driving path is used to determine lane position, then the system can operate without additional sensors, but the detection accuracy of cutting-in vehicles deteriorates, especially in close range
Solution Approach 1:
The patent combines camera-based image analysis with radar distance measurement to determine lateral position. The camera captures lane markings and vehicles, while radar provides accurate distance data. By merging these two sensor types, the system achieves both acceptable complexity and high measurement precision for detecting cutting-in vehicles.
Solution Approach 2:
The patent introduces camera-based lane marking detection as an intermediary to bridge the gap between virtual driving paths and actual lane positions. The camera captures real-time images of lane markings, and image processing algorithms extract actual lane boundary positions, serving as a mediator that corrects the inaccuracies of virtual path-based detection.
2Reliability
If radar sensors are used for distance measurement, then the system can detect vehicles ahead, but the lateral position determination remains uncertain
Solution Approach 1:
The patent merges radar distance measurement with camera-based lateral position detection. Radar provides reliable distance to the vehicle ahead, while the camera simultaneously captures lane markings to determine actual lateral position. This combination allows the system to achieve both reliable vehicle detection and precise lateral positioning.
Solution Approach 2:
The patent transitions from one-dimensional radar distance measurement to two-dimensional spatial positioning by incorporating camera images. The camera provides lateral dimension information through lane marking detection, complementing the radar's longitudinal distance measurement and enabling accurate determination of both distance and lateral position.
3Area of stationary object
If the detection range is extended to close proximity, then more vehicles can be monitored, but the uncertainty in lane assignment increases
Solution Approach 1:
The patent uses camera-based lane marking detection as an intermediary that remains effective at close ranges. While radar performance degrades at very close distances, the camera continuously captures lane markings even when vehicles are nearby, providing accurate lateral position information that maintains lane assignment precision across all detection ranges.
Solution Approach 2:
The patent changes the detection parameter from radar-based distance measurement to camera-based visual detection for lateral positioning. This parameter change allows the system to maintain high precision in lane assignment at close ranges, where camera resolution is sufficient to clearly distinguish lane markings, whereas radar-based lateral estimation becomes increasingly uncertain.
Data Source
Figure 1~2
Figure 3~4
AI summary
The invention relates to a method for detecting a lane-change or lane-change maneuver by a vehicle ahead of the driver's own vehicle, and to a corresponding vehicle. The driver's own vehicle comprises a driver assistance system for longitudinal guidance and a camera that captures the area in front of the vehicle in the form of two-dimensional images. To detect a lane-change or lane-change maneuver, the distance between a lane marking (12) delimiting the driver's own lane and a pixel (15) or image segment defined by a vehicle ahead is detected in a captured image, and the actual horizontal distance is determined.