Collision Avoidance Steering with Partial Lane Boundary Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional collision avoidance assist systems fail to perform automatic steering control when a target object is positioned on one of the dividing lines of a traveling lane, due to incomplete detection of lane boundaries, leading to potential collisions.
Innovation Solution
A collision avoidance assist apparatus that includes an object information acquiring system, a steered angle actuator, and a control unit capable of determining a collision index value and identifying a target object, which assesses the existence of an avoidance route by considering the detection of lane boundaries and the movement direction of the target object relative to the lane, enabling proper automatic steering control even when one lane boundary is partially undetected.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the conventional apparatus requires both left and right dividing lines to be detected as continuous lines to determine avoidance space, then the reliability of lane boundary detection is improved, but the productivity of collision avoidance control is deteriorated due to false negative detections when target objects are positioned on dividing lines
Solution Approach 1:
The patent applies local quality by making the detection continuity requirement selective rather than uniform. Specifically, when a target object is detected on or near a dividing line, the system relaxes the continuity requirement for that specific dividing line segment while maintaining strict requirements for other segments. This allows the system to locally adapt the detection criteria based on the presence and position of target objects, preventing false negative decisions while maintaining overall detection reliability.
Solution Approach 2:
The patent implements dynamics by making the avoidance space determination criteria flexible and adaptive rather than static. The system dynamically adjusts the continuity requirement for dividing lines based on real-time detection of target objects. When a pedestrian or other target is detected on a dividing line, the system automatically modifies the detection thresholds and continuity requirements for that specific region, enabling the collision avoidance control to be executed appropriately without being constrained by rigid pre-set criteria.
2Measurement precision
If the conventional apparatus uses strict continuous line detection for both dividing lines, then the measurement precision of lane boundary detection is improved, but the adaptability to complex road scenarios is deteriorated
Solution Approach 1:
The patent applies local quality by implementing position-dependent detection criteria. Different detection precision requirements are applied to different spatial regions and contexts. When target objects are detected on dividing lines, the system applies relaxed continuity requirements specifically to those dividing line segments, while maintaining high precision requirements for other regions. This localized adaptation enables the system to handle diverse scenarios such as pedestrians on road edges, vehicles in adjacent lanes, and complex road markings without compromising overall detection accuracy.
Solution Approach 2:
The patent utilizes parameter changes by dynamically modifying detection parameters such as continuity thresholds, detection sensitivity, and line segment weighting based on the detected scene configuration. When target objects are positioned on or near dividing lines, the system changes the parameter values for line detection in those regions, allowing for partial or intermittent line detection to be accepted as valid. This parameter adaptation enables the system to maintain measurement precision in critical regions while improving adaptability to complex and variable road scenarios.
Data Source
AI summary
Collision avoidance assist apparatus comprises an object information acquiring apparatus for acquiring, as object information, information on a three-dimensional object as well as dividing lines defining a travelling lane in front of a vehicle, and a control unit configured to, when a collision possibility with a three-dimensional object is high, identify this object as a target object. When one of the dividing lines has been detected without any discontinuation in a predetermined region as well as a part of an other dividing line has not been detected therein, if the control unit determines, based on the object information, that a second condition which is satisfied when the target object is positioned on the other dividing line and is moving in a same direction as an extending direction is satisfied and an avoidance route exists, the control unit performs automatic steering control so that the vehicle travels along the avoidance route.