Vehicle Controller Recognition Zoning for Curved-Route Obstacle Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous driving systems often detect obstacles outside the vehicle's travel route, leading to unnecessary speed control, especially in curved routes where obstacles within the inner region do not affect the vehicle's travel.
Innovation Solution
A vehicle controller that recognizes a state of objects in a sectoral-shaped recognizable area in front of the vehicle, distinguishing between regular and quasi-recognition areas. The controller performs vehicle control based on the recognition result, maintaining a specified distance to obstacles in regular areas and to boundary positions in quasi-recognition areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the recognition area includes the inner region of curved travel route, then the obstacle detection coverage is improved, but unnecessary speed control is triggered when obstacles are detected in areas that do not affect vehicle traveling
Solution Approach 1:
The recognition area is segmented into two distinct zones: a first recognition area (inner region of curved route) and a second recognition area (outer region of curved route). This segmentation allows the system to apply different control strategies to different spatial zones, preventing unnecessary speed control when obstacles are detected in the inner region while maintaining safety in the outer region.
Solution Approach 2:
Different quality standards are applied to different parts of the recognition area. The inner region (first recognition area) uses a more lenient detection criterion that does not trigger speed control, while the outer region (second recognition area) uses a stricter criterion that does trigger speed control. This local differentiation resolves the contradiction between comprehensive detection and efficient traveling.
2Reliability
If speed control is performed based on all detected obstacles in the recognition area, then safety is improved, but vehicle traveling efficiency deteriorates due to unnecessary deceleration
Solution Approach 1:
The recognition area is divided into a first recognition area (inner region) and a second recognition area (outer region). Obstacles in the first area do not trigger speed control, while obstacles in the second area do trigger speed control. This segmentation maintains safety for critical obstacles while avoiding unnecessary time loss from non-critical detections.
Solution Approach 2:
Different safety response qualities are applied to different spatial zones. The inner region adopts a permissive safety policy that allows smoother traveling, while the outer region adopts a conservative safety policy that prioritizes obstacle avoidance. This resolves the contradiction between overall safety and travel efficiency.
3Area of stationary object
If the recognition area is set to cover the entire front side of the vehicle, then the field of view is improved, but false obstacle detection increases in curved route scenarios
Solution Approach 1:
The broad recognition area is segmented into a first recognition area (inner region of curve) and a second recognition area (outer region of curve). This segmentation allows the system to maintain wide coverage while improving precision by distinguishing between relevant and irrelevant obstacles within different zones.
Solution Approach 2:
Different precision standards are applied to different parts of the recognition area. The inner region uses a precision standard that tolerates false positives, while the outer region uses a precision standard that requires higher confidence. This local differentiation maintains comprehensive coverage while reducing overall false detection rates.
Data Source
AI summary
A vehicle controller is configured to execute a recognition process of recognizing a state of an object in a recognizable area set on a front side of a host vehicle, and a driving control process of performing a vehicle control of causing the host vehicle to travel based on a recognition result of the recognition process. The recognizable area includes a regular recognition area and a quasi-recognition area when a band-shaped scheduled traveling area along a travel route on which the host vehicle travels is curved. The driving control process includes performing the vehicle control according to a first driving mode when the recognition result of the recognition process indicates that an obstacle exists in the regular recognition area, and performing the vehicle control according to a second driving mode when the recognition result of the recognition process indicates that an obstacle exists in the quasi-recognition area.