Autonomous Vehicle Occlusion Avoidance for Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous driving technologies face challenges in acquiring accurate information about surrounding objects due to occlusion by other objects or environments, which can hinder smooth and safe vehicle operation.
Innovation Solution
A vehicle system comprising a camera, sensor unit, and controller that identifies and separates occlusion areas from identification areas by adjusting velocity, steering, and distance to prevent overlaps, allowing for accurate object recognition and prediction of movement paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the vehicle maintains a fixed position and orientation, then the system structure is simple, but the identification area overlaps with occlusion area causing inaccurate object recognition
Solution Approach 1:
The patent applies dynamics by making the vehicle's position and orientation adjustable rather than fixed. The controller dynamically changes the vehicle's longitudinal position, lateral position, and/or orientation based on detected occlusion areas, allowing the identification area to be separated from occlusion areas and improve object recognition accuracy.
Solution Approach 2:
The patent changes the parameters of the vehicle's spatial configuration (longitudinal position, lateral position, orientation) to resolve the contradiction. By adjusting these parameters, the system separates the identification area from occlusion areas, improving measurement precision without requiring complex additional hardware.
2Measurement precision
If the vehicle changes position and orientation frequently to avoid occlusion, then object recognition accuracy improves, but the control system complexity increases
Solution Approach 1:
The system performs self-service by automatically detecting occlusion areas and autonomously adjusting its own position and orientation to separate the identification area from occlusion areas. This self-adjusting capability improves information acquisition accuracy while maintaining operational simplicity, as the vehicle handles the complexity internally without requiring complex external control operations.
3Measurement precision
If the vehicle adjusts velocity and position to separate identification area from occlusion area, then object recognition accuracy improves, but the driving time increases
Solution Approach 1:
The patent applies partial action by making selective adjustments to velocity and position only when and to the extent necessary to separate the identification area from occlusion areas. Rather than continuous adjustments, the system performs targeted partial adjustments that achieve the necessary separation while minimizing additional driving time and maintaining efficient progression toward the destination.
Data Source
AI summary
Provided is a vehicle capable of performing safe and efficient autonomous driving by acquiring accurate information about surrounding objects based on behavior control of the vehicle. The vehicle includes: a camera configured to acquire a surrounding image of the vehicle which includes one or more objects; at least one sensor configured to acquire position data of the one or more objects; and a controller configured to identify identification areas respectively corresponding to the one or more objects based on the surrounding image and the position data, and if a part of the identification area overlaps an occlusion area, control the vehicle to separate the identification area and the occlusion area from each other, to perform autonomous driving.