Vehicle Object Detection With Corner-Based Coverage Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor systems for vehicles struggle to efficiently and reliably capture objects in their surroundings, particularly in urban environments, due to obscuration issues that lead to inaccurate detection and potential collisions.
Innovation Solution
A method that determines the corners of objects using sensor data, assesses their relative orientation, and checks for obscuration by other objects, allowing for efficient determination of capturable regions and tracking, reducing computing effort.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If ray tracking model is used separately for each vicinity sensor to detect objects, then object detection capability is improved, but computing effort increases significantly
Solution Approach 1:
The patent merges the ray tracking processing for multiple vicinity sensors into a single unified process. Instead of performing separate ray tracking for each sensor, the system combines the sensor data and performs one integrated ray tracking operation that accounts for all sensors simultaneously. This reduces the computational complexity from O(n²) to O(n) where n is the number of sensors, while maintaining the ability to detect objects obscured by other objects.
2Measurement precision
If probability of existence is increased for objects in registration region, then detection sensitivity is improved, but obscured objects are falsely identified as faulty measurements
Solution Approach 1:
The patent converts the harmful effect of obscuration into a beneficial feature by using the ray tracking model to explicitly calculate which objects are obscured by others. Instead of treating obscured objects as faulty measurements, the system uses the obscuration information to maintain their detection status. The ray tracking rays that pass through or around obscuring objects provide the necessary measurement data to confirm the existence of obscured objects, transforming what was previously a source of errors into a reliable detection mechanism.
3Reliability
If object tracking is maintained during obscuration, then collision prevention is improved, but system complexity increases
Solution Approach 1:
The patent performs preliminary ray tracking calculations to determine in advance which objects may be obscured by others in the scene. By pre-calculating the obscuration relationships and maintaining tracking for objects that are predicted to be obscured, the system avoids the need for complex real-time adjustments when obscuration occurs. The tracking algorithm is prepared in advance to handle obscured objects appropriately, reducing the computational burden during actual operation while maintaining collision prevention capability.
Data Source
AI summary
A method for detecting objects in the surroundings of a vehicle includes receiving sensor data which describe the objects in the surroundings from a surroundings sensor of the vehicle; determining each of the corners of the objects, with the corners describing outer delimitations of each object; determining a relative position of each of the corners to the surroundings sensor; sorting the determined corners in a predetermined angle direction; checking, for each of the corners in the angle direction, whether the corners are covered by another of the objects, on the basis of the relative position of each of the corners to the surroundings sensor; and determining regions of each of the objects that can be detected by the surroundings sensor, on the basis of checking the corners.