Vehicle Collision Detection Using Spatial Uncertainty Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current driver assistance systems for motor vehicles face challenges in reliably detecting potential collisions with objects due to spatial uncertainty in object positioning, which can lead to inaccurate collision predictions.
Innovation Solution
A method that determines a future driving path and receives sensor data to identify an object area with spatial uncertainty, calculating the distance from the driving path to the furthest point of the object area outside the path, and assessing collision probability based on this distance and the portion of the object area within the path, using a control device to output warnings or intervene in steering and braking systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor data is used to determine object position, then object detection capability is improved, but spatial uncertainty in positioning increases
Solution Approach 1:
The object area is divided into two segments: a first area inside the driving path and a second area outside the driving path. This segmentation allows the system to separately evaluate the collision risk by analyzing the proportion of the object area that overlaps with the driving path, thereby resolving the contradiction between detection capability and prediction reliability.
2Measurement precision
If object area with spatial uncertainty is considered, then detection accuracy is improved, but collision prediction reliability deteriorates
Solution Approach 1:
The evaluation of collision risk is made local by focusing on the proportion of the object area that lies inside the driving path rather than treating the entire object area uniformly. This local quality approach allows the system to account for spatial uncertainty while maintaining reliable collision predictions by concentrating on the relevant portion of the object area.
3Measurement precision
If distance to furthest point outside driving path is calculated, then spatial uncertainty analysis is improved, but collision detection reliability is enhanced
Solution Approach 1:
The system transitions from a binary collision detection approach to a probabilistic approach by introducing a new dimension of evaluation: the proportion of object area inside the driving path. This dimensional change allows the system to incorporate spatial uncertainty analysis while enhancing collision detection reliability through a more nuanced assessment.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
The invention relates to a method for identifying a possible collision between a motor vehicle (1) and an object (8), in which a driving path (10) is determined, wherein the driving path (10) describes an area in which the motor vehicle (1) is moved during a future movement, sensor data that describe the object (8) are received from a sensor (4), the received sensor data are taken as a basis for determining an object area (9) that describes a position of the object (8) including a spatial uncertainty, and if part of the object area (9) is arranged within the driving path (10) then the possible collision is identified on the basis of a relative situation of the driving path (10) in relation to the object area (9), wherein a distance (d) between the driving path (10) and a point (P) of the object area (9) that is arranged outside the driving path (10) is determined and the possible collision is identified on the basis of the determined distance (d).