Vehicle Collision Detection Using Spatial Uncertainty Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate 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

VSEngineering 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

Engineering Contradiction:
Improveobject position detectionVSAvoidcollision prediction accuracy
Core Design Contradiction:
Measurement precisionVSReliability

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.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If object area with spatial uncertainty is considered, then detection accuracy is improved, but collision prediction reliability deteriorates

Engineering Contradiction:
Improveobject position descriptionVSAvoidcollision detection reliability
Core Design Contradiction:
Measurement precisionVSReliability

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.

Inventive Principle:
Principle #3Local quality

3Measurement precision

If distance to furthest point outside driving path is calculated, then spatial uncertainty analysis is improved, but collision detection reliability is enhanced

Engineering Contradiction:
Improvespatial uncertainty analysisVSAvoidcollision detection reliability
Core Design Contradiction:
Measurement precisionVSReliability

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.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP3386825B1Method for identifying a possible collision between a motor vehicle and an object by taking into account a spatial uncertainty, control device, driver assistance system and motor vehicle
Publication Date: 2019.09.18 VALEO SCHALTER & SENSOREN GMBH
  • EP3386825B1 patent drawingFigure 1
  • EP3386825B1 patent drawingFigure 2~3
  • EP3386825B1 patent drawingFigure 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).