Rear Cross-Traffic Detection With Reflection-Aware False Alert Filtering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional rear cross traffic alert systems often generate false warnings due to reflections from other objects, leading to unnecessary alarms when detecting non-existent target objects behind and lateral to the subject vehicle.

Innovation Solution

A vehicle object detection system that evaluates the presence of multiple stationary objects between the subject vehicle and the detected target, the duration of detection, reflection strength, and position within a specified area to differentiate between actual and reflected objects, suppressing warnings for mirror-like objects that do not meet these criteria.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the radar infrastructure detects objects in the detection area, then the system can identify potential target objects behind and lateral of the subject vehicle, but false warnings are generated due to reflections from other objects

Engineering Contradiction:
Improvedetection accuracyVSAvoidfalse alarms
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The system continuously monitors detection results over time and uses this feedback to distinguish between real objects and reflections. By analyzing whether detected objects persist across multiple detection cycles, the system can identify false detections caused by reflections and suppress corresponding false alarms.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent introduces intermediate evaluation criteria between raw detection data and warning output. These criteria include checking detection duration, position consistency, and presence of multiple stationary objects between the subject vehicle and detected target, serving as mediators to filter out false detections before generating warnings.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Object-generated harmful factors

If the system suppresses warnings for detected objects, then false alarms are reduced, but actual threats may be missed

Engineering Contradiction:
Improvefalse alarmsVSAvoidthreat detection capability
Core Design Contradiction:
Object-generated harmful factorsVSReliability

Solution Approach 1:

The system applies different evaluation criteria to different aspects of object detection. Instead of uniformly suppressing or alerting on all detections, it locally evaluates each detection based on specific characteristics such as detection duration, position, and environmental context, allowing selective suppression of false alarms while maintaining alertness to real threats.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system changes detection parameters dynamically based on environmental conditions and detection history. By monitoring parameters like detection time period, object position, and reflection strength, the system adjusts its warning threshold to maintain high reliability while reducing false alarms.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If multiple detection criteria are evaluated, then false detections are filtered more accurately, but the system complexity increases

Engineering Contradiction:
Improveobject identification accuracyVSAvoidjudging system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The judging system is segmented into multiple independent evaluation modules, each responsible for checking a specific criterion (detection duration, position validation, stationary object detection). This modular segmentation allows the system to achieve high measurement precision through multiple checks while managing complexity through organized, independent evaluation functions.

Inventive Principle:
Principle #1Segmentation

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

Effectively reduces false alarms by accurately identifying and suppressing warnings for non-existent target objects caused by reflections, ensuring only genuine threats trigger alerts.

Implementation Method 1

the respective rear cross traffic alert (RCTA) systems use a radar infrastructure like radar sensors for detecting vehicles or objects

Methodology Applied
Scientific EffectRadar: Radar

Implementation Method 2

a target object detected in the detection area is in fact a false detection caused by reflections on other objects

Methodology Applied
Scientific EffectReflection: Reflection

Data Source

PatentEP4358064A1Vehicle object detection system and method for detecting a target object in a detection area located behind and lateral of a subject vehicle
Publication Date: 2024.04.24 AUMOVIO AUTONOMOUS MOBILITY GERMANY GMBH
  • EP4358064A1 patent drawingFigure 1
  • EP4358064A1 patent drawingFigure 2
  • EP4358064A1 patent drawingFigure 3

AI summary

The invention relates to a vehicle object detection system for detecting a target object in a detection area located behind and/or lateral of a subject vehicle, the vehicle object detection system comprising: judging means configured to generate a judgement on whether a target object which has been detected in the detection area is an alert object, and to output the judgement to warning means which is configured to output a warning to a driver of the subject vehicle based on the judgement indicating that the target object is an alert object, wherein the judging means is configured to generate the judgement based on: one or more positions of one or more other objects other than the target object being detected to be located between the target object and the subject vehicle, and a time period during which the target object is detected, being smaller than a specified period of time. Moreover, the invention relates to a corresponding computer-implemented method for detecting a target object in a detection area located behind and/or lateral of a subject vehicle, particularly for controlling a vehicle object detection system according to the invention.