Sensor Fusion Uncertainty Matching for Vehicle Data Validation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The integration of information from vehicle-to-X communication with environmental sensors is challenging due to differing measurement inaccuracies and varying detection cycles, making it difficult to assign and validate data effectively for reliable vehicle control systems.

Innovation Solution

A method that compares the fuzziness of sensor data from different types of sensors, allowing for the validation and confirmation of information based on matching uncertainty values, thereby optimizing the use of information and reducing computational effort by only assigning information with similar uncertainty levels.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If information from vehicle-to-X communication and environmental sensors is merged, then the reliability of sensor fusion is improved, but the difficulty of assigning and validating data increases due to differing measurement inaccuracies

Engineering Contradiction:
Improvereliability of sensor fusionVSAvoiddifficulty of assigning and validating data
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies parameter changes by comparing uncertainty values (a specific parameter) of sensor data from different sources. When uncertainty values match within a threshold, the system merges information; when they differ significantly, it validates information. This dynamic parameter-based approach resolves the contradiction by adapting the integration strategy based on the actual quality characteristics of the data being processed.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent segments the information integration process into two distinct paths: merging (when uncertainty values match) and validation (when uncertainty values differ). This segmentation allows the system to handle different types of sensor data appropriately, reducing the overall complexity by breaking down the challenging integration task into manageable, condition-based subtasks.

Inventive Principle:
Principle #1Segmentation

2Loss of information

If all sensor information is assigned and merged, then the completeness of information usage is improved, but the computational effort increases significantly

Engineering Contradiction:
Improvecompleteness of information usageVSAvoidcomputational effort
Core Design Contradiction:
Loss of informationVSPower

Solution Approach 1:

The patent applies partial action by selectively merging only those sensor information pairs whose uncertainty values match within a specified threshold. Instead of attempting to merge all possible sensor data combinations, the system performs partial merging operations on qualifying data pairs, thereby reducing computational effort while still achieving comprehensive information usage through the combination of merging and validation paths.

Inventive Principle:
Principle #16Partial or excessive action

3Quantity of substance

If sensor data with different uncertainty values is merged, then the quantity of usable information is improved, but the measurement precision decreases

Engineering Contradiction:
Improvequantity of usable informationVSAvoidmeasurement precision
Core Design Contradiction:
Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent applies dynamics by making the information integration strategy adaptive based on the uncertainty values of the sensor data. The system dynamically switches between merging (when uncertainties match) and validation (when uncertainties differ), allowing it to optimize both the quantity of usable information and the measurement precision depending on the specific characteristics of the data being processed.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP2867093B1Method and system for information usage
Publication Date: 2020.07.29 CONTINENTAL TEVES AG & CO OHG
  • EP2867093B1 patent drawingFigure 1
  • EP2867093B1 patent drawingFigure 2~3

AI summary

The invention relates to a method for information usage, in which by means of sensors of a first and second sensor type first and second sensor data are detected and items of information describing identical objects and/or object attributes are allocated to one another in the first and second sensor data, wherein, furthermore, the first and second sensor data are encumbered by first and second ambiguities and wherein, furthermore values of the second encumbering ambiguities are at least as great as values of the first encumbering ambiguities. The method is characterised in that a comparison of the values of the first and second encumbering ambiguities of items of information describing identical objects and/or object attributes takes place in the first and second sensor data, wherein in the case of values of substantially the same size, the items of information describing identical objects and/or object attributes are allocated to one another in the first and second sensor data and wherein in the case of values which are not of substantially the same size the items of information describing identical objects and/or object attributes are confirmed or discarded in the first sensor data by the second sensor data. The invention further relates to a corresponding system and use thereof.