Reliability Determination for Autonomous Vehicle Object Identification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing systems for automatic vehicle operation lack a reliable method to determine the reliability degree of object identification information, which is crucial for accurate decision-making in autonomous driving scenarios.

Innovation Solution

A processing device that combines sensor information, environmental data, and individual sensor characteristics to determine the reliability degree of object identification information, using a measurement time recording unit, reliability degree determination unit, and composition unit to generate and output comprehensive reliability and object identification data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If sensor information is used to generate object identification information, then automatic operation function is enabled, but reliability of object identification information cannot be determined

Engineering Contradiction:
Improveautomatic operation functionVSAvoidreliability degree of object identification information
Core Design Contradiction:
Extent of automationVSReliability

Solution Approach 1:

The patent segments the reliability determination into multiple independent components: sensor information quality, environmental condition assessment, and object identification information reliability. Each component is evaluated separately and then integrated to provide comprehensive reliability assessment.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system implements feedback mechanisms by evaluating sensor information quality and environmental conditions, then using this feedback to determine and transmit reliability degrees back to the automatic operation system for decision-making.

Inventive Principle:
Principle #23Feedback

2Measurement precision

If sensor information quality is evaluated, then data transmission control is improved, but object identification information reliability determination is not addressed

Engineering Contradiction:
Improvesensor information quality evaluationVSAvoidobject identification information reliability
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent merges sensor information quality evaluation with object identification information reliability determination into a unified assessment system. Environmental condition assessment is also integrated to provide comprehensive reliability evaluation.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The system adds a new dimension of assessment by evaluating environmental conditions (weather, road conditions, lighting) in addition to sensor information quality, creating a multi-dimensional reliability assessment framework.

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

3Measurement precision

If multiple information pieces are combined for object identification, then accuracy is improved, but complexity of reliability determination increases

Engineering Contradiction:
Improveobject identification accuracyVSAvoidreliability determination complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the complexity by dividing reliability determination into distinct modules: sensor information evaluation, environmental condition assessment, and integrated reliability determination. This modular approach manages complexity while maintaining comprehensive assessment.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3605489B1Processing device and method for generating object identification information
Publication Date: 2022.09.07 MITSUBISHI ELECTRIC CORP
  • EP3605489B1 patent drawingFigure 1
  • EP3605489B1 patent drawingFigure 2
  • EP3605489B1 patent drawingFigure 3

AI summary

A processing device (10) according to the present invention includes an object identification unit (103) that generates, based on data measured by sensors (1-1 to 1-n), an object identification information that is information indicative of a state of an object, and a reliability degree determination unit (102) that determines a reliability degree of the object identification information using at least one of individual information of the sensors (1-1 to 1-n), environmental information obtained when the sensors (1-1 to 1-n) have performed the measurement, and time information that is a time at which the sensors (1-1 to 1-n) have performed measurement.