Multi-Sensor Object Recognition Reliability via Discrepancy Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional object recognition methods in automated driving systems face challenges in achieving high reliability when integrating detection results from multiple sensors, leading to differing determination results.
Innovation Solution
An object recognition apparatus comprising multiple sensor units, including cameras and LiDAR/Radar systems, calculates differences in detection results and decides reliabilities to enhance recognition accuracy by integrating data from redundant systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple sensor systems are integrated for object determination, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The system divides object determination into multiple independent determination units, each processing data from specific sensor combinations. This segmentation allows complex multi-sensor integration to be broken down into manageable modules, improving measurement precision through redundant verification while controlling device complexity through structured organization.
Solution Approach 2:
A reliability decision unit acts as an intermediary that evaluates and compares determination results from multiple object determination units. This intermediary component coordinates the complex interactions between multiple sensors and determination units, selecting results based on calculated reliabilities, thereby improving object recognition reliability while managing system complexity through centralized coordination.
2Reliability
If multiple determination units are used to integrate sensor data, then object recognition reliability is improved, but determination result consistency deteriorates
Solution Approach 1:
The system implements feedback through reliability calculation and comparison. Each object determination unit's results are evaluated by the reliability decision unit, which calculates reliabilities based on sensor data quality and determination consistency. This feedback mechanism allows the system to select the most reliable determination result, improving overall object recognition reliability while maintaining determination result consistency through systematic selection criteria.
Solution Approach 2:
The reliability decision unit dynamically changes selection parameters based on calculated reliabilities. Instead of using fixed rules for selecting determination results, the system adjusts selection criteria according to real-time reliability assessments of different sensor combinations and determination units, thereby improving object recognition reliability while maintaining consistency through adaptive parameter adjustment.
Data Source
AI summary
An object recognition apparatus for recognizing an object is provided. The apparatus includes first and second object determinations units configured to determine the object based on at least detection results of the object by first and second sensors and at least detection results of the object by third and fourth sensors, respectively; an object recognition unit configured to recognize the object based on a determination result by the first and/or second object determination unit; first and second calculation units configured to calculate a difference between the detection result by the first sensor and the detection result by the second sensor and a difference between the detection results by the third and fourth sensor, respectively; and a reliability decision unit configured to decide reliabilities of the determination results by the object determination units based on calculation results by the calculation units.