Dual Neural Network Recognition Apparatus for Adaptive Learning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing vehicle AI systems lack an efficient method to collect and adapt to changing usage environment data, such as new automobile designs and dangers, which is essential for real-time recognition and judgment during travel.

Innovation Solution

A recognition apparatus comprising a first and second neural network with different structures, a comparison unit, and a communication unit that wirelessly transmits data to a host system when output differences exceed a predetermined standard, enabling efficient collection and adaptation of learning data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple neural networks with different structures are used to collect learning data, then the adaptability to changing usage environments is improved, but the device complexity increases

Engineering Contradiction:

Improveadaptability to changing usage environmentVSAvoidcomplexity of neural network structure

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system segments the learning data collection function by dividing it into multiple neural networks with different structures (first neural network and second neural network). Each neural network processes data differently, and their output results are compared to identify learning data. This segmentation allows the system to adapt to changing environments through diverse processing approaches while managing complexity through functional division.

Inventive Principle:

Principle #1Segmentation

2Measurement precision

If learning data is collected through comprehensive analysis, then the measurement precision of usage environment changes is improved, but the processing time increases

Engineering Contradiction:

Improveprecision of learning data identificationVSAvoidtime for data collection and processing

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

Instead of comprehensively analyzing all data through all neural networks, the system performs partial action by comparing output results from different neural networks and selectively identifying learning data based on comparison results. This approach achieves sufficient precision for identifying meaningful learning data while significantly reducing processing time by avoiding exhaustive analysis of all possible data aspects.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS11341398B2Recognition apparatus and learning system using neural networks

Publication Date: 2022.05.24 HITACHI LTD

AI summary

Learning data of a usage environment can be efficiently collected. A recognition apparatus includes: a first neural network configured to receive input of data; a second neural network configured to receive input of the data, the second neural network having a different structure from a structure of the first neural network; a comparison unit configured to compare a first output result of the first neural network and a second output result of the second neural network; and a communication unit configured to wirelessly transmit the data to a host system configured to learn the data when a comparison result between the first output result and the second output result is different by a predetermined standard or more.