Distributed IoT Neural Network Using RFID Iotons
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing approaches to implement artificial neural networks fail to replicate the sophistication and complexity of biological neural networks, lacking sufficient memory, sensory inputs, adaptive connectivity, and the large number of neurons needed for effective performance.
Innovation Solution
An advanced neural network is implemented using an internet-of-things methodology, where ordinary items equipped with RFID tags and DPU form artificial iotons that connect and adapt, providing memory, sensory inputs, and a vast number of neurons, exceeding biological networks in processing power.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional computers are used to implement neural networks, then the system structure is simple and easy to manufacture, but the processing power and complexity are insufficient to replicate biological neural networks
Solution Approach 1:
The patent segments the neural network implementation into distributed autonomous agents (iotons) that operate independently but connect to form the overall network. Each ioton is a simple unit with basic sensory and motor capabilities, while the collective network achieves complex neural network functionality through the interactions of these segmented units.
Solution Approach 2:
The patent transitions from traditional single-dimensional computational models to a multi-dimensional distributed system where iotons operate in spatial, temporal, and organizational dimensions simultaneously. This dimensional expansion enables the system to achieve biological-like complexity without requiring each individual component to be highly complex.
2Productivity
If a large number of neurons are implemented to match biological networks, then the processing power increases, but the memory requirements and system resources become excessive
Solution Approach 1:
The patent merges multiple functions into each ioton unit, combining sensory input processing, memory storage, computation, and motor output capabilities within single autonomous agents. This functional merging reduces the need for separate dedicated memory resources while maintaining high processing power through the distributed network of multi-functional units.
Solution Approach 2:
Each ioton is designed as a self-sufficient autonomous unit that manages its own memory, processing, and decision-making without requiring centralized resource allocation. This self-service capability allows the system to scale to large numbers of neurons without proportionally increasing overall system resource requirements, as each unit operates independently with its own localized resources.
Data Source
AI summary
Disclosed is an improved approach to implement artificial neural networks. According to some approaches, an advanced neural network is implemented using an internet-of-things methodology, in which a large number of ordinary items having RFID technology are utilized as the vast infrastructure of a neural network.