AIoT processing techniques for efficiently extracting and analyzing large amounts of IoT information

다국어 초록 (Multilingual Abstract)

Background/Objectives: Recently, as the usability of IoT devices increases, the kinds of information processed by IoT devices are also diversifying. For this reason, research to ensure the high-speed processing and integrity of information generated from IoT devices continues steadily. In particular, research using various hash techniques is actively underway to minimize IoT information errors. Methods/Statistical analysis: In this paper, instead of sending IoT information directly to the server (data center), we propose an AIoT technique that allows AIoT to pre-analyze and delivers only important information. Findings: The proposed technique allows easy control of IoT information operation after analyzing patterns of information collected from IoT devices to extract and analyze large amounts of IoT information. Furthermore, the proposed technique minimizes network delay as well as minimizes server (data center) processing and analysis time by reducing network traffic that can occur when information collected from numerous IoT devices is delivered to the server (data center). Improvements/Applications: AIoT, IoT information, Extract and Analysis, Artificial, Depp Learning, hash techniques Performance evaluation compared the storage efficiency of servers and the number of transactions of IoT information sent and received per second between IoT devices and AIoT devices while varying the hash length of IoT information from 16 bits to 128 bits.

목차 (Table of Contents)

• Abstract1
• I. INTRODUCTION
• II. BACKGROUND
• III. DATA EXTRACTION TECHNIQUES USING AIOT DEVICES
• A. Overview
• B. Link IoT information in AIoT devices using hierarchical subnets
• C. Scheduling of IoT information
• D. Calculate the importance of IoT information
• IV. PERFORMANCE EVALUATION
• A. Simulation Environment Settings
• B. Evaluation
• V. CONCLUSION
• REFERENCES