Janghun Lee (Undergraduate Student)

main 

Undergraduate Student (B.S), Embedded System-on-Chip Integrator
AI-Embedded System/Software on Chip (AI-SoC) Lab
School of Electronics Engineering, Kyungpook National University
Phone: +82 053 940 8648
E-mail: 1dlwkdgns [@] naver[DOT] com
[Homepage] [Google Scholar] [SVN]

Repository Commit History

main 

Introduction

Full Bio Sketch

Mr.Lee is currently doing his undergraduate degree in Electronics Engineering at Kyungpook National University, Daegu, Republic of Korea. His research interests include confidential and reliable Over-The-Air(OTA) update, both entire and partial, in OS environment and embedded system architecture using multimedia.

Research Topic

Fire Detection, Notification and Suppression System using YOLOv8

The currently widely used fire suppression system is the sprinkler. When a fire occurs and the temperature sensor exceeds the threshold, the sprinkler operates. Depending on the environment, this method may take a long time for the sprinkler to operate from the fime the fire occurs. As an alternative to this problem, I propose a fire detection, notification, and suppression system using AI. Using AI, immediate detection, notification and suppression are possible.

In this study, the following structure was designed. Each edge device continuously obtains one frame of image from the real-time video obtained through the camera. The YOLOv8 model trained to detect fire determines whether a fire has occurred in the captured image. If a fire has occurred, a rectangle is marked on the captured image at the location of the fire. After that, the device number and the corresponding image are transmitted to the central device, CM4 sensing, using RS485 communication. CM4 sensing transmits the received data to the server computer using socket communication, and the server transmits the received data to the mobile phone. The mobile phone generates a notification that a fire has occurred using a dedicated app. When you click the fire suppression button in the app, it will send a command to the edge device to suppress the fire by going back to the path where the fire occurrence data was transmitted. The edge device will spray water at the location of the fire shown in the image. In this study, spraying water is replaced by a laser pointer. This system uses RS485 wired communication, making it noise-resistant and ensuring integrity in communication. Additionally, This system can be expanded by installing additional edge devices.

Over-The-Air update using IrDA Communication in an OS environment

Recently, many devices use Over-The-Air(OTA) technology to update internal software using wireless communication. The advantage of performing OTA updates using the Internet is that many devices can be updated at the same time. However, OTA updates using the Internet have security vulnerabilities. There is a risk that unauthorized external users can eavesdrop on the contents of the OTA update software or arbitrarily update the software of the device.

As an alternative to this problem, I propose OTA updates using IrDA (Infrared Data Association) communication. IrDA has fewer security vulnerabilities during data transmission due to the directivity and short-range nature of infrared. It also allows for the creation of highly secure and water proof devices since it does not require a physical connection port.

In this study, OTA using IrDA is performed in an environment where an OS exists. AES is used for confidentiality, and CRC is used for integrity. The host divides the update file into packets of a certain size and transmits them to the target device. The target device checks the CRC and requests the host to retransmit the packet if there is a problem with the packet.

For small programs, it is okay to change the entire program, but for large modularized programs, especially programs that use multiple dynamic libraries, it is possible to update the program by changing only the dynamic libraries that need to be modified. In addition, it is possible to change the dynamic libraries at runtime, so that the program can be updated without stopping.

Publications

Conference Publications

  • Janghun Lee and Daejin Park. Fast Embedded Software Management based on Dynamic Partial Update Techniques via IR for Intelligent AI-Driven Systems (Under Review) In IEEE Symposium on Multimedia, Robotic Computing (SMRC), 2024.

Participation in International Conference

  • IEEE SMRC, Tokyo, Korea

Last Updated, 2024.9.30