I appreciate you providing the title, but there is no Korean text in your message to translate. "Industrial Edge AI Solution Challenge 2026" is already in English. Please provide the Korean text you would like me to translate, and I'll render it into natural, technical-industry register English while preserving all HTML tags and structure.

Industrial Edge AI Solution Challenge 2026 is

a problem-solving focused AI contest that tackles practical issues remaining on manufacturing floors

using on-device AI, edge AI, and embedded technologies, now that smart factories have reached a certain level of adoption.

This contest is not a simple algorithm performance competition or theory-focused AI competition,

but aims to "directly create AI solutions that can be used in real factories."

Through implementation of complete Physical AI solutions connecting sensors, data, AI models, and actual control/decision-making,

we seek to discover next-generation AI technologies and talent needed by Korean manufacturing industries.

◆ Reasons for Conducting This Contest

Smart factories have rapidly spread, but the following questions continue to be repeated in manufacturing sites.

• Can we detect equipment anomalies before they fail?

• Can AI replace human visual inspection for defect detection?

• Can we standardize expert experience as an AI model?

• Why can't AI be directly applied to collected process data?

• Can we implement secure AI without using the cloud?

◆ Contest Core Missions (6 Problem Areas)

Participants can freely create works by selecting one of the 6 manufacturing site problem areas below.

① Equipment/Process-Customized AI Models: AI prediction and analysis models reflecting characteristics by process and equipment

② Edge-Based Real-Time Anomaly Detection: AI that detects anomalies in motors, bearings, pumps, and equipment in real-time

③ AI-Based Quality and Defect Inspection Automation: Vision and sensor-based defect judgment and quality inspection AI

④ Data Pipeline Automatic Management and Data Integrity Assurance: Automation solution that makes manufacturing data immediately usable by AI

⑤ Process Control and Autonomous Optimization Loop: Autonomous system where AI analyzes processes and performs direct control

⑥ Security Enhancement + On-Device AI: Security-focused AI system operating locally without cloud

◆ What Kind of Projects Can You Create?

Participants can implement practical Edge AI projects as follows.

• Equipment failure prediction system based on bearing noise

• Automatic scratch and assembly defect inspection device using AI cameras

• AI-based reflow oven/temperature chamber automatic control system

• Manufacturing sensor data automatic data integrity and calibration solution

• On-device AI analysis box operating in closed-network environments

👉 As long as your project can demonstrate a working demo, technology stack and implementation methods can be freely chosen.

◆ What Hardware and Software Should I Use?

You can use any board you have, such as Raspberry Pi, ESP32, STM32, Jetson Nano, etc.,

and can directly implement on-device AI devices by deploying lightweight models like YOLO, Edge Impulse, and Tiny LLM on actual devices.

◆ Who Can Participate?

Anyone from students to practicing engineers with interest in practical problem-solving can participate

• Embedded, AI, robotics, and automation developers

• Engineers interested in manufacturing and industrial AI

• University (graduate) students, researchers, startups, makers

• All teams/individuals who want to create industry problem-solving AI projects

◆ "Total 5 Sessions of Edge AI Practical Webinars" Provided for Participants!

With DigiKey sponsorship, the entire process from basics → practice → real-world application is guided through 5 educational sessions to help participants complete the contest.

1. Understanding Edge AI & Physical AI

• On-device AI structure
• Physical AI concept integrating sensor+AI+motor
• Application cases in industrial, robotics, and mobility sectors

2. Edge AI Implementation Based on STM32 N6

• Structure for executing AI on ultra-low-power MCU
• Sensor-based anomaly detection

3. ESP32 Edge AI Hands-On Practice

• Image/gesture classification with ESP32-S3 + camera
• TFLite Micro / FOMO execution

4. Edge Impulse Learning Workflow

• Data collection → model training → device deployment
• Automatic generation of audio/image/IMU models

5. Running YOLO on Jetson Nano

• YOLO nano model deployment
• Real-time recognition based on TensorRT acceleration

These 5 educational sessions are designed so anyone can follow along and create a finished work.

DigiKey Parts Refund (Supporting 40 Teams Total)

• First-come, first-served 10 teams → Up to 100,000 KRW refund

• Following 30 teams → Up to 50,000 KRW refund

In Quest 1, only invoices are submitted, and refunds are paid only to teams that reach Quest 3 (final submission).

Any board you have is OK, such as Raspberry Pi, ESP32, STM32, Jetson Nano, etc.!

This is a contest where you directly implement on-device AI devices by deploying lightweight models like YOLO, Edge Impulse, and Tiny LLM on actual devices.

Important Notes

If the number of contest applicants falls below a certain standard,

the contest schedule may be changed or the contest may be cancelled or suspended for operational reasons.

• Registration Period: Thursday, December 18, 2025 ~ Friday, January 16, 2026
• Quest 1 Period: Thursday, December 18, 2025 ~ Friday, January 16, 2026
• Quest 2 Period: Saturday, January 17, 2026 ~ Sunday, February 15, 2026
• Quest 3 Period: Monday, February 16, 2026 ~ Sunday, March 1, 2026

※ Contest registration is available during Quest 1 period.