ST Live: ST's AI on the Edge - Anomaly Detection Solution Based on NanoEdge AI

eric_wang

ST Live: ST's AI on the Edge - Anomaly Detection Solution Based on NanoEdge AI [Copy link]

 

Live Broadcast Introduction

NanoEdge AI studio is a software tool that can help users automatically generate AI models. It does not require users to have professional knowledge in the field of AI. It only needs to provide a small amount of signal data to quickly create an AI model that meets application requirements. This course mainly introduces how to use NanoEdge AI Studio to automatically create an AI model based on three-axis acceleration sensor signals and quickly deploy it to the ST SensorTile.box development board for anomaly detection.

Live broadcast time

August 10, 2022, 15:00-17:00

>>>Click here to sign up<<<

Live broadcast arrangement

This course will first give a comprehensive introduction to NanoEdge AI Studio, including application areas, problems that can be solved, supported functions, supported AI algorithms, related demos, and actual customer use cases. Then, it will demonstrate step by step how to automatically collect sensor data through NanoEdge AI Studio, and automatically create AI models to solve related problems based on the collected data, and finally quickly integrate the generated AI library into the SensorTile.box development board for functional verification. In the hands-on practice session, we will also lead everyone to complete the above steps step by step, so that everyone can quickly create their own AI applications.

Teaching part

1) Market overview of STM32-based AI solutions

2) Introduction to NanoEdge AI Studio

Experimental Section

Hands-on practice: Implement an AI anomaly detection application based on accelerometer sensor on SensorTile.box.

Experimental Platform

This experiment is based on ST's SensorTile.box (STEVAL-MKSBOX1V1) development board. It integrates a low-power MCU STM32L4R9 and many high-precision sensors, including six-axis inertial sensor LSM6DSOX, three-axis acceleration sensors LIS2DW12 and LIS3DHH, digital temperature sensor STTS751, audio sensor MP23ABS1, etc. In the experiment, we will create an AI model based on the three-axis acceleration signal of LSM6DSOX and perform anomaly detection.

Preparation

To ensure the smooth progress of the experimental course, please prepare in advance:

1) NanoEdge AI Studio v3.1.0 (for automatic creation of AI models)

2) STM32CubeIDE V1.6.1 (for compiling and downloading code)

3) ST-LINK V2 or V3 (for downloading code, please pay attention to the debug interface type of SensorTile.box and prepare the corresponding debug adapter cable)

4) USB to MicroUSB cable (for powering the board and collecting sensor data)

5) Phillips screwdriver (for opening the case of the SensorTile.box)

"Follow the example" activity

• Step 1: Follow the course on August 10th - interact with ST instructors online to learn how to use Nano edge AI to build projects through PC and demo board;
• Step 2: August 10-18 Conception - Through the online training course, propose your ideas about what kind of project you will use Sensortile.box (STEVAL-MKSBOX1V1), the basic ideas of the project, the expected results, etc.
• Step 3: Selection from August 19 to 26 - ST will select the project ideas of interest and provide free development boards for the selected candidates to complete their projects;
• Step 4: August 27 - September 30: Create a project video - Complete the project using Sensortile.box (STEVAL-MKSBOX1V1) based on the original project idea and save the project video;
• Step 5: After October, multi-platform promotion - ST official platforms (official WeChat, video account, Bilibili account, etc.) will simultaneously promote excellent project case videos and reward an STM32 T-shirt

soso

The whole learning process is quite friendly to beginners.