How to deploy the AI ​​model on the STEVAL-STWINKT1B board?

[Felipe_Nascimento]

I'm trying to develop an AI model and integrate it into the STEVAL-STWINKT1B board to classify the noise of industrial engines.

How do I implement the generated library in nano edge ai so that the board shows the classification by obtaining data from sensors in real time?

I followed the steps in this documentation, but I was unsuccessful. https://wiki.st.com/stm32mcu/wiki/AI:How_to_perform_anomaly_detection_using_FP-AI-MONITOR1

[SimonePradolini]

Hello @Felipe_Nascimento 

If you are interested in the FP-AI-MONITOR1 demo specifically, can you explain to me better where you are failing to replicate the demo?

In the meantime, I can suggest you also another approach, by using ST tools available through ST Edge AI Suite.

The idea behind Edge AI Suite is to provide a comprehensive solution to help developers accelerate their product transformation. Here, developers can find the tools to optimize and deploy machine learning algorithms, from data collection to final deployment on hardware. 

This presentation can guide you to select the tool you need.

I guess you are searching for the High Speed Datalog tool, that allows users to manage the acquisition and labelling of sensor datasets via SD card or USB for different STEVAL and NUCLEO boards, included STWIN board (STEVAL-STWINKT1B). The firmware is released both in source code or precompiled, so if you are not interested in embedded programming you are free to flash the precompiled binary directly. It offers also a Python SDK to collect and manage your dataset and convert them in the preferred format. Converters to CSV, Unico and NanoEdge format are natively supported.

I can suggest 2 approaches:

1. STWIN board mounts also ISM330DHCX sensor, a 6-axis IMU with Machine Learning Core (MLC), a dedicated core for machine learning processing in the sensor. See the datasheet for the full description. High Speed Datalog natively supports this feature, so I suggest you first collect data on your setup by enabling ISM330DHCX only, then convert the dataset for Unico and generate a configuration for MLC in UCF format. Once you have your UCF, you can use STWIN with the same Datalog firmware and load the UCF file. The outcome of the ISM330DHCX MLC will be available in the High Speed Datalog tool. This Quick Start Guide and this User Manual can guide you during the procedure.  Here and here we provide an example based on STWIN.box + ISM330BX. Board and sensor are different, but the procedure is the same. You can use it as a reference.

2. If you are interested in NanoEdge AI technology, you can acquire the dataset by enabling the sensors you wish, then convert the dataset for NanoEdge, load it into NanoEdge AI Studio, and let the tool select the AI solution you need.

Let me know it these solutions can help you.

Best regards

Simone

[Felipe_Nascimento]

I wanted a step-by-step guide on how to integrate the library onto the STEVAL-STWINKT1B board after generating the AI ​​model 

According to the topic bellow

2.3. Installing the NanoEdge^TM Machine Learning library

[Julian E. · Accepted solution]

Hello @Felipe_Nascimento ,

I think you will find everything that you need here:

AI:How to Build an Anomaly Detection Project for Predictive Maintenance with NanoEdge AI Studio - stm32mcu

In the end of the tutorial, we use a git repository to use the datalogging code to create our final application. I believe that the STEVAL-STWINKT1B is available.

You can also take a look at the other tutorials.

Have a good day,

Julian

[Felipe_Nascimento]

In this case, it needs to be on the STEVAL-STWINKT1B board because of the project, as I was already able to acquire data on the SD Card and generate the library on the Nano Edge Ai.

Now we need to integrate the library generated on the STEVAL-STWINKT1B board, already doing the classification in real time to at least show it on the terminal.