Developing with MEMS-Studio and the latest-generation sensors

[Laurids_PETERSEN]

Learn how to use the X-NUCLEO-IKS4A1, the latest STM32ODE expansion board equipped with motion MEMS and environmental sensors. 

You’ll interface with various MEMS sensors, visualize and analyze output data, and use advanced digital features from the machine learning core and intelligent sensor processing unit.

You’ll use the MEMS-Studio desktop software solution for evaluating and programming all MEMS sensors. MEMS-Studio introduces innovative solutions for expanding the functions of established applications such as Unico-GUI, Unicleo-GUI, and AlgoBuilder.

Prerequisites

• STM32 Nucleo-64 development board with STM32F401RE MCU (NUCLEO-F401RE)
• Motion MEMS and environmental sensor expansion board for STM32 Nucleo (X-NUCLEO-IKS4A1)
• Mini-B USB cable

What you'll learn

• How to use MEMS-Studio
• MEMS data capture and analysis
• Machine learning core configuration and visualization
• How to use the machine learning core to classify activity through motion
• Implementing an application using AlgoBuilder within MEMS-Studio
• Designing an application with an IMU equipped with an Intelligent Sensor Processing

Ready to get started?

The videos are available in our YouTube playlist embedded below, or at the following URL:
https://www.youtube.com/playlist?list=PLTJzs51NlEIDjjXEtibeY0zQSFRa0ARYG

Join the discussion

Share your thoughts, ask questions, and engage with fellow developers. Our experts are active in this post, and they are excited to hear your feedback.

Additional resources

• Webinar: In-sensor monitoring with intelligent MEMS sensors 
• Software: X-CUBE-MEMS1
• Software: MEMS-Studio
• Purchase an X-NUCLEO-IKS4A1 
• Purchase a NUCLEO-F401RE
• Knowledge article: Exploring the configurable options of the new IMU LSM6DSV32X in MEMS-Studio
• Knowledge article: How to start and improve your workflow with MEMS-Studio
• Knowledge article: Accurate sensor calibration and SFLP algorithms with X-NUCLEO-IKS4A1
• Knowledge article: How to augment your anomaly detection application with ISM330BX machine learning core

[LPetr.1]

I have collected some samples of different movements:
1. Flipping down the device

2. Flipping up the device

3. Shaking the device

4. Device stationary on the table

 

When I upload all CSV's to the MLC tab Data patterns menu, something is not right.

First thing I notice is that when I am uploading shaking/stationary samples, it loads very fast. When I select flip_down or flip_up samples and click "Load" button the MEMS Studio hangs for up to 1 minute to load 3-4 50KB files which indicates that something may be wrong. I did try to inspect the CSV's using text editor and they all look fine and simillar.

Second thing, I dont understand what is the issue with the flip_up and flip_down samples. Why the Time option is different between these and stationary/shaking? As a result of that, I cannot proceed with ARFF generation.

 

I have explained more about the issues I am having in the community forum post below:
https://community.st.com/t5/mems-sensors/using-mems-studio-to-generate-custom-machine-learning-decision/td-p/892810

hopefully you can clarify and explain the issues I am having