Exploring the configurable options of the new IMU LSM6DSV32X in MEMS-Studio

Denise SANFILIPPO · ‎2024-07-03

Introduction

In this article, we explore the configurable options of LSM6DSV32X in MEMS-Studio, a software tool running on your PC. The hardware setup is made up of the professional MEMS tool board STEVAL-MKI109V3 with the DIL24 adapter of LSM6DSV32X (STEVAL-MKI240KA).

Hands-on

We build a new configuration for LSM6DSV32X with the dedicated tool for MEMS sensors: MEMS-Studio, which is a desktop software solution for all the MEMS sensors in the ST portfolio. Through this development environment, you can easily evaluate and program all the MEMS sensors. You can also develop embedded AI features, evaluate libraries, analyze data, and design algorithms even without writing code. MEMS-Studio is an all-in-one solution that includes all the features, which were previously available in other tools like Unico-GUI, Unicleo-GUI, and AlgoBuilder.
You can find more information about MEMS-Studio n the user manual UM3233 available on st.com.

The first step consists of downloading the latest version of MEMS-Studio for the three operating systems supported Windows, Linux and iOS. Once you have downloaded the software, you can install it on your computer and start the application.

After selecting LSM6DSV32X as device, in the advanced feature section of MEMS-Studio, you have the possibility to configure different features available in the sensor.

The [Pedometer] tool, for instance, can be used to configure and test the pedometer embedded in the device.

There are three different tabs for this tool:

[Configuration tab], which allows to easily configure the pedometer with its default configuration and evaluate the outputs in terms of steps detected.

Figure 1: Pedometer - Configuration tab

[Debug tab], which is used to load the data pattern into the device to test the pedometer on the data pattern loaded.

Figure 2: Pedometer - Debug tab

[Regression tab], which allows finding an optimal configuration based on the dataset you can provide.

Figure 3: Pedometer - Regression Tool

Another important tool in MEMS-Studio is the Finite State Machine [FSM] tool. This tool allows to configure different state machines embedded in the sensor from the [Configuration] tab.
Here you can see an example of the shake detection, which considers the lsm6dsv32x_shake.ucf file available on GitHub.

Figure 4: FSM - Configuration tab

In the [Testing] tab, you can evaluate the configuration and check the generated interrupts and output registers information.

Figure 5: FSM - Testing tab

In the [Debug] tab, you can inject samples in the device to evaluate the state machine state by state.

Figure 6: FSM - Debug tab

The machine learning core [MLC] tool instead allows the user to configure the MSC feature embedded in some devices. The configuration procedure is divided into several steps appearing in the different tabs.

The [Data patterns] tab allows managing the data patterns to be used to and assigning a label to each data pattern loaded.

Figure 7: MLC - Data patterns

Then [ARFF generation] tab allows configure inputs, filters, and features to generate an ARFF file.

Figure 8: MLC - ARFF generation

Then the [Decision tree generation] tab allows configuration of the decision tree outputs and generating the decision trees for the device.

Figure 9: MLC - Decision tree generation

The [UCF generation] tab instead allows setting the meta classifier and generating the configuration file, which contains a configuration for the sensor.

Figure 10: MLC - UCF generation

The [Datalog analysis] tab allows processing acquired data and recommends parameters like windows length filters and features that can be used for decision tree generation.

Figure 11: MLC - Datalog analysis

The [Decision tree output viewer] tab to test the configuration by looking at the decision tree outputs.

Figure 12: MLC - Decision tree output viewer

The [Data injection] tab to debug the configuration sample by sample.

Figure 13: MLC - Data injection

The STEVAL-MKI240KA kit contains the IMU LSM6DSV32X plus some electrodes for touch, swipe, and other applications. These applications are possible thanks to the Qvar embedded features in the sensor, where Qvar stands for electrostatic charge variation.

Inside the [Quick Setup] of the [Sensor Evaluation] tab of MEMS-Studio, you can enable the embedded Qvar feature clicking on the button [Enable AH_Qvar].

Figure 14: Sensor Evaluation tab - Quick Setup

In the [Features demo] section, you can select touch or swipe detection and evaluate. You can plot a single touch event, double touch, triple touch, single double, and long press inside the chart.

Figure 15: Sensor Evaluation - Features Demo

The following table shows other configuration examples that can be created. It is not exclusive to wearable applications configurations like gym activity recognition, gesture recognition and so on. It contains an asset tracking application using machine learning core and finite state machine available in ST MEMS sensors, such as LSM6DSV32X.

Figure 16: Wearable applications and asset tracking examples

Using the edge processing capabilities in ST MEMS sensors you can minimize the power consumption of your application. In fact, since the application runs on the sensor, those solutions require a current of just a few microamperes (μA).

Figure 17: Power consumption referred to MLC/FSM only

We encourage you to try MEMS-Studio with the new IMU LSM6DSV32X!

Exploring the configurable options of the new IMU LSM6DSV32X in MEMS-Studio

Introduction

Hands-on

Related links