ST STEVAL-MKSBOX1V1 User manual

Introduction
The STEVAL-MKSBOX1V1 (SensorTile.box) is a ready-to-use box kit with wireless IoT sensor platform designed to help you
build apps that use motion and environmental sensors, regardless of your level of expertise.
The hardware node is a board that fits into a small plastic case (IP54) with a rechargeable battery. You can connect with your
smartphone to the board via Bluetooth by using the ST BLE Sensor app (available both on Google Play and Apple Store) and
immediately build your own apps through a special interface that offers beginner and expert level functionality. This multi-sensor
kit therefore allows you to design wireless IoT and wearable sensor applications quickly and easily, without performing any
programming.
SensorTile.box includes a firmware programming and debugging interface that allows professional developers to engage in
more complex firmware customization using the STM32 Open Development Environment (STM32 ODE), which includes a
sensing AI function pack with neural network libraries.
The kit board includes an embedded SPBTLE-1S Bluetooth SMART application processor that is compliant with BT
specification v4.2. This transmitter module is FCC (ID:S9NSPBTLE1S) certified and IC (IC:8976-SPBTLE1S) certified.
Figure 1. STEVAL-MKSBOX1V1 (SensorTile.box) multi sensor development kit
How to use the wireless multi sensor development kit with customizable app for
IoT and wearable sensor applications
UM2580
User manual
UM2580 - Rev 5 - May 2020
For further information contact your local STMicroelectronics sales office.
www.st.com

1How to set up the hardware
Important:
Before you begin, please check the insert card that comes with the SensorTile.box blister pack. If it doesn't show a procedure
for battery connection similar to the steps below, then your device is supplied with the battery already connected to the board.
In this case, you only need to connect the device via USB to wake it up the first time.
If the insert card has a similar procedure to the steps below, your device is supplied with the battery disconnected and you
should follow this procedure to connect the battery and wake the device up.
Step 1. Remove the SensorTile.box contents from its package.
Step 2. Unscrew the shroud cover.
You should have the following items:
– An evaluation board in a plastic shroud
– A LiPo battery
Step 3. Slide the male battery connector vertically into the female connector on the board.
You will hear a light click when the connector is attached correctly.
Figure 2. STEVAL-MKSBOX1V1 battery connection
Step 4. Re-position the circuit with the battery below it and the close the shroud with one of the following types
of lid:
–with flanges
– without flanges
Step 5. If necessary, charge the battery via a USB cable.
The blinking of the red LED indicates the battery charging status.
UM2580
How to set up the hardware
UM2580 - Rev 5 page 2/41

2How to use ST BLE Sensor app with SensorTile.box
Before you begin, you need to download and install our ST BLE Sensor app on your smartphone. The app is
available from the Google and Apple online stores.
Step 1. Launch the app on your smartphone.
Figure 3. ST BLE Sensor app main screen
UM2580
How to use ST BLE Sensor app with SensorTile.box
UM2580 - Rev 5 page 3/41

Step 2. Select [CREATE A NEW APP].
The Example Apps screen that follows lists the preloaded apps that you can use immediately.
Figure 4. Example Apps screen
UM2580
How to use ST BLE Sensor app with SensorTile.box
UM2580 - Rev 5 page 4/41

Step 3. Select one of the apps with the icon from the list.
After you select the app, ST BLE Sensor will scan for available SensorTile.box devices in range.
Figure 5. Board selection
Step 4. Select the appropriate SensorTile.box device from the Board screen.
A blue LED on the SensorTile.box device will flash slowly to confirm Bluetooth pairing.
A pop up message in ST BLE Sensor will prompt you to confirm loading the new app in replacement of
any previously opened apps.
Step 5. Select the appropriate SensorTile.box device from the Device List.
The app will commence monitoring or logging activity and return real time feedback data to the
corresponding app screen in ST BLE Sensor.
UM2580
How to use ST BLE Sensor app with SensorTile.box
UM2580 - Rev 5 page 5/41

3Application descriptions
3.1 Mode 1 example apps
The ST BLE Sensor bundles the following ready-to-use app scenarios:
• Baby Crying Detector
• Barometer
• Compass and Level
• Data Recorder
• Human Activity recognition
• In-Vehicle Baby Alarm
• Pedometer
• Sensor Fusion - Quaternion
• Vibration Monitor - Training
• Vibration Monitor - Compare
Figure 6. Apps screen
App scenarios with the icon produce immediate outputs on your smartphone in real time.
App scenarios with the icon store sample data on the internal micro SD card.
App scenarios with the icon are reserved for Expert mode.
UM2580
Application descriptions
UM2580 - Rev 5 page 6/41

RELATED LINKS
2 How to use ST BLE Sensor app with SensorTile.box on page 3
3.1.1 Baby crying detector
The baby crying detector (BCD) app implements the Fast Fourier Transform (FFT) and artificial intelligence
processing to detect baby crying events using the Sensortile.box on-board microphone.
The analysis of the acquired audio is based on the FFT that converts a signal from its original time domain to a
representation in the frequency domain.
Figure 7. FFT analysis - from time to frequency domain
The FFT of the audio signal is the result of all the contributions of each frequency and the related magnitude
factor generated by the audio signal.
Figure 8. FFT analysis - principles
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 7/41

The FFT feature extraction of the acquired signal is processed by the STM32 MCU which calculates the MEL FFT
and the MFCC (MEL frequency cepstral coefficient) parameters sent to the implemented MCU neural network: if a
baby crying event is detected, the green user LED on the Sensortile.box board lights up and a warning is sent to
the smartphone via Bluetooth.
Figure 9. STBLESensor - baby crying detection process
The neural network is classified as a deep feed forward neural network and its structure is composed of 2 hidden
nodes of 100 neurons each.
The tool used to develop the neural network is Keras with an open source high level library written in Python.
Optimization and loading of the neural network on the Sensortile.box has been performed using
STM32CubeMX.AI.
Figure 10. STBLESensor - baby crying app neural network
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 8/41

The baby crying app works with the following ST high sensitivity audio sensor and operating parameter settings:
•APP DATA INPUT: microphone audio acquisition
•SENSOR USED: MP23ABS1 MEMS microphone
•SENSOR SETTINGS: 16 KHz sample acquisition
•APP DATA OUTPUT: baby crying/not crying icon
RELATED LINKS
Appendix A ARMA filter coefficient calculation on page 36
3.1.2 Barometer app
The Barometer app uses the Sensortile.box on-board environmental sensors (STTS751, LPS22HH and HTS221).
Figure 11. Barometer app screen
The Barometer app monitors the environmental information in real-time and shows the data on your smartphone
as icons or graph plot.
The LPS22HH pressure sensor embeds another built-in sensor for temperature compensation (0.5 hPa with a
range of 260-1260 hPa of absolute pressure).
The STTS751 digital temperature sensor has an operating temperature range of -40/+125 °C, with maximum
resolution 0.0625 °C/LSB and precision of ± 0.5 °C (typ.).
The HTS221 digital relative humidity and temperature sensor has a relative humidity range of 0/100%, a
sensitivity of 0.004% RH/LSB, a humidity accuracy of ± 3.5% RH, 20-80% RH and an accuracy in temperature of
± 0.5 °C (typ,) in the range of 15/+40 °C.
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 9/41

When you run the Barometer app and connect the Sensortile.box device, the ST BLE Sensor app shows a
monitoring screen for the environmental sensors.
Figure 12. Environmental screen
You can access other output options from the menu icon in the top left of the screen.
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 10/41

Figure 13. Plot Data screen - humidity
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 11/41

Figure 14. Plot Data screen - temperature
Thanks to the low power sensors, low output data rate and low power MCU, this app is highly suitable for battery-
based projects with very low power consumption.
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 12/41

The Barometer app sets the following operating parameter settings for the following ST high accuracy
environmental sensors:
•APP DATA INPUT: pressure, temperature and humidity values
•SENSORS USED:
–LPS22HH (absolute pressure MEMS digital sensor)
–STTS751 (temperature digital sensor)
–HTS221 (relative humidity and temperature digital sensor)
•SENSOR SETTINGS:
–LPS22HH settings:
◦ Power mode: Low Noise
◦ Output Data Rate: 1 Hz
◦ Filter: ODR/2
–STTS751 settings:
◦ Low power mode
◦ Output data rate: 1 Hz
–HTS221 settings:
◦ Low power mode
◦ Output data rate: 1 Hz
•APP DATA OUTPUT:
– Relative humidity (%)
– Temperature (°C)
– Absolute Pressure (mBar)
– Plot collected data
3.1.3 Compass and Level app
The Compass and Level app shows the orientation estimation of the Sensortile.box in relation to the Earth
magnetic North or level indication.
The app is based on the sensor fusion firmware algorithm (MotionFX library) embedded in the Sensortile.box
MCU.
The sensor fusion algorithm is an extended Kalman filter (EKF) that minimizes sensor inaccuracies and includes
gyroscope calibration and magnetometer calibration (to compensate the magnetometer offset).
Figure 15. MotionFX algorithm flow
The algorithm uses the LSM6DSOX iNemo 6-axis accelerometer and gyroscope data and the LIS2MDL 3-axis
compensated magnetometer data as inputs, combining the two sensors in a virtual 9-axis sensor.
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 13/41

The magnetometer indicates yaw angle and heading, but only if hard-iron offset is compensated and when there
is no additional magnetic field around the Sensortile.box disturbing the measurement.
Figure 16. MotionFX algorithm flow - magnetometer function
Data regarding yaw and angle heading are also given when tilt is compensated by the accelerometer.
Figure 17. MotionFX algorithm flow - accelerometer tilt compensation
The gyroscope indicates the new orientation based on the previous one, when its bias is compensated by the
accelerometer. The gyroscope can detect the static condition of the Sensortile.box,
Figure 18. MotionFX algorithm flow - gyroscope function
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 14/41

To enable the magnetometer calibration, you need to touch the symbol highlighted in the picture below.
Figure 19. STBLESensor Compass and Level app - calibration icon
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 15/41

Then, move the Sensortile.box in a 8-pattern figure as shown below; the calibration is completed when the icon
becomes green.
Figure 20. STBLESensor Compass and Level app - starting the calibration
The MotionFX library provides orientation estimation, magnetometer hard-iron offset compensation,
accelerometer vibration rejection and gyroscope bias compensation.
The sensor fusion algorithm calculates the quaternion coefficient and the Euler angles to detect the right
orientation of the Sensortile.box represented by a compass rotation or level indicator.
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 16/41

Figure 21. STBLESensor Compass and Level app - orientation screen
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 17/41

Figure 22. STBLESensor Compass and Level app - offset example 1
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 18/41

Figure 23. STBLESensor Compass and Level app - offset example 2
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 19/41

The Compass and Level app works with the following ST high accuracy motion sensors and operating parameter
settings:
•APP DATA INPUT: Accelerometer, gyroscope and magnetometer values
•SENSORS USED:
–LSM6DSOX (acceleration and gyroscope sensor)
–LIS2MDL (compensated magnetometer)
•SENSOR SETTINGS:
–LSM6DSOX:
◦ Low power mode
◦ Output data rate: 52 Hz
◦ Low pass filter: 700 Hz
◦ Full scale: 2 g for accelerometer, 2000 dps for gyroscope
–LIS2MDL settings:
◦ Low power mode
◦ Output data rate: 50 Hz
◦ Full scale: 50 gauss
•APP DATA OUTPUT:
– Compass orientation model
– Level indication model
◦ Quaternion values
◦ Heading
◦ Euler angles
◦ Plot collected data
3.1.4 Data Recorder app
Data recorder can be used to monitor and record movements and/or environmental conditions that parcels or
objects are subjected to during movement or shipping.
The data can be used to verify whether a parcel has suffered shocks or undesirable temperatures that could
damage the goods, or if a vehicle has been driven according to appropriate speed and safety parameters.
Certain sensors are enabled according to what is being monitored, and data is stored in the internal memory card
for later retrieval and analysis. Motion sensors are set to Low Power Mode with a data rate of around 50 to
100 Hz, while a data rate of 1 Hz is appropriate for environmental sensors.
3.1.5 Human Activity Recognition app
The Human Activity Recognition app uses the Sensortile.boxLSM6DSOX MEMS accelerometer and the
embedded Machine Learning Core (MLC).
The following activities can be recognized by icons on the screen, independently from the Sensortile.box
orientation:
• Stationary
• Walking
• Fast walking
• Jogging
• Biking
• Driving
UM2580
Mode 1 example apps
UM2580 - Rev 5 page 20/41