alediclem / adcnavinertial

ARM microcontroller board for inertial navigation algorithms

adcNAVinertial (current version v11_20)

Project Summary

adcNAVinertial is a simple 4-layer PCB development board designed to fulfill several purposes:

• Evaluate sensors performance for applications in inertial navigation;
• Sensor modeling and calibration procedures;
• Real-time full-speed USB sensor data streaming;
• Implementation of sensor fusion algorithms on embedded ARM MCU and evaluation of code-generation capabilities of MATLAB/Simulink environment;
• Lay the foundation for a navigation computer board for small autonomous multirotors;

Board details

The main component of the board is a STM32F4 Arm Cortex-M4 microcontroller. The board can mount a total of three sensor types which are connected to separate SPI peripherals of the processor. Some types of sensors have multiple footprint choices:

• The inertial measurement unit (IMU) sensor can be mounted on a LGA-14L or LGA-16 footprint;
• The magnetometer sensor can be mounted on a LGA-12 footprint;
• The absolute pressure and temperature sensor can be mounted on a HLGA-10L or LGA-10 footprint.

The MCU can be clocked by a 16MHz external crystal oscillator; a reset button and one status RGB LED may help debugging.

The board has a SWD connector for MCU debugging and programming, one Micro-B USB connector with ESD diode clamping connected to the MCU full-speed peripheral, a 6-pin PicoBlade connector socket for external embedded devices, and one 2-pin PicoBlade connector for external power supply with reverse polarity and overcurrent protection. Spare pins of the microcontroller can be accessed via test points or through-hole pads.

Assembly and test

A board prototype was easily assembled using reflow soldering; magnetometer IC was out of stock from distributors due to 2020 semiconductor shortage.

C drivers for sensors and USB were written in order to stream data to a host computer using a virtual COM port. The following Simulink model can be used for real-time visualization and processing up to the maximum sensor sample rate of 6.6 kHz:

The following scope and spectrogram capture shows an example of the accelerometer output for a sample rate of 104 Hz:

For the x axis:

Future versions

• Add ToF sensor;
• Remove unused footprints