Inertial platform

The Tightly Coupled GNSS Data Logger includes two inertial platforms. They are based on the Xsens MTI-3 and BerryGPS-IMU v4 respectivly. They provide inertial measurement and attitude estimate at 100 Hz that are published to the ROS network. The inertial units are located on the third level in the data logger.

Xsens MTI-3 AHRS

Hardware

The Xsens MTI-3 is a fully integrated Attitude and Heading Reference System (AHRS) delivered as a system on a module. It is mounted on a custom carrier board made to interface with a Raspberry Pi (4). Gerber files for the PCB is available in the repository. The PCB is designed to enable a multitude of communication options. The MTI-3 can be connected either by a USB cable (via a USB-to-Serial adapter on board), directly to the serial port of the Raspberry Pi or alternatively with SPI (not tested).

Software

The software to used to connect the MTI-3 to the ROS network is supplied by Xsense in their SDK.

BerryGPS-IMU v4

Hardware

The BerryGPS-IMU is a hardware platform developed by Ozzmakers for Attitude and Position estimation. It is designed to for Raspberry Pi and includes a range of sensors. Data from the sensors are either read from an I2C bus, SPI bus or UART.

Software

Since the BerryGPS-IMU does not include any software, a 3rd party library was used to provide Attitude and Heading estemates from the measured data. The library used is a fork of RTIMULib2 that is included in one of the submodules of this repository. The library can provide estimates using either a Kalman-Filter or RTQF. Since the library did not include any drivers for the sensors on BerryGPS-IMU, they where developed by us. To connect the IMU to the ROS network a bare-bone ROS publisher where written based on this guide.

Usage

Xsens MTI-3 AHRS

To use the Xsens MTI-3 AHRS in condjunction with ROS you need:

  • Xsens MTI-3 SoM assembled to the carrier PCB

  • Raspberry Pi with ROS Nuetic installed and fully setup

  • The Xsens MT SDK downloaded to the Raspberry Pi and installed accourding to this guide

Connect the carrier PCB to the Raspberry Pi with a USB cable. Select power option with JP1. Navigate to catkin_ws/src and run the following commands to source the workspace and launch the ROS node

source devel/setup.bash
roslaunch xsens_mti_driver xsens_mti_node.launch

The ROS node shuld launch and connect to an avalible ROS master or launch a local master if no other is found. If everything works as intended you shuld see

Measuring...

In your terminal window. If not, then refer to the troubelshoting on Xsens webpage.

BerryGPS-IMU v4

Warning

In the current version of the software, data from the magnetometer is not accurate and causes the attitude to be wrong. Presumably this is due to the callibration not working. A workaround is to disabel magnetic data in the estimation.

To use the BerryGPS-IMU v4 in conjunction with ROS you need:

You can use this repository to get all the necessary software listed above.

Initial setup

Clone the above repository to you local ROS workspace by issiung the following commands:

cd catkin_ws
git clone URL ./src
cd src
git submodule init
git submodule update

Follow by building your catkin workspace:

cd ~/catkin_ws
catkin_make

Also make sure to build and install the RTIMULib which includes the hardware drivers and sensor fusion algorithms:

cd ~/catkin_ws/src/RTIMULib2/Linux
mkdir build
cd build
cmake ..
make -j4
make install
ldcofig

Copy the RTIMUConfig file to your catkin workspace

cp ~/catkin_ws/src/BerryIMU/RTIMULib.ini ~/catkin_ws/

Run the ROS node by sourcing your workspace and then launching the node:

source devel/setup.bash
rosrun BerryIMU BerryIMU_node

If the node finds a ROS master it will launch and you will see

Measuring...

In your terminal. If not, contact the software maintainer.