You'll need the submodules and dependencies, so clone like so:
git clone --recurse-submodules REPO_ADDRESS
Or, after cloning, run:
git submodule update --init --recursive
To get the F’ dependencies on Ubuntu (all dependencies are available through Homebrew on Mac):
sudo apt-get install pkg-config g++ python-pip python-lxml python-tk python-dev
On all platforms (modify the pip command if you don’t want a global install)
sudo -H pip install -r Gse/bin/required.txt
Go to the ROS/fprime_ws folder, and build the workspace with your ROS environment sourced (tested with catkin):
catkin init
catkin build
From the top directory of the repo, run:
touch ROS/fprime_ws/src/fprime/mod.mk
For a particular deployment (e.g. SDREF
, HEXREF
), go to that folder and run make dict_install
Then, run touch Gse/generated/DEPLOYMENT/serializable/ROS/__init__.py
See quest-fw#3
- Acts as a shim to transport data to and from low-level processor
- Interfaces with ROS and F' ground station
- On Snapdragon Flight, loads
HEXREF
onto DSP and interfaces with that code - On any Linux environment, interfaces to low-level processor over two UARTS (one data and one debug)
To set up a new installation, download Flight_3.1.3_qrlSDK.tgz
from Intrinsyc and qualcomm_hexagon_sdk_lnx_3_0_eval.bin
from Qualcomm, and place in cross_toolchain/downloads
.
Run ./bootstrap.bash
from the top level of the repository. This will set up the Hexagon toolchain using install.sh
from cross_toolchain
, install all required dependencies, create a Linaro Indigo ARM sysroot and install ROS in it using proot
for later cross-compilation include and link steps.
Tested on Linux and Mac with TI Code Composer Studio version 8.1.0
, with ARM compiler version 18.1.2
To build the simulation example, go to the SIMREF folder, and run the following:
make gen_make
make
Then, you can run the executable like so (after starting ROS master):
ROS_NAMESPACE=firefly ./linux-linux-x86-debug-gnu-bin/SIMREF
When you start the RotorS (https://github.com/ethz-asl/rotors_simulator) firefly example, SIMREF will use the /clock message to carry out control cycles. This parallels what happens on hardware targets, where the IMU data-ready interrupt triggers the estimation and control loops.
Works out of the box with https://github.com/genemerewether/ethzasl_sensor_fusion for testing high-level filter updates, but can be easily adapted to simulated sensors in Gazebo. Just run additional ROS nodes as necessary, and remap the pose or position sensor topics of the sensor fusion packages. Or, publish the mav_msgs/ImuStateUpdate
message (see the mav_msgs
submodule of this repo) from an appropriate filter.
F' Release Notes
Release 1.0:
This is the initial release of the software to open source. See the license file for terms of use.
An architectural overview of the software can be found here.
A user's guide can be found here.
Documentation for the Reference example can be found here.
Release 1.01
Updated contributor list. No code changes.