Simbotic Cortex
Machine learning, physics and controls development environment:
- airsim rosnode
- python airsim and tensorflow scripts
- C++ and tensorflow applications
- AirSim lib developemnt - physics models and APIs
- PX4 missions
Setup
Clone this repo recursively:
git clone --recursively git@github.com:VertexStudio/simbotic-cortex.git
Docker Setup
CPU-based docker build:
./docker_build.sh
./docker_run.sh
GPU-based docker build. You will need nvidia-docker.
./docker_build.sh gpu
./docker_run.sh gpu
Inside Docker Image
Python Client
cd $AIRSIM_ROOT/scripts
python hello_drone.py
C++ Client
Setup AirSim:
$AIRSIM_ROOT/setup.sh
Build C++ app:
cd $AIRSIM_ROOT/app
./cmake_build/bin/cmake -G "Unix Makefiles"
make -j`nproc`
./air-app
Advanced Docker Setup
CPU-based image
cd simbotic-cortex # into this repo clone
docker build --build-arg USER_ID=$(id -u ${USER}) --build-arg GROUP_ID=$(id -g ${USER}) -t simbotic-cortex/cpu .
docker run --runtime=nvidia -ti -v $(pwd):/sim --network=host -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix --cap-add=SYS_PTRACE simbotic-cortex/cpu
GPU-based image
You will need nvidia-docker.
cd simbotic-cortex # into this repo clone
docker build --build-arg USER_ID=$(id -u ${USER}) --build-arg GROUP_ID=$(id -g ${USER}) -t simbotic-cortex/gpu .
docker run --runtime=nvidia -ti -v $(pwd):/sim --network=host -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix --cap-add=SYS_PTRACE simbotic-cortex/gpu