From gmyoungblood-parc's github repository. Feel free to fork and go crazy.
Docker container setup for ArduPilot Software-In-The-Loop on Alpine Linux
This container establishes an environment for running ArduPilot, which includes a number of large packages such SciPy and OpenCV in a development supporting environment with gcc and python 2.7.
ArduPilot is set to run in ArduPlane using JSBsim for flight dynamics.
This container is currently 1.19 GB in size (stripped down from 1.95 GB) when built locally and squashed (e.g., docker-alpine-ardupilot$ build --squash -t ardupilot-contained .). It is listed on Docker Hub as gmyoungblood/ardupilot-sitl and weighs in, sadly due to caching, still at 1.95 GB. So, if you need a smaller version then build it locally.
The details: ArduPilot and JSBsim comprise ~360 MB, supporting (mostly python) /usr/lib libraries consume ~400 MB of which scipy, MAVproxy, and pymavlink take over half, and in /usr/libexec gcc and git (required) take up ~100 MB.
Use the environmental variable in setup, SIM_OPTIONS and the information below to configure the simulator.
Usage: sim_vehicle.py
Options:
-h, --help show this help message and exit
-v VEHICLE, --vehicle=VEHICLE
vehicle type (ArduCopter|AntennaTracker|APMrover2|Ardu
Sub|ArduPlane)
-f FRAME, --frame=FRAME
set vehicle frame type
ArduCopter: octa-quad|tri|singlecopter|firefly|gazebo-
iris|calibration|hexa|heli|+|heli-compound|dodeca-
hexa|heli-dual|coaxcopter|X|quad|y6|IrisRos|octa
AntennaTracker: tracker
APMrover2: rover|gazebo-rover|rover-skid|calibration
ArduSub: vectored
ArduPlane: gazebo-zephyr|CRRCSim|last_letter|plane-
vtail|plane|quadplane-tilttri|quadplane|quadplane-
tilttrivec|calibration|plane-elevon|plane-
tailsitter|plane-dspoilers|quadplane-tri
|quadplane-cl84|jsbsim
-C, --sim_vehicle_sh_compatible
be compatible with the way sim_vehicle.sh works; make
this the first option
-H, --hil start HIL
Build options:
-N, --no-rebuild don't rebuild before starting ardupilot
-D, --debug build with debugging
-c, --clean do a make clean before building
-j JOBS, --jobs=JOBS
number of processors to use during build (default for
waf : number of processor, for make : 1)
-b BUILD_TARGET, --build-target=BUILD_TARGET
override SITL build target
-s BUILD_SYSTEM, --build-system=BUILD_SYSTEM
build system to use
--rebuild-on-failure
if build fails, do not clean and rebuild
--waf-configure-arg=WAF_CONFIGURE_ARGS
extra arguments to pass to waf in its configure step
--waf-build-arg=WAF_BUILD_ARGS
extra arguments to pass to waf in its build step
Simulation options:
-I INSTANCE, --instance=INSTANCE
instance of simulator
-V, --valgrind enable valgrind for memory access checking (very
slow!)
-T, --tracker start an antenna tracker instance
-A SITL_INSTANCE_ARGS, --sitl-instance-args=SITL_INSTANCE_ARGS
pass arguments to SITL instance
-G, --gdb use gdb for debugging ardupilot
-g, --gdb-stopped use gdb for debugging ardupilot (no auto-start)
-d DELAY_START, --delay-start=DELAY_START
delays the start of mavproxy by the number of seconds
-B BREAKPOINT, --breakpoint=BREAKPOINT
add a breakpoint at given location in debugger
-M, --mavlink-gimbal
enable MAVLink gimbal
-L LOCATION, --location=LOCATION
select start location from
Tools/autotest/locations.txt
-l CUSTOM_LOCATION, --custom-location=CUSTOM_LOCATION
set custom start location
-S SPEEDUP, --speedup=SPEEDUP
set simulation speedup (1 for wall clock time)
-t TRACKER_LOCATION, --tracker-location=TRACKER_LOCATION
set antenna tracker start location
-w, --wipe-eeprom wipe EEPROM and reload parameters
-m MAVPROXY_ARGS, --mavproxy-args=MAVPROXY_ARGS
additional arguments to pass to mavproxy.py
--strace strace the ArduPilot binary
--model=MODEL Override simulation model to use
--use-dir=USE_DIR Store SITL state and output in named directory
--no-mavproxy Don't launch MAVProxy
Compatibility MAVProxy options (consider using --mavproxy-args instead):
--out=OUT create an additional mavlink output
--map load map module on startup
--console load console module on startup
--aircraft=AIRCRAFT
store state and logs in named directory
eeprom.bin in the starting directory contains the parameters for your simulated vehicle. Always start from the same directory. It is recommended that you start in the main vehicle directory for the vehicle you are simulating, for example, start in the ArduPlane directory to simulate ArduPlane
Set the container SIM_OPTIONS to stream towards your GCS using --out=udpout:<GCS hostname or IP>:<port>
The Mac has a changing IP address (or none if you have no network access). From Docker 17.06 onwards the recommendation is to connect to the special Mac-only DNS name docker.for.mac.localhost, which will resolve to the internal IP address used by the host.
For example: --out=udpout:docker.for.mac.localhost:14553 on the container and set the GCS to connect to localhost on port 14553 for UDP.
See Networking features in Docker for Mac
See Docker container networking
*** We do not test on Windows, so please provide feedback on suggestions for fixes for any issues ***
In Docker for Windows, the container communicates through a vEthernet adapter called DockerNAT. The host ports are available on the default gateway of the container network interface. Access the Docker Engine API on the host.
C:\> ipconfig
Windows IP Configuration
Ethernet adapter vEthernet (Temp Nic Name):
Connection-specific DNS Suffix . :
Link-local IPv6 Address . . . . . : fe80::99d:bf5e:8700:56df%26
IPv4 Address. . . . . . . . . . . : 172.27.219.121
Subnet Mask . . . . . . . . . . . : 255.255.240.0
Default Gateway . . . . . . . . . : 172.27.208.1
C:\> curl http://172.27.208.1:2375/info -UseBasicParsing
StatusCode : 200
StatusDescription : OK
...
The ArduPilot-SITL Docker Container is maintained by G. Michael Youngblood at the Palo Alto Research Center (PARC, a Xerox company) in California. It is licensed under GPLv3.