UTM Multi-Features Docker Based TestBed

This is an academic project that aims to implement a docker based simulation infrastructure that allow to represent multiple features of an Unmanned Aircraft System Traffic Management (UTM), from the low level drone control system up to the airspace management algorithm.

The proposed TestBed is under construction and the current version was built through the integration of some OpenSource projects, as presented below:

Reproducing the TestBed

All tests were done on Ubuntu 20.04 LTS for the infrastructure and Windows 10 for the visualization tool.

Building the Infrastructure

The base for the resources for the TestBed are:

DOCKER CONTAINERS: https://www.docker.com/
CONTAINERNET (MININET): https://containernet.github.io/ MAVSDK: https://github.com/mavlink/MAVSDK
PX4 AUTOPILOT: https://github.com/PX4/PX4-Autopilot
imGUI: https://github.com/ocornut/imgui

STEP 1: Install DOCKER

REF: https://docs.docker.com/engine/install/ubuntu/

sudo apt-get update
sudo apt-get install \
    ca-certificates \
    curl \
    gnupg \
    lsb-release

curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /usr/share/keyrings/docker-archive-keyring.gpg
echo \
  "deb [arch=$(dpkg --print-architecture) signed-by=/usr/share/keyrings/docker-archive-keyring.gpg] https://download.docker.com/linux/ubuntu \
  $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null

sudo apt-get update
sudo apt-get install docker-ce docker-ce-cli containerd.io

STEP 2: Install CONTAINERNET

REF: https://containernet.github.io/#installation

sudo apt-get install ansible
git clone https://github.com/containernet/containernet.git
sudo ansible-playbook -i "localhost," -c local containernet/ansible/install.yml

Test CONTAINERNET

# run interactive container and directly start containernet example
docker run --name containernet -it --rm --privileged --pid='host' -v /var/run/docker.sock:/var/run/docker.sock containernet/containernet

STEP 3: Get PX4 SITL HEADLESS FLEX

The Drone simulation solution selected for this first version of this TestBed was the SITL from the PX4 Autopilot (https://px4.io/) using Gazebo (http://gazebosim.org/) simulation. To make it possible, we start from the image px4-gazebo-headless (https://github.com/JonasVautherin/px4-gazebo-headless), but we need to make some modifications to have enough flexibility. The new container project created aims to add more flexibility to change the SITL params when running docker image with the latest version of PX4 and to allow initialization in different conditions and configurations, becoming more suitable for a wider range of applications (https://hub.docker.com/r/andrekuros/px4-sitl-headless-flex).

Getting px4-headless-flex

docker pull andrekuros/px4-sitl-headless-flex

Test Drone Simulation (Single Run Example)

sudo docker run --rm -it px4-sitl-headless-flex -sysid 11 -speed 1\  
        -gip 192.168.0.6 -gport 14550 -aip 192.168.0.6 -aport 14542\  
        -param "MPC_XY_CRUISE 20" -param "MIS_DIST_1WP 3000"\  
        -param "MPC_XY_VEL_MAX 20" -param "MIS_DIST_WPS 3000"

Change the -gip to the IP of your ground station host (for example QGroundControl) to connect to the simulated drone.
See more options in https://github.com/andrekuros/px4-sitl-headless-flex.

Run your first Simulation

Using the ContainerNet you can start a simulated Mininet Network with multiple drones throught an Python. To do this first run the Containernet Docker Image an then run the python script.

sudo docker run --name containernet -it --rm --privileged --pid='host' -v /home/user/share:/containernet/examples -v /var/run/docker.sock:/var/run/docker.sock containernet/containernet  /bin/bash   

containernet> python3 test1.py

#Case: Initial Test - 03/07/2021
#test1.py
#Andre Kuros

from mininet.log import lg, info
from mininet.net import Mininet, Containernet
from mininet.topolib import TreeNet, TreeContainerNet
from mininet.node import Controller
from mininet.cli import CLI
from mininet.link import TCLink
from mininet.log import info, setLogLevel
from mininet.nodelib import NAT


sitlImage="andrekuros/px4-sitl-headless-flex"
simSpeed= "1"
gcsIp= "10.0.0.99"
mavsdkIp="10.0.0.99"
params  = " --param \"MPC_XY_CRUISE 20\" --param \"MIS_DIST_1WP 3000\" --param \"MPC_XY_VEL_MAX 20\" --param \"MIS_DIST_WPS 3000\""#--param \"IMU_INTEG_RATE 10\""

numCompanies = 3
dronesInCompanies = [3,3,3]
delaysCompaniesUAVs = ['0ms', '0ms', '0ms']
delaysCompaniesSwitch = ['0ms', '0ms', '0ms']
bwCompanies = [1000,1000,1000]

lg.setLogLevel('info')

net = Containernet( topo=None)
net.addController('c0')
s1 = net.addSwitch("s1")

h_gcs = net.addDocker("h_gcs", ip="10.0.0.99", dimage="mavsdk-kuros-gcs" , network_mode='none',  dcmd="/bin/bash", volumes=["shareAll:/shareAll"])
net.addLink(h_gcs,s1,cls=TCLink, delay='0ms', bw=1000)

#Create drones as hosts
for c in range(1, numCompanies + 1):
    sc = net.addSwitch("sc" + str(c))    
    net.addLink(sc,s1,cls=TCLink, delay=delaysCompaniesSwitch[c-1], bw=1000)
    for h in range(dronesInCompanies[c-1]):
        h_new = net.addDocker("c" + str(c) + "-n" + str(h) ,dimage=sitlImage, network_mode='none', dcmd="/bin/bash /root/entrypoint.sh --sysid " + str(10*c + h) +  " --gip " + gcsIp + " --gport 1455" + str(c-1) + " --aip " + mavsdkIp + " --aport 1454" + str(c-1) + " --speed " + simSpeed + params )
        net.addLink(h_new,sc,cls=TCLink, delay=delaysCompaniesUAVs[c-1], bw=bwCompanies[c-1])

net.build()

# Add NAT connectivity
#net.addNAT().configDefault()

#start Network
net.start()
CLI( net )

# Shut down network
net.stop()

---

The UTM Controller

The UTM Controller is the solution created to allow to test algorithms to control the Airspace, receiving informations from all drones in the simulation and sending commands and authorizations. The solution is divided in two main modules, the GCS_UTM and the GridView. The GCS_UTM has the background to communicate with the drones and the interfaces to interact with them. The GridView is an solution to allow the visualization of the Airspace and add the possibility to interact in realtime.

Using the GCS_UTM

The current version of the GCS_UTM was tested Windows or Linux and can be started inside or outside of the Mininet, allowing to monitor and control the drones from an external host.

To activate or not teh GridView, comment/umcomment "#Define GRIDVIEW" in the UTM_APP.cpp

Setting GCS_UTM

Some params allow to configure the GCS_UTM, they can be set in the file "config.json". Basically, at this moment, is only possible to set the params of each company and set to turn on/of the GridView (only for Windows).

Param	Description
gridView	on/off (1/0) !Is not working for now, use #Defnie GRIDVIEW in UTM_APP.cpp
companies	Array with the companies configurations
name	given name for the company
location	position in the airspace, representing (x,y) cells
cod	unique code for the company
port	port that the MavLink will be transmitted
timeout	time in seconds to consider that the connection is down

**Attention: the number of companies should match with the initialization in the ContainerNet python script

Config.json example:

{    
  "gridview" : 1,
  "companies": [
    {
      "name": "company1",
      "location": [ 2, 2 ],
      "cod": 1,
      "port": 14540,
      "timeout": 3
    },
    {
      "name": "company2",
      "location": [ 19, 3 ],
      "cod": 2,
      "port": 14541,
      "timeout": 3
    },
    {
      "name": "company3",
      "location": [ 15, 18 ],
      "cod": 1,
      "port": 14542,
      "timeout": 3
    }
  ]

}

---

Evaluation Tests

At this moment the model of the tests are hardcoded in C++ in the GCS_UTM.cpp at the Method GCS_UTM::runTests. The tests are called from the UTM_APP or from the GridView in case this is activated. The method GCS_UTM::generateStats create a CSV file to allow the analysis.

Running the tests:

• Call the pyhton scrypt to create the ContainerNet environment:

  • ex: python3 demo01.py

• Open the GCS_UTM

  • Wait to the Mavsdk to connect to all drones (with 30 drones in some cases 45 seconds):
    • With GridView -> Click "Update Systems"
    • Without GridView -> Press Enter
  • Wait for all systems to be recognized (with 30 drones in some cases 30 seconds)
    • With GridView -> Click "Run Test 1"
    • Without GridView -> Press Enter

The results will be in the file "name_of_the_test.csv", as example:

cod	company	actionRequests	actionSuccess	missionRequests	missionsSuccess	busy	cancel	timeouts
c1-n10	1	3	0	21	21	0	0	0
c1-n11	1	3	0	21	21	0	0	0
c1-n12	1	3	0	21	21	0	0	0
c1-n13	1	3	0	21	21	0	0	0
c1-n14	1	3	0	21	21	0	0	0
c1-n15	1	3	0	21	21	0	0	0

---

Some known errors:

When Running the Python ContainerNet Scrypt:

FileNotFoundError: [Errno 2] No such file or directory: '/etc/network/interfaces'

Solution:

sudo apt-get autoremove --purge ifupdown
sudo apt-get install ifupdown
sudo mn -c

Exception: Error creating interface pair (sc1-eth1,s1-eth1): RTNETLINK answers: File exists

Solution:

sudo mn -c

Version Docker <= 4.10 in Python

sudo python -m pip install docker==4.1.0