1.5
1
0.5
0 0 2 4 6 8
0.5
1
1.5
]s/dar[
yticolev
1
0.8
0.6 0.4
0.2
0 0.20 2 4 6 8
0.4 0.6
0.8
1
Joint 2 !!elbow
0!ms
2!ms
5!ms
7!ms
10!ms
time [s]
]s/dar[
yticolev
Joint 0 !!shoulder pan
0.001
0. 00!m00 s5
2!ms
5!ms0 0 2 4 6 8
7!ms 0.1000!m05s
time [s] 0.001
]s/dar[
yticolev
Joint 1 !!shoulder lift
0!ms
2!ms
5!ms
7!ms 10!ms
time [s]
2.5
2
1.5
1
0.5
0
0.50 2 4 6 8
1
1.5
2
]s/dar[
yticolev
Joint 3 !!wrist 1
0!ms
2!ms
5!ms
7!ms
10!ms
time [s]
2
1.5
1 0.5 0 0.50 2 4 6 8
1
1.5
2
]s/dar[ yticolev
Joint 4 !!wrist 2
1.5
1
0.05!ms 2!ms 50!m0s 2 4 6 8
0.75!ms
110!ms
time [s] 1.5
]s/dar[ yticolev
Joint 5 !!wrist 3
0!ms 2!ms 5!ms
7!ms
10!ms
time [s]
Figure 4: The velocity commands sent to the robot
50
45
40
35
30
25
20
15
10
5
0
0 2 4 6 8
sm
0
ot
ecnereffid
yticolev
detalumuc
]s/dar[
esac
ycnetal
pluginfitsintothemodulardesignofGazebo. Asofthe
interface is available, it eases to test various network
effects on the robot control. Based on our preliminary
evaluations it does affect the QoC KPIs of the robot
control.
The evaluation showed behavior which is expected and
reasonable, but also cases which show that the whole
systemneedsfine-tuning. Weplantoevaluatethework-
ing mechanism of the system with the help of the ROS,
gazebo and research communities. We plan to do more
extensivemeasurementswiththetool. Weplantointer-
faceitwithvariousradionetworksimulatorsandseethe
effects of the radio on the QoC KPIs. In a similar way,
weplantoinvestigatethehowthesystembehaveswhen
taking into account not only the network links charac- teristics but also the protocols for message exchanging.
We also plan to compare the level of similarity of the
simulation to real robot HW controlled in a real radio
network. WearetakingpartintheARIACcompetition
and we plan to evaluate if the tool can provide any ad-
vantageforusinanyoftheusecasesofthecompetition.
REFERENCES
Andersen T.T., 2015. Optimizing the Universal Robots ROS
2 ms driver. Tech. rep., Technical University of Denmark, Depart-
5 ms mentofElectricalEngineering.
7 ms
10 ms ARIAC,2017. Agile Robotics for Industrial Automation Compe-
tition(ARIAC). URLhttp://gazebosim.org/ariac.
DARPA, 2017. Network Emulation at
time [s] the DARPA Robotics Challenge. URL
https://iwl.com/white-papers/network-emulation-at-the-darpa-roboti
Figure5: The cumulateddifferenceofthe velocitycom-
mands comparing to the reference scenario Gazebo, 2017. Gazebo. URLhttp://gazebosim.org/.
IvaldiS.;PetersJ.;PadoisV.;andNoriF.,2014. Toolsforsimu-
latinghumanoid robot dynamics: Asurveybasedonuserfeed-
This is likely due to the various updating frequency pa- back. In 2014 IEEE-RAS International Conference on Hu-
rametersthatGazeboemploystorunthesimulation. It manoid Robots.842–849.
needs definitely further work to make it clear how the
MoveIt,2017. MoveIt. URLhttp://moveit.ros.org/.
introduced latency affects other characteristics or be-
haviors,suchastherobotcommandingfrequency,whole ns3,2017. ns-3. URLhttps://www.nsnam.org/.
physical simulation steps, internal message timings. nwnstc, 2014. Network Namespaces and Traffic Control. URL
Figure 5 shows the cumulated difference of the velocity http://gigawhitlocks.com/2014/08/18/network-namespaces.html.
commands comparing to the reference scenario. The
omnetros, 2017. How to link OM-
2ms latencyscenarioistheclosesttothe referenceasit NET++/Castalia with ROS. URL
isexpected. Inthefirst3secofthetrajectoryexecution http://cpham.perso.univ-pau.fr/WSN-MODEL/castalia-ros.html.
the5msscenarioisclosertothereferencethanthe7ms
OurPlugin, 2017. Gazebo latency plugin. URL
scenario,but around6 sec,the 5 ms scenariocollects so https://github.com/Ericsson/robot_hw_sim_latency.
much error that shows bigger deviation than the 7 sec
roscontrol, 2017. Data flow of ros control and Gazebo. URL
scenario. The 10 sec scenario has another magnitude of
http://gazebosim.org/tutorials/?tut=ros_control.
error,andthuscutoffthediagramafterthefirstsecond.
rostopics,2017.ROSTopics.URLhttp://wiki.ros.org/Topics.
CONCLUSION AND FURTHER WORK
UR5,2017. UR5. URLwww.universal-robots.com/ur5-robots.
URSim, 2017. URSim. URL
In this paper, we proposed a plugin (OurPlugin 2017)
https://www.universal-robots.com/download/?option=28545#section16
to extendthe capabilities ofthe currentGazeborobotic
simulatorandturnitintoaCPSsystem. Therealization
of the proposed method is a plugin to Gazebo. The