IEEE Robotics & Automation Magazine - December 2018 - 21

an inspection sensor, i.e., an ultra-
sound or eddy-current sensor to per-
form industrial contact inspections
● a ground interface, with a camera
and sensor device, allowing the
operator to monitor and modify all
inspection parameters.
The procedure to perform inspections
with this novel aerial vehicle is as follows.
1) The aerial robot takes off and ap-
proaches the area to be inspected.
This phase could be autonomous or
teleoperated and may eventually
benefit from the previously men-
tioned trajectory planning.
2) The machine comes into contact
with the pipe to be inspected. This
phase can also be autonomous or
teleoperated and benefits from the
perception system.
3) The operator switches the aerial
vehicle into the inspection mode,
during which the robot automatical-
ly maintains its position with
respect to the point of contact. From
this moment, the inspection opera-
tor has full control of the system.
a) The operator, looking at the cam-
era, can move the sensor along the
pipe, disregarding the aerial robot's
individual degrees of freedom,
benefiting from the task-priority
controller, and eventually resorting
to shared-control methods.
b) Once at the area of interest, the
operator uses the sensor control
interface to perform the inspection.
4) The operator can then move the end
effector to the next area of interest
or can finalize the inspection.
5) After the full inspection is complete,
the pilot switches back to the flying
mode and detaches the aerial robot
from the pipe.

Platform
Altitude (rad)

●

0.1
0
-0.1
700

705

710

715
Time (s)
Roll

720

725

720

725

720

725

720

725

Pitch

x Relative
Displacement
of the Platform (m)

(a)
-1.72
-1.74
-1.76
700

705

710

715
Time (s)

y Relative
Displacement
of the Platform (m)

(b)
0.3
0.25
0.2
0.15
700

705

710

715
Time (s)

z Relative
Displacement
of the Platform (m)

(c)
0.1
0.05
0
700

705

710

Reference

715
Time (s)
Estimation

Ground Truth

Euclidean
Position Error (m)

(d)
0.04
0.02
0
700

705

710

715
Time (s)
(e)

720

725

Lessons Learned
Figure 11. The results of an indoor autonomous contact operation telemetry experiment
Experiments in industrial settings can during which 17 strong perturbations up to 100 N were injected: (a) the small effect of
be performed only after the robustness the disturbances on the robot's attitude, (b) x control and estimation, (c) y control and
of the proposed solutions have been estimation, (d) z control and estimation, and (e) the robot's 3-D position estimated
error (on average lower than 1 cm, even with the perturbations).
extensively tested. In this testing, the
main lessons learned were the following.
● Wind and aerodynamic perturbations generated by
direction, as required for contact inspection. Compli-
nearby surfaces should be compensated for by the con-
ance in the arms is also highly relevant to maintaining
troller quickly enough to maintain safety and achieve
safety. Currently available light servocontrollers impose
accuracy and relative stability. Multidirectional-thrust
limitations on the accuracy that can be obtained by the
platforms are better suited for maintaining forces in any
end effector.
december 2018

*

IEEE ROBOTICS & AUTOMATION MAGAZINE

*

21
IEEE Robotics & Automation Magazine - December 2018

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - December 2018
