IEEE Systems, Man and Cybernetics Magazine - October 2020 - 21

between the UGV's actual and desired heading angles as
expressed in the following formula:
}(k) = K pa e a (k) + K ia | e a (k) T + K da Te a (k).

(4)

Results
To test the formation controller, the open-source software
was used because it was easier to interface with the robots.
The first experiment focused on the leader UGV driving in a
straight line with the follower UGVs staying in formation. As
seen in Figure 12, the follower UGVs were successfully able
to follow the leader UGV when driving in a straight line.
The next experiment included the leader UGV driving
in an S-shaped pattern. As a result, the followers should
also drive in an S-shaped pattern tracking the path of the

0.6
Leader UGV
Follower UGV
Follower UGV

0.4

Y Position (m)

0.2
0
-0.2
-0.4
-0.6
-0.8
-2

-1

1
2
X Position (m)

Figure 12. The results of the formation controller

when the leader UGV is driving in a straight line.

1.5
1
0.5

Y Position (m)

-0.5
Leader UGV
Follower UGV
Follower UGV

-1.5
-2
-2.5
-3
-2

-1

1
2
X Position (m)

Figure 13. The results of the formation controller

when the leader UGV drives in an S shape.

Conclusion
As verified by the hardware experiments, VIVEPOSE is
a good system for indoor localization. Although it does
suffer from the radial drift issue, this problem is not as
significant as in the proprietary HTC Vive software. The
experimental results also show how VIVEPOSE can be
utilized for robotics applications, such as indoor formation control. In the future, we plan to compensate for
the radial drift problem by utilizing filtering techniques
such as the EKF and using a more advanced optimizer
for the position estimation. We also plan on improving
the formation controller by using intra-agent communication rather than relying solely on the VIVEPOSE position estimates.
Acknowledgments
This work was supported by the Air Force Research
Laboratory under grant FA8750-15-2-0116 and by the
Office of Student Development through a contract
with the North Carolina Agricultural and Technical
State University.

-1

leader. However, as seen in Figure 13, the formation controller does not work as well with this pattern. This result
is likely due to the following problems:
◆ the radial drift phenomena of the VIVEPOSE
◆ the FOV for the projected light patterns from the
lighthouse
◆ the lighthouse and Vive tracker positioning due to vehicle size.
Improvements to better mitigate these factors are
planned for future work, which include
◆ changing the optimization algorithm used for determining position
◆ altering the formation-control algorithm to better
account for the angular velocity of the leader UGV
◆ extending the distance between the lighthouse and
ground via a larger leader vehicle.
These considerations are necessary to develop a more
robust formation control with this system in a 2D planar
environment. The tests performed were operated in the
regular ambient lighting conditions of a laboratory environment, in the presence of both white and yellow ceiling
lights. Although the setup was primarily focused to serve
as a means to localize indoors, to test the robustness of
the system, the setup was moved outdoors and reliable
tracking with the sensors was achieved in an outdoor environment under the presence of ambient sunlight.

About the Authors
Jonathan Lwowski (jonathan.lwowski@gmail.com) is
with the Department of Electrical and Computer Engineering, The University of Texas at San Antonio. He is a Student
Member of IEEE.
Abhijit Majumdar (abhijit.g.majumdar@gmail.com) is
with the Department of Electrical and Computer Engineering,
O c tob e r 2020

IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE

IEEE Systems, Man and Cybernetics Magazine - October 2020

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - October 2020