Paolo Pirjanian, Ph.D.
Chief Scientist,
Evolution Robotics, Inc.
Mauldin Auditorium (NSH 1305)
Refreshments 3:15 pm
Talk 3:30 pm
Over the recent years we have witnessed a growing number of
robotic products make their way into consumer markets. While
these products can arguably be viewed as robots, they provide
very little autonomous or intelligent functionality. All
robotics vacuum cleaners, for instance, use random navigation
for floor coverage because they cannot rely on odometry for
localization. Meanwhile, the research community is
developing sophisticated techniques for localization and
mapping. How can we close the gap between the technologies
that come out of the research labs and the technologies that
go into consumer and commercial products? I will discuss
the challenges of developing technologies for affordable,
useful, and reliable consumer robotic products. I will
present our solution to visual, simultaneous localization and
mapping called vSLAM. vSLAM is a first-of-a-kind,
breakthrough technology for localization and mapping that
requires only one low-cost camera and odometry (e.g., wheel
encoders). This solution provides up to 100 times reduction
in cost when compared to more traditional range-based SLAM
techniques which use sonars or laser range finders. Because
of its low cost and robust performance in realistic
environments vSLAM is particularly well-suited for use in
consumer and commercial products.
Paolo Pirjanian is the Chief Scientist at Evolution Robotics,
Inc and heads the R&D efforts in developing core technologies
for consumer and commercial robotics products. His main work
is focused around vision-based navigation, control
architectures, and human-robot interaction. Paolo is also a
part-time lecturer at the Computer Science department of
University of Southern California. Prior to joining
Evolution Robotics, Paolo initiated several research thrusts
on multi-robot systems for space exploration at the Jet
Propulsion Laboratory, NASA. This research resulted in a
multi-robot control architecture known as CAMPOUT and the
Cliffbot, an ensemble of three robots for high-risk access on
cliff-walls. He received his Ph.D. degree in robotics from
Aalborg University, Denmark where he also served as a research
professor.
For appointments, please
contact Illah Nourbakhsh (illah@cs.cmu.edu)
The Robotics Institute is part of the School of Computer Science, Carnegie Mellon University.