Dear all,
It’s my pleasure to announce a talk by my former postdoc Paul Furgale (ETH Zurich) this coming Friday 5th December about “Lifelong Visual Localization: Scaling up algorithms for lifelong mobile robot autonomy” It will take place in Huxley 217 at 5 pm. I’m sorry for the short notice, hoping that some of you might still make it to this certainly interesting talk!
Regards,
Stefan
Lifelong Visual Localization: Scaling up algorithms for lifelong mobile robot autonomy
Abstract:
The current state-of-the-art in mobile robot localization and mapping is
able to produce large-scale geometrically consistent maps representing
the world as perceived over a short window of time. The resulting maps
can be used for accurate localization of robots or mobile phones.
However, the evolution of a robot’s map over time—including
detection and representation of periodic changes in appearance, and the
remapping of places when appearances change—has not been adequately
addressed in the literature. As a result, localization performance
begins to drop as soon as a map is produced. Performance degradation
continues until the map no longer reflects the appearance of the world
and localization fails.
The near future should see the wide deployment of mobile service robots,
automated vehicles, and richly instrumented smartphones. This raises the
possibility that data collection for the maintenance of maps could be
crowdsourced. For example, as the cars navigate through the city, they
could assess their surroundings for novelty by comparing the expected
appearance (from the map) with the observed appearance. Raw sensor data
from novel experiences could be uploaded for a central server that would
then update the map to reflect the changes. This talk will focus on the
efforts of the Autonomous Systems Lab to build a pipeline to support
long term map maintenance from a wide variety of data streams to support
efficient and accurate localization both indoors and outdoors.
Bio:
Paul Furgale is the Deputy Director of the Autonomous Systems Lab at the Swiss Federal Institute of Technology in Zurich (ETH Zurich). His current research is focused on long-term autonomy for mobile robotic systems, including perception, mapping, localization, and planning over long timescales and in highly dynamic environments. He is the scientific coordinator for V-Charge, a European project and industry/academic collaboration that seeks to develop automated valet parking and charging of electric vehicles in mixed traffic. He received a PhD (2011) from the University of Toronto Institute for Aerospace Studies (UTIAS) where he developed algorithms to support over-the-horizon sample return for planetary exploration rovers as part of the Autonomous Space Robotics Lab. His PhD work was field tested in the Canadian High Arctic and subsequently integrated into several Canadian Space Agency rover prototypes.