Thesis Defense - Trajectory Bundle Estimation For Perception-Driven Planning

SHARE:

Event Speaker: 

Abraham Bachrach

Event Location: 

33-206

Event Date/Time: 

Friday, December 14, 2012 - 9:30am

When operating in unknown environments, autonomous vehicles must perceive and
understand the environment ahead in order to make effective navigation
decisions. Today, autonomous vehicles tend to rely exclusively on metric
representations built using range sensors to plan paths. However, such sensors
are limited by their maximum range, field of view, and occluding obstacles in
the foreground, thereby limiting the effectiveness of planners that depend on
them.

If we wish to develop autonomous vehicles that are able to navigate directly
toward a goal at high speeds through unknown environments, then we must move
beyond the simple range-sensor based techniques. We must develop algorithms
that enable autonomous agents to harness knowledge about the structure of the
world to interpret noisy and ambiguous sensor information, and make inferences
about about parts of the world that cannot be directly observed.

In this thesis we develop a new representation based around a library of
trajectory bundles, that makes this challenging task more tractable. Rather
than attempt to explicitly model the geometry of the world in front of the
vehicle (which can be incredibly complex), we reason about the world in terms
of abstract notions about what the vehicle can and cannot do. Trajectory
bundles provide a lens through which we can look at perception tasks for
navigation, allowing us to leverage machine learning tools in much more
effective ways. We develop a Bayesian filtering framework that enables us to
estimate a belief over which trajectory bundles are feasible based on the
history of actions and observations of the vehicle, resulting in improved
navigation performance.

Thesis Advisor: Nicholas Roy

Relevant URL(S):
For more information please contact: Abraham Bachrach, 510-541-5439, abachrach@gmail.com