Estimation for autonomous mobile robot navigation
localization and mapping
localization for mobile sensor networks
sensing and estimation under processing and communication
My Ph. D. thesis focuses on resource-aware
estimation for single- and multi-robot localization/mapping. In
the presence of time-varying processing and communication-bandwidth
constraints, the objective of my work is to design estimators
that can optimally utilize these time-varying resources in order to
maximize estimation accuracy.
2013: I am back to working on SLAM! Specifically, I am working on batch
MAP estimators (commonly known as bundle adjustment) using inertial
sensors and cameras and/or depth sensors. We are currently
developing a novel and promising keyframe-based batch-MAP approach
called C-KLAM (Constrained Keyframe Localization and Mapping) for
real-time localization and mapping. We will be presenting our
preliminary results at the "Long-Term Autonomy: Navigation and Mapping for Real-World Applications" workshop on 10 May 2013 at ICRA in Karlsruhe, Germany.
2013: Currently, I am working on Covariance Intersection algorithms for
multi-robot cooperative localization. Our proposed algorithm, submitted
to IROS 2013, can handle both limited processing resources and
asynchronous communication constraints.
Dec 2012 : Recently, I have been working on the extremely interesting
and challenging problem of designing estimators capable of
processing a combination of analog and quantized measurements.