VANDANA "VANDI" VERMA
My general research areas are Robotics, Artificial Intelligence, and Machine Learning. In particular, I am interested in pushing the limits of computational efficiency and optimality of modeling, state estimation and control of increasingly complex autonomous systems in uncertain environments. I am interested in the development of new fundamental theories, as well as in the physical demonstrations of their utility.
My thesis is titled Partile Filters for Robot Fault Diagnosis.
There are numerous examples that indicate that real-time methods that are capable of fault diagnosis and adaptive control in the presence of significant uncertainty hold the key to highly autonomous systems that are successful and appear intelligent (and are accepted by humans).
In my thesis, fault detection and identification is represented as a hybrid (discrete/continuous) state-estimation problem. An exact Bayesian solution for this problem is intractable. My thesis work is makes contributions to the theory of probabilistic state-estimation to solve this problem.
RECENT PROJECTS AND LABS
Robotic search for Life in the Atacama: This project aims to investigate life in extreme environments. I implemented on-board health monitoring for the rover to autonomously detect anomalies such as non-traversable terrain, insufficient power, unreasonable deviation from navigation plans, and dozens of other faults. I also participated in the field test in the Atacama desert in Chile in April 2003. Currently we are analysing data from the field test, building a new rover, and improving health monitoring and autonomous recovery from faults among other things.
OTHER PAST PROJECTS
|Sun-Synchronous Navigation||Mars Autonomy||Robotic Antarctic Meteorite Search|
Morphin is on Mars
with the MER rover Spirit!
...albeit much updated,
and now called GESTALT.
I developed the pan tilt camera,
...and infinite respect for anyone
who works with VxWorks.