Multiple Extended Target Tracking
Target tracking is an important practical problem, particularly in surveillance and defense applications. One of the main problems encountered during target tracking is the association of measurements with unique targets. Additional complexity is added when multiple measurements can be obtained for each target; a common occurrence when using high resolution sensors. My work is to estimate the kinematic state and extent (size and shape) of each target in such scenarios. Furthermore, my work is concerned with estimating the number of targets in the surveillance region, and their associated probabilities of existence.
Simultaneous Localization and Mapping
In order for a vehicle to traverse through an environment safely, both the vehicle state and the state of each object in the environment must be known with certainty. Simultaneous Localization and Mapping (SLAM) attempts to obtain such knowledge, given only noisy measurements and control torques.
My research is concerned with deriving a robust SLAM method that can be used in complex dynamic environments; commonly referred to as the holy grail of autonomy within the academic community.