Extended target tracking, or ETT, is the problem of estimating the kinematic state and extent (size and shape) of each target in a surveillance region, given only noisy measurement sets. One of the main problems encountered during ETT involves the association of measurements with unique targets. This problem is further compounded when multiple measurement are obtained for each target; for example, the use of X-band radar in the tracking of surface-vessels. In this talk, we show how the integration of random matrices and finite set statistics alleviates such problems, leading to robust and reliable estimations.