R-FISST and the Data Association Problem with applications to Space Situational Awareness Conference Paper uri icon

abstract

  • 2016, American Institute of Aeronautics and Astronautics Inc, AIAA. All rights reserved. In multi-object tracking the Data Association Problem (DAP) occurs when one receives noisy measurement data and is uncertain of which object the data belongs to. This is further complicated when considering situations where the measurement could be produced by a source of clutter or a previously unknown object. As the number of possible associations increases so does the number of possible hypotheses. This paper considers scenarios in which generating all possible hypotheses is computationally intractable. The authors propose using a Monte Carlo Markov Chain (MCMC) based random sampling technique called Randomized Finite Set Statistics (R-FISST) to sample and generate the highly probable hypotheses. This allows one to avoid exhaustively generating all hypotheses and keeps the multi-object tracking problem computationally tractable. A Space Situational Awareness (SSA) example is used to create the mentioned scenario and illustrate the benefits of the R-FISST technique.

name of conference

  • AIAA/AAS Astrodynamics Specialist Conference

published proceedings

  • AIAA/AAS Astrodynamics Specialist Conference

author list (cited authors)

  • Faber, W. R., Chakravorty, S., & Hussein, I. I.

citation count

  • 5

complete list of authors

  • Faber, Weston R||Chakravorty, Suman||Hussein, Islam I

publication date

  • January 2016