AUTHOR=Carlson Logan , Navalta Dalton , Nicolescu Monica , Nicolescu Mircea , Woodward Gail 

TITLE=Early Classification of Intent for Maritime Domains Using Multinomial Hidden Markov Models

JOURNAL=Frontiers in Artificial Intelligence

VOLUME=4

YEAR=2021

URL=https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2021.702153

DOI=10.3389/frai.2021.702153

ISSN=2624-8212

ABSTRACT=<p>The need for increased maritime security has prompted research focus on intent recognition solutions for the naval domain. We consider the problem of <italic>early classification of the hostile behavior of agents in a dynamic maritime domain</italic> and propose our solution using multinomial hidden Markov models (HMMs). Our contribution stems from <italic>a novel encoding of observable symbols as the rate of change</italic> (instead of static values) for parameters relevant to the task, which enables the early classification of hostile behaviors, well before the behavior has been finalized. We discuss our implementation of a one-versus-all intent classifier using multinomial HMMs and present the performance of our system for three types of hostile behaviors (ram, herd, block) and a benign behavior.</p>