The Situation Awareness Window: a Hidden Markov Model for analyzing Maritime Surveillance missions

Abstract

In recent years, the use of Maritime Surveillance (MS) systems has increased in both defense and civilian domains. A demanding workload is placed upon operators of these systems, including the need to perform simultaneous information fusion from a number of sources to enable rapid decision throughput based upon Situation Awareness (SA). We have developed a method to objectively encode, summarize, and analyze airborne MS crew activities to gain insights into what is attended to in the execution of surveillance requirements. We label this method the “Situation Awareness Window” (SAW), which integrates sensor and tactical information with kinematics to define key attention and decision components of the operators that emerge over the surveillance mission. The SAW is defined with respect to the objects that are surveyed, the surveillance activities, and their chronological order. A SAW Hidden Markov Model (SAW-HMM) operates upon the surveillance mission activity encoder, resulting in a probabilistic relationship between the attention switching across sensor types and surveyed objects over the entire mission. That is, to implement the SAW-HMM we encoded the selection of sensors and surveillance decisions using a novel “encoder-interface” that allows users to probe many different features, observations, and states of a given mission. Ultimately the SAW will provide automated, objective, and insightful post mission debriefing technologies for operators and mission planners to encapsulate task demands and SA features over the mission.