Message Passing Algorithms for Scalable Multitarget Tracking

An adaptive version of the single-model joint anomaly detection and tracking filter, based on the multiple-model approach, has been derived to handle unknown parameters in the underlying vessel dynamics.

“This Bayesian framework has been extended in [34] where an adaptive version, based on the multiple-model approach [36], of the single-model joint anomaly detection and tracking filter has been derived to handle unknown parameters in the underlying vessel dynamics.”

A probabilistic joint anomaly detection and tracking methodology for sequential detection of maritime anomalous deviations and simultaneous vessel tracking allows a mathematical definition of nominal vessel behaviour based on the combination of the OU dynamic model and context information extracted from historical data.

“This method allows for a mathematical definition of nominal vessel behaviour based on the combination of the OU dynamic model and context information extracted from historical data.”

SPA-based MTT algorithms can achieve an attractive performance-complexity compromise, making them suitable for large-scale tracking scenarios involving a large number of targets, sensors, and measurements, and allowing their use on resource-limited devices.

“thereby, SPA-based MTT algorithms can achieve an attractive performance-complexity compromise (see [39] and references therein), making them suitable for large-scale tracking scenarios involving a large number of targets, sensors, and measurements, and allowing their use on resource-limited devices.”

In the probabilistic joint anomaly detection and tracking methodology, any anomalous deviation is represented as a switching unknown control input that modifies the nominal object dynamics, i.e., the OU mean velocity parameter.

“The key idea behind this approach is to represent any anomalous deviation as a switching unknown control input that goes into action by modifying the nominal object dynamics, i.e., the OU mean velocity parameter.”

An emerging approach to MTT and information fusion, based on a factor graph and the sum-product algorithm (SPA), has flexibility, low complexity and useful scalability.

“An emerging approach to MTT and information fusion - one with flexibility, low complexity and useful scalability - is based on a factor graph and the sum-product algorithm (SPA) [38], [39], [49].”

A major advantage of the SPA is its ability to exploit conditional independence properties of random variables for a drastic reduction of complexity.

“A major advantage of the SPA is its ability to exploit conditional independence properties of random variables for a drastic reduction of complexity”