Surface Composite Tracker Component Navy SBIR FY2014.1 Sol No.: Navy SBIR FY2014.1 Topic No.: N141-036 Topic Title: Surface Composite Tracker Component Proposal No.: N141-036-0375 Firm: Lakota Technical Solutions, Inc. PO Box 2309 Columbia, Maryland 21045 Contact: William Farrell Phone: (410) 381-9780 Web Site: www.lakota-tsi.com Abstract: Near-shore, littoral surface tracking is challenging due to the dynamic and inhomogeneous sea surface clutter as well as a diverse and dense target environment that leads to incomplete, non-contiguous, intermittent, and degraded tracking ability of any single sensor. Composite tracking provides a way to achieve a more continuous, complete, and unambiguous track picture utilizing data from multiple sensors. Lakota proposes to develop a composite tracker, the Adaptive Multi-frame Parameterized Tracker (AMPT), which is innovative in its ability to adapt to a wide range of ocean environments, target densities, and target types. AMPT provides a novel solution by uniquely combining the following algorithmic techniques: Multi-Frame Data Association (MFA), Sequential Probability Ratio Testing (SPRT), Interacting Multiple Model (IMM) state estimator, Covariance Intersection (CI), and Maximum Likelihood Activity Estimation (MLAE). The MFA algorithm employs a maximum time-depth sliding window of data from each sensor source to associate its data with the composite track picture. Each sensor's cost function considers the sensor's statistical characteristics (contact vs. track) and estimates of the local clutter/track density to dynamically select and integrate information from different sensors into the composite track picture. The SPRT is a modified version of a Neyman-Pearson hypothesis testing procedure for track confirmation/initiation that uses adaptive test thresholds and an additional penalty term inspired by the Minimum Description Length (MDL) principle for Information Encoding. To support a wide range of potential target dynamics, an IMM state estimator is employed with a unique combination of Kalman filters and Covariance Intersection filters. Finally, MLAE adaptively estimates local clutter/track densities for SPRT threshold selection and cost function calculations. AMPT will improve Situation Awareness (SA) within the maritime littoral environment by generating a surface track picture that is more complete, continuous, unambiguous, accurate, and precise than its contributing sensors. Benefits: The proposed AMPT approach is beneficial because it implements an efficient multi-hypothesis tracker (MHT) using an adjustable sliding window of sensor inputs with heterogeneous sensor performance characteristics and statistical correlations. Additionally, AMPT is parameterized in order to decouple the sensor-specific characteristics from the core processing including data association, track management, and state estimation. This enables rapid and non-invasive introduction of new sensor sources while still maintaining an efficient MHT implementation. Finally, AMPT is beneficial because it estimates inhomogeneous clutter and track densities and adapts the data association and track confirmation parameters across the battlespace to achieve uniform performance. Return