NodeStar is a next-generation Target Motion Analysis (TMA) algorithm based on rigorous Bayesian mathematics. It combines uncertain information from multiple sources on a contact to produce probability distributions on target state which explicitly include uncertainty. The mathematical theory behind NodeStar is based on the multiplication of arbitrary non-linear likelihood functions into the target’s prior state combined with a maneuvering motion model for the target.

NodeStar refers to both an experimental codebase used in the Advanced Processor Build (APB) process as well as a derived operational software product. It processes information from acoustic sensors such as the sphere, MF active, hull, sail (HF), HF active, WLR9, Wide Aperature Array (WAA), TB16, TB23, and TB29 arrays as well as information from non-acoustic sensors such as radar, visual (periscope), and ESM. NodeStar is not limited to bearings information and can process range, range rate, bearing rate, inverse range, speed, aspect (AOB) and more. Input does not have to be accurate as long as the errors can be modelled.


The NodeStar experimental codebase is involved in ongoing efforts to incorporate new sources of information and bring more accurate error models to bear on existing sources. One area of research is detailed environmental modeling combined with realtime estimation of target source level. Another is investigating using NodeStar solutions in an operator feedback loop to leverage operator knowledge and expertise.