Hybrid high-fidelity modeling of radar scenarios using atemporal, discrete-event, and time-step simulation
Loading...
Authors
Cheng, Yuan-Pin
Subjects
Discrete Even Simulation (DES)
radar range equation
Simkit
Visual Simkit
atemporal model
hybrid sensor model
DES event graph
Nearly Orthogonal Latin Hyper Cube (NOLH)
regression analysis
cross-validation
radar range equation
Simkit
Visual Simkit
atemporal model
hybrid sensor model
DES event graph
Nearly Orthogonal Latin Hyper Cube (NOLH)
regression analysis
cross-validation
Advisors
Brutzman, Donald P.
Pace, Phillip E.
Date of Issue
2016-12
Date
Dec-16
Publisher
Monterey, California: Naval Postgraduate School
Language
Abstract
Many simulation scenarios attempt to seek a balance between model fidelity and computational efficiency. Unfortunately, scenario realism and model level of detail are often reduced or eliminated due to the complexity of experimental design and corresponding limitations of computational power. Such simplifications can produce misleading results. For example, Radar Cross Section (RCS) effects in response to time-varying target aspect angle are dominant yet often ignored. This work investigates a new approach. A hybrid, high-fidelity sensor model can be achieved by using a Time-Step (TS) approach with precomputed atemporal response factors (such as RCS), each situated on active simulation entities that interact within an overall Discrete Event Simulation (DES) framework. This work further applies regression analysis to the cumulative results of 1,281 design points multiplied by 100 replications each to provide additional insight from massive results. Through a rigorous cross-validation (CV) for ensuring proper model selection, this new methodology adapts the best aspects of temporal control by each simulation approach, integrating multiple high-fidelity physically based models in a variety of tactical scenarios with tractable computational complexity. The meta-model generates statistically identical results as the baseline hybrid sensor model in mean with higher variation. This outcome sensor model performs in an efficient DES architecture by predicting only probability of detection (Pd) and time-to-detect (TTD) of the targets with the consideration of the range and the aspect relationship between the sensor and the target. The range and the aspect factors were ignored in current existing DES sensor models due to the overly simplified detection algorithm and the constraint of modeling complexity.
Type
Thesis
Description
Series/Report No
Department
Computer Science (CS)
Modeling, Virtual Environments, and Simulation Institute (MOVES)
Organization
Identifiers
NPS Report Number
Sponsors
Funder
Format
Citation
Distribution Statement
Approved for public release; distribution is unlimited.
Rights
Copyright is reserved by the copyright owner.