Terrain database interoperability issues in training with distributed interactive simulation

Abstract

In Distributed Interactive Simulation (DIS), each participating node is responsible for maintaining its own model of the synthetic environment. Problems may arise if significant inconsistencies are allowed to exist between these separate world views, resulting in unrealistic simulation results or negative training, and a corresponding degradation of interoperability in a DIS simulation exercise. In the DIS community, this is known as the simulator terrain database (TDB) correlation problem. This is part of the larger synthetic environment correlation problem in DIS, which includes atmosphere, ocean, space, and a wide variety of dynamic effects, behaviors and models. In this article, we investigate the terrain database correlation problem and the resultant effects on interoperability in DIS systems. The fundamental elements of terrain databases designed for real-time distributed simulation are introduced. A generic data pipeline for terrain database generation systems is developed for the purpose of illustrating causes of the correlation problem and issues of terrain database fidelity. Implications of the problem are discussed, and testing methodologies are recommended for its mitigation. Several statistical methods have been developed to analyze consistency between various elements of the synthetic environment across DIS platforms. Correlation metrics have been formulated for terrain elevations and features. Comparisons and consistency of final rendered images have been addressed. Finally, a suite of software tools that has been developed for interoperability investigations and visual comparison of terrain databases is presented.

A network-centric terrain database regeneration architecture

“…This article proposes a bi-directional procedural terrain database (TDB) generation architecture for future automated incremental updates of large heterogeneous TDB formats supporting modeling and simulation (M&S) within a network-centric military environment and describes the challenges facing it. Within the context of this article, an environmental database is an integrated set of data elements that describes a geographical region which may include terrain, natural and man-made features and atmospheric, oceanographic and space for M&S. 1–4 Our research focus is on the representation of terrain and man-made features within the aforementioned M&S network context. For demonstration purposes only, we implement a limited automated prototype of the theoretical architecture in a significantly scaled-down M&S network context and report on some of our findings.…”

“…Traditionally, military TDB formats have been generated using the one-way downstream pipeline paradigm shown in Figure 1, which gathers source data upstream, converts the source data using a terrain generation system, and publishes various target formats downstream for different M&S user needs. 4…”

“…13 Some of the correlation errors include numerical inaccuracy, as well as algorithmic, parametric, and temporal inconsistency. 4 De-confliction and conflation of features, 14 as well as feature-to-image mapping, may be necessary to ensure that features on images and vectors are correlated.…”

Real-time simulation of dust behavior generated by a fast traveling vehicle

“…In many virtual environments and distributed interactive simulations, it is desirable to simulate trucks, armored vehicles, bulldozers, and other ground-based moving objects [Balci 1997;Chen et al 1997;Cremer et al 1995;Schiavone et al 1997;IST 1994;Li and Moshell 1993]. Typically, dust behavior is not simulated when these objects travel on an unpaved road.…”

Considering Modeling and Simulation of DOD Response to Accidental Nuclear Disaster

“…Advances in simulation interoperability by the Department of Defense reduced stovepipes within the defense modeling and simulation community [32][33][34][35]. Contrasting with high levels of interoperability discussion within the defense community, nuclear industry discussion of interoperability training is limited [36,37] between the SCDAP/ RELAP5, MAAP, Rascal, and Sandia National Lab's MELCOR simulations that serve the nuclear community [38][39][40].…”