©2001 CRC Press LLC
related to the estimation of target error and how to treat the codependence between sensors. They extend
this work to nonlinear systems in Chapter 13. Finally, in Chapter 14, Ronald Mahler provides a very
extensive discussion of multitarget, multisensor tracking using an approach based on random set theory.
Part III of this book addresses issues of the design and development of data fusion systems. It begins
with Chapter 15 by Ed Waltz and David L. Hall, and describes a systemic approach for deriving data
fusion system requirements. Chapter 16 by Christopher Bowman and Alan Steinberg provides a general
discussion of the systems engineering process for data fusion systems including the selection of appro-
priate architectures. In Chapter 17, David L. Hall, James Llinas, Christopher L. Bowman, Lori McConnel,
and Paul Applegate provide engineering guidelines for the selection of data fusion algorithms. In Chapter
18, Richard Antony presents a discussion of database management support, with applications to tactical
data fusion. New concepts for designing human-computer interfaces (HCI) for data fusion systems are
summarized in Chapter 19 by Mary Jane Hall, Sonya Hall, and Timothy Tate. Performance assessment
issues are described by James Llinas in Chapter 20. Finally, in Chapter 21, David L. Hall and Alan N.
Steinberg present the
dirty secrets
of data fusion. The experience of implementing data fusion systems
described in this section was primarily gained on DoD applications; however, the lessons learned should
be of value to system designers for any application.
Part IV of this book provides a taste of the breadth of applications to which data fusion technology
can be applied. Mary L. Nichols, in Chapter 22, presents a limited survey of some DoD fusion systems.
In Chapter 23, Carl S. Byington and Amulya K. Garga describe the use of data fusion to improve the
ability to monitor complex mechanical systems. Robert J. Hansen, Daniel Cooke, Kenneth Ford, and
Steven Zornetzer provide an overview of data fusion applications at the National Aeronautics and Space
Administration (NASA) in Chapter 24. In Chapter 25, Richard R. Brooks describes an application of
data fusion funded by DARPA. Finally, in Chapter 26, Hans Keithley describes how to determine the
utility of data fusion for C4ISR. This fourth part of the book is not by any means intended to be a
comprehensive survey of data fusion applications. Instead, it is included to provide the reader with a
sense of different types of applications. Finally, Part V of this book provides a list of Internet Web sites
and news groups related to multisensor data fusion.
The editors hope that this handbook will be a valuable addition to the bookshelves of data fusion
researchers and system designers. We remind the reader that data fusion remains an evolving discipline.
Even for classic problems, such as multisensor, multitarget tracking, competing approaches exist. The book
has sought to identify and provide a representation of the leading methods in data fusion. The reader
should be advised, however, that there are disagreements in the data fusion community (especially by
some of the contributors to this book) concerning which method is
best
. It is interesting to read the
descriptions that the authors in this book present concerning the relationship between their own techniques
and those of the other authors. Many of this book’s contributors have written recent texts that advocate
a particular method. These authors have condensed or summarized that information as a chapter here.
We take the view that each competing method must be considered in the context of a specific
application. We believe that there is no such thing as a generic data fusion system. Instead, there are
numerous applications to which data fusion techniques can be applied. In our view, there is no such
thing as a magic approach or technique. Even very sophisticated algorithms may be corrupted by a lack
of
a priori
information or incorrect information concerning sensor performance. Thus, we advise the
reader to become a knowledgeable and demanding consumer of fusion algorithms.
We hope that this text will become a companion to other texts on data fusion methods and techniques,
and that it assists the data fusion community in its continuing maturation process.