xvi PrefacePreface
AFRL ATR technology) whose guidance and thoughts propelled us to write this
book. We acknowledge DARPA program manager Dr. John Gorman for collabora-
tive research with the multi-faceted ATR deep learning (MADRLEARNING) and
Target Recognition and Adaption in Contested Environmnets (TRACE) programs
that were the first two major programs that embraced the DL technology for SAR
imagery classification. We also appreciate DARPA PM Mr. Paul Tilghman and his
SETA Dr. Esko Jaska for collaborative research on the RFMLS program. Over the
years, we were privileged to work with many great students of which their work is
referenced in the text. Specifically, we greatly appreciate the contributions of our
recent AFRL interns Nathan “Nate” Inkawhich and Matthew “Matt” Inkawhich
and their professors Dr. Yiran Chen and Dr. Hai Li. Nate and Matt are phenom-
enal doctoral students at Duke University that perfected the DL approaches. They
worked with us extensively on RF ML research. We considered them as “Innova-
tive Superstars” for the advancement of next generation AI/ML technology. We
also acknowledge the contributions of Chris Capraro, Eric Davis, and Darrek Is-
ereau of SRC, Inc. for contributions in the areas of SAR imaging and adversarial
machine learning, Dr. Mark Peot, Dr. John Spear of Teledyne Scientific and Imag-
ing for transfer learning research, and Prof. Dhireesha Kudithipudi of University of
Texas and Syed Humza of Rochester Institute of Technology for contributions in
near real-time training algorithms development. We are grateful to AFRL Informa-
tion Directorate leaderships, especially, Dr. Mike Hayduk, Mr. Joe Caroli, and Dr.
Don Telesca for their encouragement and support.
We appreciate the mentorship of Dr. Vince Velten, Dr. Mike Bryant, Dr. Mike
Minardi, Patti Ryan Westerkamp, and especially Dr. Tim Ross from the AFRL Sen-
sors Directorate–COMprehensive Performance Assessment of Sensor Exploitation
(COMPASE) center.
Current AFRL Sensors Directorate researchers contributing to SAR ATR re-
search include: Dr. Eric Branch, Dr. Theresa Scarnati, Benjamin Lewis, Dr. Christo-
pher Paulson, Daniel Uppenkamp, Elizabeth Sudkamp, Dr. Paul Sotirelis, Dr. Linda
Moore, Dr. LeRoy Gorham, Steven Scarborough, and Dr. John Nehrbass.
As the book captures many developments, lessons learned, and practical results
not least of which is the data, we acknowledge those who created the MSTAR vi-
sion. Parts of the MSTAR vision were shared by a number of individuals through-
out DoD, but the important fact was that Dr. Richard Wishner of DARPA was
primarily responsible for the belief, construction, and implementation of the vision.
He convinced Mr. Larry Lynn, the Director of DARPA, to fund the MSTAR pro-
gram and then constructed the program true to the tenets of his vision. The prefer-
ence for a model-based approach was motivated by the excellent work performed
by Professor Tom Binford at Stanford University. Both Richard Wishner and Ed
Zelnio were introduced to the model-based approach by Professor Binford.
On the government side, several individuals played key roles. Major Tom
Burns started as program manager for AFRL and then became the DARPA pro-
gram manager later in the MSTAR program. Mr. Martin Justice took over for Ma-
jor Burns and continued as program manager until the program ended. Mr. Mark
Minardi served as the program technical director, Dr. Bill Pierson served as the
ATR performance evaluation lead, and Dr. Tim Ross spearheaded the system engi-
neering function key to the program success. AFRL provided continuity through-
out the program while MSTAR was guided by four consecutive DARPA program