invited Speakers

Fred Daum, Raytheon Company
Erik Blasch, AFRL/RI
Pierre Valin, DRDC Valcartier
Tony Ponsford, Raytheon Canada

fred daum

Industrial Strength Nonlinear Filters


The extended Kalman filter (EKF) is the workhorse algorithm for tracking, navigation, multi-sensor data fusion, robotics and many other applications. The EKF often has good estimation accuracy, but it is surprisingly bad for some important real World systems. Particle filters hold the promise of improved accuracy, but at the cost of extremely high real time computational complexity. Unfortunately, particle filters generally suffer from the curse of dimensionality. In this talk, we will show the root cause of this problem, and we will explain how to improve the situation by roughly seven orders of magnitude. We have solved the well known and important problem of “particle degeneracy” using a new theory, called particle flow. Our filter is four orders of magnitude faster than standard particle filters for any given number of particles, and we require many orders of magnitude fewer particles to achieve the same filter accuracy. Therefore, the net reduction in computational complexity is roughly seven orders of magnitude relative to standard particle filters for the same accuracy. Moreover, our filter beats the EKF accuracy by several orders of magnitude for difficult nonlinear problems. We show many numerical results for various nonlinearities, with both stable and unstable plants, varying process noise, measurement noise, initial uncertainty of the state vector, and dimension of the plant from d = 1 to 30. Our theory uses exact particle flow to compute Bayes’ rule, rather than a pointwise multiply. We do not use resampling or proposal densities or importance sampling or any other MCMC method. Resampling is the bottleneck to efficient parallel processing for standard particle filters, whereas our filter never resamples, and hence it has no bottleneck for parallel processing. We design the
particle flow with the solution of a linear first order highly underdetermined PDE, like the Gauss divergence law in electromagnetics. We study over a dozen methods to solve this PDE, mostly inspired by physics and fluid dynamics. This talk is for normal engineers who do not have log-homotopy for breakfast.


Fred Daum is an IEEE Fellow, and he is a graduate of Harvard University . Fred is also a senior principal Fellow at Raytheon. Fred was awarded the Tom Phillips prize for technical excellence, in recognition of his ability to make complex radar systems work in the real world. He developed, analyzed and tested the real time algorithms for essentially all the large long range phased array radars built by the USA in the last four decades. These real time algorithms include: extended Kalman filters, radar waveform scheduling, Bayesian discrimination, data association, track initiation, discrimination of satellites from missiles, calibration of tropospheric and ionospheric refraction, and target object mapping. Fred's nonlinear filter theory has been applied by engineers at Boeing for the boost phase intercept problem, with results that are vastly superior to the extended Kalman filter. Fred's nonlinear filter theory generalizes the Kalman and Beneš filters. He has published nearly one hundred technical papers, and he has given invited lectures at MIT, Harvard, Yale, Brown, Georgia Tech., University of Connecticut, Univ. of Minnesota, Melbourne Univ., Univ. of New South Wales, Univ. of Illinois at Chicago and Northeastern University.

erik blasch

Performance Evaluation of Seismic and Acoustic Track and Identification Fusion


To determine the quality of information fusion systems, there is a need to investigate methods of performance evaluation to include data quality, models, system performance, and gains from information fusion processing. Methods of performance evaluation for Level 1 object assessment include the simultaneous track and identification (STID) of targets. The performance evaluation of STID results support Level 4 sensor management and Level 5 user refinement and decision making. In this presentation, we discuss metrics, data sets, and evaluation methods that support emerging techniques in performance evaluation of MOVINT characteristics with references to ongoing work. Applications include radar, electro-optical, and seismic/acoustic STID methods of various tracking methods along with classification techniques. For tracking, we compare linear belief filtering approaches and non-linear particle filtering approaches. For identification methods, we discuss information theory, support vector machines, and belief methods. The presentation motivates and seeks to initiate coordination of performance evaluation methods of the information fusion community to include data sets, metrics, and evaluation techniques.


Erik Blasch is currently on exchange to Defence R&D Canada at Valcartier, Quebec in the Future Command and Control (C2) Concepts and Structures Group of the C2 Decision Support Systems Section. Prior to the sabbatical, Dr. Blasch was the Information Fusion Evaluation Tech Lead for the Air Force Research Laboratory - COMprehensive Performance Assessment of Sensor Exploitation (COMPASE) Center (AFRL/RYAA) and an Adjunct Electrical Engineering and Biomedical Engineering Professor at Wright State University (WSU) and the Air Force Institute of Technology (AFIT) in Dayton, Ohio. He is also a reserve Maj with the Air Force Office of Scientific Research (AFRL/AFOSR) in Washington, DC. He was a founding member of the International Society of Information Fusion (ISIF) in 1998 and the 2007 ISIF President. Dr. Blasch has focused on Automatic Target Recognition, Targeting Tracking, and Information Fusion research compiling 300+ scientific papers and book chapters. He is active in ISIF, IEEE (AES and SMC), and SPIE societies. Dr. Blasch received his B.S. in Mechanical Engineering from the Massachusetts Institute of Technology in 1992 and Master’s Degrees in Mechanical (‘94), Health Science (‘95), and Industrial Engineering (Human Factors) (‘95) from Georgia Tech and attended the University of Wisconsin for an MD/PHD in Mech. Eng/Neurosciences until being called to Active Duty in 1996 to the United States Air Force. He completed an MBA, MSEE, MS Econ, MS/PhD Psychology (ABD), and a PhD in Electrical Engineering from Wright State University and is a graduate of Air War College. He is a Fellow of SPIE. He recently completed a trip to Gaspesie.

Pierre Valin

High Level Information Fusion Developments, Issues, and Grand Challenges


This talk will present the evolution of fusion levels from the early days of the JDL, the current DFIG view, and the future that is promised with the newly created ISIF working group FPMFWG. From this discussion, one can adopt a definition of High-Level Information Fusion (HLIF) that aims to provide Situation Awareness and perform Threat Assessment. The Canadian context of NEOps (NCW or NEC in other countries) will be discussed, in particular the CF mission-critical outcomes related to HLIF. The role of Decision Support Systems, the OODA loop for operations and the intelligence cycle will be presented, as well as its relevance to other formalisms. The role of ontologies in recognition and identification (R&I) will be stressed. Issues with reliability and credibility according to various standards (NATO and US) will be highlighted, as they impact the quality of fusion through MOPs and MOEs. HLIF grand challenges, tradeoffs and the State Transition Data Fusion will be explained. Finally a presentation of the C2DSS DRDC Valcartier section and its groups will show that they are ideally structured to address HLIF problems through their mandate. A short overview of the methodology, the tools used and the M&S approach, and the integration in a new LIDS facility to showcase the results of ARPs and TDPs, will conclude the talk.


Dr. Pierre Valin received his Ph.D. in Theoretical High Energy Physics from Harvard in 1980, and then taught physics in universities before joining Lockheed Martin Canada in 1993. In 2004, he became a Defence Scientist at DRDC Valcartier, where he currently leads a research group in Future C2 Concepts & Structures, and is Thrust Leader for Air Command at DRDC since 2007. Dr. Valin has been particularly active in ISIF though the organization of FUSION 2001 and 2007, as an ISIF board member since 2003, VP-Membership since 2004, and President in 2006, as well as associate editor for JAIF. His interests include Multi-Sensor Data Fusion requirements and design, use of a priori information databases, imagery classifiers and their fusion, reasoning techniques for recognition and identification, distributed information fusion, dynamic resource management.

tony ponsford

Effective Maritime Domain Awareness Based on Appropriate Layered Surveillance and a Multilevel Decision Support System


An effective Maritime Domain Awareness system for tracking both collaborative and non-collaborative vessels, within a multi-agency environment, is presented. Operational examples of the “Force Multiplier” effect that be achieved are presented.

This presenation reviews the requirements and options for providing effective Maritime Domain Awareness (MDA) from the shore line out to, and beyond, the 200 nm Exclusive Economic Zone (EEZ). It is shown that effective MDA can be accomplished by implementing a Decision Support System (DSS) system based on a collaborative, layered approach. The MDA system takes advantage of the ability to provide wide-area persistent surveillance by new and emerging sensors such as High Frequency Radar and Space-based interception of Automatic Identification broadcasts.

A value proposition is presented for the usefulness of different sensor and data sources, and their combinations, based on specific mission requirements. A focus is placed on the value of real-time, persistent surveillance of the EEZ using High Frequency radar in association with other sensor and data sources. It is shown that when associated within the MDA system that anomalous or unusual vessel behaviour can be identified and an appropriately response taken.

Effective MDA requires improved collaboration between multiple-Government agencies that operate at Local, Regional and National levels as well as interacting with the private sector and international agencies. The architecture of DSS is discussed in the context of the requirement for tactical, operational and strategic operations in a collaborative multi agency environment.


Dr. A.M. (Tony) Ponsford is an Engineering Fellow and Technical Director Maritime Domain Awareness (MDA) at Raytheon Canada. Tony joined Raytheon Canada Limited in 1991 and established the company’s MDA group building on his experience with High Frequency Surface Wave Radar (HFSWR) and Integrated Maritime Surveillance (IMS) technology as applied to surveillance of the 200 nautical mile Economic Exclusion Zone (EEZ). Raytheon Company (NYSE: RTN), with 2009 sales of $25 billion, is a technology and innovation leader specializing in defense, homeland security and other government markets throughout the world. With headquarters in Waltham, Mass., Raytheon employs 75,000 people worldwide.

Tony’s career in Maritime Domain Awareness started whilst serving in the Merchant Marine where he worked with Shell Tankers in developing concepts of MDA. In the 1980’s, working as a research associate at the University of Birmingham, he initiated the development of HFSWR for persistent surveillance of a nation’s 200 nautical mile (nm) Exclusive Economic Zone (EEZ).

Dr Ponsford moved to Canada in 1987 and joined NORDCO Limited in St. John’s, Newfoundland as the Senior Scientist, Manager and Technical Director of the newly formed Integrated Maritime Surveillance business unit. In this capacity, Tony established Canada’s first HFSWR test bed facility at Cape Bonavista. This development progressed into the world’s first shore-based, real-time, EEZ surveillance sensor that provided continuous, all-weather, tracking of ships, icebergs and aircraft throughout the EEZ. Two prototype systems were subsequently deployed on the Canadian East Coast. A third transportable unit was also produced and evaluated on a number of Canadian and US programs. The success of the Canadian demonstration has resulted in a number of international sales.

In 1977 Tony graduated with distinction from Plymouth Navy College (UK). In 1982 he earned a Bachelor of Science degree in Maritime Technology, graduating with first class honours, from the University of Wales Institute of Science and Technology. Whilst working at the University of Birmingham (UK), Tony was awarded a doctorate in philosophy under special regulations of the University, in recognition of his pioneering work in High Frequency Surface wave Radar.

Dr Ponsford is a Senior Member of the IEEE and is Co-Chair of the IEEE AESS Ottawa Chapter. He is also the General-Chair of the Organizing Committee for IEEE Radar Conference 2013.