Conference Program

Fusion 2007 Technical Program (Updated June 30)

Monday, 9 July 2007

Room 410 Leduc 418 Morrice Lismer Borduas Suzor-Côté
  Tutorials Workshops
08:00 Ontology Based Situation Awareness and High Level Fusion: Methods and Tools
Mieczyslaw M. Kokar		 Advances and Applications of DSmT for Information Fusion
Jean Dézert, Florentin Smarandache		 Logic-based Approaches to Information and Knowledge Fusion
Éric Grégoire		 Introduction to Social Network Analysis and Complex Networks
Pontus Svenson		 Distributed Sensor Network
Office of Naval Research (ONR) 		 
09:30 Coffee Break in room 414 and in foyer room
10:00 Ontology Based Situation Awareness and High Level Fusion: Methods and Tools
continued		 Advances and Applications of DSmT for Information Fusion
continued		 Logic-based Approaches to Information and Knowledge Fusion
continued		 Introduction to Social Network Analysis and Complex Networks
continued		 Distributed Sensor Network
continued		  
11:30 Lunch time (Lunch NOT INCLUDED)
12:30 Evaluation of Information Fusion Systems
Erik Blasch		 Multitarget Tracking and Multisensor Fusion – Part 1
Yaakov Bar-Shalom		 Possibilistic Information Fusion
Henri Prade		 Distributed Robotic Sensor Systems
Katia Sycara, Bin Yu		 Distributed Sensor Network
continued		 Data/Information Fusion Research in Canadian Academia
Mathematics of Information Technology and Complex Systems and Lockheed Martin Canada
15:30 Coffee Break in room 414 and in foyer room
16:00   Multitarget Tracking and Multisensor Fusion – Part II
Yaakov Bar-Shalom		 Computational Approaches to Situation Assessment and Decision Support
Subrata Das		 What is risk? What is probability?
Glenn Shafer		   Data/Information Fusion Research in Canadian Academia
continued
19:00 Cocktail in Foyer room


Tuesday, 10 July 2007

  Borduas Krieghoff Suzor-Côté Leduc Fortin Pilot Foyer
07:00 Breakfast in Place Montcalm room
08:00 Plenary: New Challenges for Defining Information Fusion Requirements, James Llinas (CMIF)     Poster Session A
09:30 Break
09:20 Combination in Evidence Theory
Arnaud Martin		 Tracking - 1
Peter Willett		 Nonlinear Estimation and Filtering - 1
Shozo Mori		 System Design - 1
Felix Opitz		 Sensor Management - 1
Emmanuel Duflos, Philippe Vanheeghe
10:40 Coffee Break in Foyer room
11:00 Evidence Theory and Fuzzy Sets
Frédéric Dambreville		 Tracking - 2
Chee Chong		 Nonlinear Estimation and Filtering - 2
Branko Ristic		 System Design - 2
Erik P. Blasch		 Sensor Management - 2
Emmanuel Duflos, Philippe Vanheeghe
12:20 Lunch in Place Montcalm room
13:20 Optimization – 1
Alain Appriou		 Tracking - 3
Pramod Varshney		 Nonlinear Estimation and Filtering - 3
Olivier Bilenne		 C4ISR
Angela Pawlowski		 Collaborative Sensor Systems
Katia Sycara, Bin Yu
14:40 Break
15:00 Optimization – 2
Vesselin P. Jilkov		 Tracking - 4
Stefano Coraluppi		 Nonlinear Estimation and Filtering - 4
Pierre Valin		 Security and Safety
Li Bai		  
16:00 Coffee Break in Foyer room
16:20 to 19:20 Panel : Results from Levels 2/3 Fusion Implementation: Issues, Challenges, Retrospectives and Perspectives for the Future, Ivan Kadar      


Wednesday, 11 July 2007

Borduas Krieghoff Suzor-Côté Leduc Fortin Pilot Foyer
7:00 Breakfast in Place Montcalm room
08:00 Plenary: Knowledge Propagation : A Federative Look at Developments in the Framework of Belief Function Theories, Alain Appriou (ONERA)     Poster Session B
09:00 Break
09:20 Surveillance and Situation Analysis – 1
Stefano Coraluppi		 Track Fusion
Huimin Chen		 Particle Filtering - 1
Yvo Boers		 Distributed Systems- 1
Katia Sycara		 Image Fusion and Video Fusion Assessment - 1
Stavri Nikolov
10:40 Coffee Break in Foyer room
11:00 Surveillance and Situation Analysis – 2
Alan Steinberg		 Tracking - Theory
Mahendra Mallick		 Particle Filtering - 2
Gee Wah Ng		 Distributed Systems- 2
Mitch Kokar		 Image Fusion and Video Fusion Assessment - 2
Allen Waxman
12:20 Lunch in Place Montcalm room
13:20 Surveillance and Situation Analysis – 3
Jean Roy		 Tracking - Missing Data
Roy Streit		 Detection - Localization
Frederik Gustafsson		 Multiple Classifiers
Robert Lynch		 Information Fusion for Intelligent Transportation Systems
Malte Ahrholdt
14:40 Break
15:00 Surveillance and Situation Analysis – 4
Pontus Svenson		 Tracking - Maneuvering Targets
Erik Blasch		 Detection - Sensor Networks
Peter Willett		 Context Information in Data Fusion
Jesus Gracia, Jose Molina		 Information Fusion in Bioinformatics
Subrata Das
16:20 Coffee Break in Foyer room
16:40 to 18:40 Panel : Agent Based Information Fusion, Subrata Das      
19:15       Aperitif
19:45 Banquet Dinner in the Grand Ballroom


Thursday, 12 July 2007

  Borduas Krieghoff Suzor-Côté Leduc Fortin Pilot
07:00 Breakfast in Place Montcalm room
08:00 Plenary: System Architecture for Dynamic Information Fusion: Dynamic Matching and Meta Perception, Masatoshi Ishikawa (U. of Tokyo)    
09:00 Break
09:20 Information Modeling
Laurence Cholvy		 Data Association
Darko Musicki		 Image Fusion – 1
Allen Waxman		 Situation Management - 1 Modeling
Gabriel Jakobson, Lundy Lewis, John Salerno		 High Level Information Fusion on the Cognitive Arena - 1
Gee Wah Ng
10:40 Coffee Break in Foyer room
11:00 Information Retrieval
Langis Gagnon		 Multitarget Tracking
Dale Blair		 Image Fusion – 2
John Irvine		 Situation Management – 2 Reasoning
Gabriel Jakobson, Lundy Lewis, John Salerno		 High Level Information Fusion on the Cognitive Arena - 2
Gee Wah Ng
12:20 Lunch in Place Montcalm room
13:20 Semiotics and Ontologies
Eric Dorion		 Navigation, Positioning and Guidance
Jean-Pierre Le Cadre		 Image Fusion – 3
Stavri Nikolov		 Information Fusion and Pattern Recognition - 1
Anne-Laure Jousselme, Patrick Maupin		 Data Fusion Technologies for the Canadian Armed Forces - 1
Mark Edwards
14:40 Break
15:00 Probabilities and Statistics
Gregor Pavlin		 Resource Management
John Salerno		 Image and Information Fusion for Remote Sensing
Alexandre Jouan		 Information Fusion and Pattern Recognition - 2
Anne-Laure Jousselme, Patrick Maupin		 Data Fusion Technologies for the Canadian Armed Forces - 2
Mark Edwards
16:20 Coffee Break in Foyer room
16:20 to 19:20 Panel : Distributed Sensor Networking, Rabinder Madan    



DETAILED PROGRAM : attached




Special Sessions

Fusion 2007 technical program features regular and special sessions that are scheduled from 10-12 July 2007. Panel discussions and tutorials are scheduled for the first day of the conference (please refer to the appropriate item on the web site menu for details).

Prospective authors are encouraged to submit papers to the special session which theme is the most appropriate to their work. Papers submitted for special sessions will undergo the same review process than those submitted to the regular program. They should be submitted according to the same deadlines: 15 february 2007 for the draft version of the paper and 15 may 2007 for its final version.

To get more details about the special sessions, authors are encouraged to contact directly the session chairs listed below.


Special session 1: Data Fusion Technology for the Canadian Armed Forces

Session chair:
discMark Edwards, General Dynamics Canada, Email: Mark.edwards@gdcanada.com

Objective:
The objective of this session is to explore the application of Data Fusion to Military needs. As with many other fields of modern endeavor, there are times when there is too much information to sift through to make proper informed decisions. This session will address the application of techniques to various operational requirements to show the real world use of Data Fusion.

Abstract:
There are many ways that Data fusion can support military operations in areas of Situational Awareness, surveillance, and planning tasks. These requirements call upon all levels of fusion capability. Accurate and efficient management of information on the battlefield is vital for successful military operations. The process of collecting, collating, integrating and interpreting data must be achieved whilst ensuring that no information is lost. The critical area is to look at the tactical requirements for data fusion, and to ensure that the technology is supporting the military in performing their tasks in a more simplified fashion. The purpose of this session is to look at the practical application of data fusion to Canadian military requirements in a number of areas as outlined in the topics of interest below. The primary goal is to make the bridge between the theoretical aspects of data fusion and the practical aspects.

Keywords:
Military, tactical, operational, fusion, joint, Army, Navy Air Force


Special session 2: Image Fusion and Video Fusion Assessment

Session chairs:
discAlexander Toet, TNO Defense, Safety and Security, Soesterberg, The Netherlands, Email: lex.toet@tno.nl

discStavri Nikolov, Dept. of Electrical and Electronic Engineering, University of Bristol, UK, Email: stavri.nikolov@bristol.ac.uk

Objective:
The focus of this session will be on state-of-the-art methodologies, metrics and tools used to assess the information content or quality of fused imagery and video, and to evaluate the performance of image fusion and video fusion systems.

Topic of interest:
The fusion of multimodality (multi-spectral) image and video sources is emerging as a vital technique for surveillance purposes, navigation and object tracking applications, and in medical diagnostics. The main goal of image fusion is to provide a single compact representation of the input images that is more informative than each of the individual inputs. There are several potential benefits of multi-sensor image fusion: wider spectral, spatial and temporal coverage, extended range of operation, decreased uncertainty, improved reliability and increased robustness of the system performance. The combination of complementary information from a range of different sensing modalities can provide enhanced performance for visualisation, detection, tracking, identification or classification tasks.

The widespread use of image fusion methods has led to a rising demand of pertinent quality assessment tools in order to compare the results obtained with different algorithms and systems or to derive an optimal setting of parameters for a specific fusion algorithm. For man-in-the-loop applications, the performance of the fusion algorithm can be measured in terms of improvement in operator performance in different tasks like detection, recognition and classification. This approach requires a well defined task, for which quantitative measurement can be made, and it usually involves costly and time consuming field trials. Computational image fusion quality assessment metrics that relate to human observer performance are therefore of great value. The assessment can either be done by comparing the fused result with a reference image that provides the ground-truth, or (since such ground-truth is not available in most applications) by relating the fused result (or some of its features) to each of the input images (the so-called non-reference approach). Video fusion assessment is even more challenging as the spatio-temporal characteristics of the inputs need to be taken into account. This special session comprises eleven contributed papers that present some of the latest research in image and video fusion assessment and its various applications. The focus of the proposed special session will be on assessment of dynamic fused imagery.

Return to Special Sessions


Special session 3: Sensor Management

Session chairs:
discEmmanuel Duflos, INRIA-Futurs, Ecole Centrale de Lille, France, Email: emmanuel.duflos@ec-lille.fr

discPhilippe Vanheeghe, INRIA-Futurs, Ecole Centrale de Lille, France, Email: philippe.vanheeghe@ec-lille.fr

Objective:
To propose an overview of the last main research results and of their applications in this field. Sensors management must be tackled using several scientific fields like sensors modeling, optimization, Bayesian inference, detection theory, random sets, communications, sensors and data fusion or more recently reinforcement learning. Sensors management has many applications in military and civil domains. The three objectives of this session are to:

Keywords:
Sensors Management, Multisensor Systems, Data Fusion, Markov Decision Process, Reinforcement Learning, Random Sets, Bayesian Inference.


Special session 4: Collaborative Sensor Systems

Session chair:
discKatia Sycara, Carnegie Mellon University, USA Email: katia@cs.cmu.edu

discBin Yu, Quantum Leap Innovations, Inc., USA Email: byu@quantumleap.us

Objective:
Collaborative sensor systems of the near future are envisioned to consist of hundreds of unmanned vehicles such as UAVs and UGVs. These networked autonomous and geographically distributed sensors play strong roles in military and civilian operations, e.g., battlefield surveillance and disaster rescue. At the same time, sensor systems offer many exciting research challenges due to their real-world constraints such as imperfect sensor data, real-time execution, and scarce wireless communication bandwidth.

The purpose of this special session is to bring together the researchers to address the challenges and opportunities of collaborative sensor systems in mission-critical applications. The session will solicit papers addressing both theoretical and practical aspects of sensor systems, as well as their relation to level two fusion (situation awareness) and level three fusion (threat assessment). In particular, the workshop aims at bridging the gap between the vast amount of theoretical work on AI, control theory and operation research and the practical needs of sensor systems research.

The topics of interest include but are not limited to:

Keywords:
Collaborative sensor systems, data fusion, task allocation, path planning

Return to Special Sessions


Special session 5: Sensor Data Fusion for Intelligent Transportation Systems

Session chair:
discDr. Malte Ahrholdt, Volvo Technology Corporation Dept. 6320, Intelligent Vehicle Technologies 40508 Göteborg, Sweden, Email: malte.ahrholdt@volvo.com

Keywords:
Sensor Data Fusion – Automotive Applications – Intelligent Transportation Systems (ITS) – Advanced Driver Assistance Systems (ADAS) – Track-level data fusion – Early fusion – Multilevel Fusion –– Object Tracking – Data Fusion Architecture – Sensor Data refinement – Object Refinement – Situation Refinement – Data Fusion Test Methods

Abstract:
Sensors like radar, lidar or video systems are more and more employed in vehicles for advanced driver assistance systems. In this session, methods are discussed how to combine information from different automotive sensors to achieve a reliable common perception of the environment. This fusion of information about the vehicle and its surrounding is one of the basic requirements of future Intelligent Transportation Systems.

Topic of interest:
Today, more and more vehicles are equipped with sensors (like radar, lidar or video systems) observing the close environment. Maps and communication devices are used as an additional source of information. The aim is to use the knowledge about the surrounding to enable the driver to use his vehicle more comfortably and safely even in difficult situations. Proposals for Intelligent Transportation Systems aim at having an “intelligent electronic co-driver” assisting and warning the driver or even intervening when necessary.

To achieve this, a profound knowledge of the vehicle surrounding is required. In a first approach, information of a single sensor has often been directly linked to vehicle applications. As complexity and demands for reliability rise, often more than a single sensor is employed. Then the question arises how to fuse these different sources of information to a common understanding of the environment in the best way. Some solutions are already available, but further progress is needed to fulfil all requirements.

In the proposed session, different concepts and methods for combining information from automotive sensors shall be discussed. A general objective is to derive a representation of road and objects in the vehicle’s surrounding based on a set of sensor measurements. Different methods to achieve this common goal will be explained and discussed in the session.

The processing of automotive sensor data can be structured into sensor data refinement, object refinement and situation refinement. Different aspects of these processing steps shall be addressed in this session.

An additional point of interest it the development process of data fusion systems for ITS purposes. Here, one question is how to validate the result of a fusion system, e.g., as compared to a single sensor, but also with respect to expected detection properties and false alarm rates.

Using data fusion systems in automotive application implies some particular requirements. On one hand side, information complexity and density varies a lot depending on the type of considered traffic scenario and application. On the other hand, demands on perception reliability and improved real-time capability constantly increase due to the development of new active safety functions.

By taking these requirements of vehicular sensor data fusion systems into special account, this session shall contribute to an exchange of recent research developments within Intelligent Transport Systems and the broader scope of the information fusion community.

Organizer’s biography:
Malte Ahrholdt received a Ph.D. degree from the Hamburg University of Technology in Germany in the area of sensor signal processing in 2005. Since 2004, he is with the Intelligent Vehicle Technologies group at Volvo Technology in Gothenburg, Sweden. His focus are automotive advanced driver assistance application and particularly sensing and data fusion techniques. He is involved in the European PReVENT research initiative “ProFusion2” on data fusion and is coordinator of the Swedish IVSS project “SEFS – Sensor Data Fusion for Automotive Safety Systems”.

Return to Special Sessions


Special session 6: Context Information in Data Fusion

Session chairs:
discJesus Garcia, Grupo de Inteligencia Artificial Aplicada, Universidad Carlos III, Madrid, Spain, Email: jgherrer@inf.uc3m.es

discJose Manuel Molina, Grupo de Inteligencia Artificial Aplicada, Universidad Carlos III, Madrid, Spain, Email: molina@ia.uc3m.es

Abstract:
Contextualized Information Fusion groups a large variety of disciplines: distributed intelligence, data and information communication, software design, computer vision, speech recognition, robotics, information fusion, etc. The broad concept of “Ambient Intelligence” (AmI) should be here understood as one where mobile services, whichever they are, are provided to the final user in a transparent, ubiquitous, continuous and personalized way. In a broad sense, AmI paradigm deals with pervasive systems assisting the users by autonomously interpreting their intentions and trigger automatic reactions. It is focused on intelligent environments, "walls with eyes and ears" to "sense" the users and aid them in their daily tasks.

The basic idea for developing AmI application is of a distributed layered architecture enabling ubiquitous communication and an advanced human–machine communication protocol. A wide range of application domains can benefit from AmI research as appear in the European Commission document entitled Scenarios for Ambient Intelligence in 2010 written by the IST Programme Advisory Group (ISTAG) for the Fifth European Framework Programme (1998–2002). An updated version of the continuously evolving vision of the report appeared in 2003 under the title Ambient Intelligence: from vision to reality, putting forward ideas for the currently open Sixth Framework Programme (2002–2006). A system based on AmI criteria requires a bundle of technologies to implement modularity, low-power devices, distributed and high band-width heterogeneous networks of sensors and actuators. Modular units of intelligence (agents), will create a distributed layer of intelligence, from the sensing to the understanding of information. Machine understanding includes research in the fields of computer vision, machine learning, situation assessment, and any other techniques applied by the community working on information fusion.

From the conceptual point of view, this session intends to contribute in works to the DATA FUSION approach associated to ambient intelligence, from the “user/network-centric” approach of cooperation and resource-sharing on which distributed systems are based (“shared situational awareness”, users networks and ad-hoc networks). In particular, effective exploitation of contextual information represented as knowledge-bases (KB), e.g. environmental maps and characteristics of the targets of interest is a powerful tool to gain adaptability and improved performance in the fusion process and inferences about interesting events. An effective use of the KB can be exploited at various levels of the fusion architecture algorithms so as to significantly reduce the number of false alarms, missed detections, and false tracks and improve true target track life (system robustness).

Objective:
The focus of this special session is on the Data Fusion techniques within the AmI applications. This is a fundamental step to create a first level of user-awareness in the environments. The main topics would include:

Return to Special Sessions


Special session 7: High-level Information Fusion on Cognitive Arena

Session chair:
discGee Wah, NG, Advance analysis and Fusion Laboratory, DSO National Laboratories, Email: ngeewah@csail.mit.edu

Objective:
To address the issue on high-level information fusion such as situation assessment and impact assessment, in the cognitive domain.

There are increasing interests in research on high-level information fusion, particularly at the situation assessment and impact assessment. Under situation assessment, the goals include deriving higher-order relations and identifying meaningful events and activities. Impact assessment includes estimating the level of threat or danger, predicting possible outcomes of particular decisions, determining the vulnerabilities of one’s own assets and determining possible courses of actions. It is inevitable that the high-level information fusion will cover the cognitive arena, such as cognitive intelligence, cognitive architectures, cognitive decision aids, sensemaking and bio-inspired intelligent fusion process for situation awareness and impact assessment.

Topics appropriate for this special session include (but are not necessarily restricted to):

Keywords:
High-level Information Fusion, Cognitive-based Situation Assessment and Impact Assessment, Cognitive Intelligence for Intent Inference, Cognitive Architecture for high-level fusion, Multi-source integration for Sensemaking, Bio-inspired high-level information fusion.

Return to Special Sessions


Special session 8: Situation Management

Session chairs:
discGabriel Jakobson, Altusys Corp, US, Email: jakobson@altusys.com

disc Lundy Lewis, Southern New Hampshire University, US

disc John Salerno, AFRL, US, Email: john.salerno@rl.af.mil

Abstract:
Situation Management (SM) is a fairly new term applied to the management of complex dynamic systems. People, whose interest is in situations happening in dynamic systems, are more familiar with the terms of Situation Awareness, Situation Calculus, Situation Semantics, and Situational Control. Situation Management as a management concept, technology, and a reference to the domain of applications, is taking roots from all of the above-mentioned disciplines, but it also aims to build a comprehensive and more general framework of concepts and enabling technologies for understanding and affecting situations that are happening or might happen with certain systems in a specific domain of interest. The objective of this session is to provide research and applications forum for advancing the SM technology, especially the issues related to high level (cognitive). The session continues the discussions started at the SM sessions at FFUSION 2006 in Florence, Italy.

Our interest will be mostly on cognitive aspects of situation management, i.e. on aspects related to the meaning of situations, the intelligent methods of reasoning about the situations, and action planning. In order to exhibit such intelligent capabilities, the systems should possess fairly elaborate conceptual knowledge about the domain, i.e. domain ontology.

Technical Scope:

Return to Special Sessions


Special session 9: Information Fusion and Pattern Recognition

Session chairs:
discAnne-Laure Jousselme,  DRDC-Valcartier, Decision Support Systems, Canada, Email: Anne-Laure.Jousselme@drdc-rddc.gc.ca

discPatrick Maupin, DRDC-Valcartier, Decision Support Systems, Canada, Email: Patrick.Maupin@drdc-rddc.gc.ca

Objective:
This session proposes an overview of the recent trends in pattern recognition applications and related information fusion problems. More specifically the session will discuss the problems involving the relationships existing between diversity measures, margin theory, individual voting and collective decision schemes as well as multi-objective techniques used for the design of pattern recognition systems. The problem of pattern recognition in situation analysis will also be stated, formalized and discussed.

Fusion 2007 participants are invited to contact the session chairs and propose original work on the topics listed above. 

Keywords:
Voting schemes, Multi-objective optimization, Systems design, Uncertainty and knowledge formalization, Situation Analysis  

Return to Special Sessions


Tutorials

Schedule

Monday, 9 July 2007
 
08:00-11:30 a.m. J. Dezert & F. Smarandache É. Grégoire P. Svenson M. Kokar
12:30-3:30 p.m. Y. Bar-Shalom I H. Prade K. Sycara & B. Yu E. Blasch
4:00-7:00 p.m. Y. Bar-Shalom II G. Shafer S. Das  


Tutorials Description
 
Yaakov Bar-Shalom Mieczyslaw M. Kokar
Multitarget Tracking and Multisensor Fusion - Part I Ontology Based Situation Awareness and High Level Fusion: Methods and Tools
 
Yaakov Bar-Shalom Henri Prade
Multitarget Tracking and Multisensor Fusion - Part II Possibilistic Information Fusion
 
Erik Blasch Glenn Shafer
Evaluation of Information Fusion Systems What is risk? What is probability?
 
Subrata Das Pontus Svenson
Computational Approaches to Situation Assessment and Decision Support Introduction to Social Network Analysis and Complex Networks
 
Jean Dezert and Florentin Smarandache Katia Sycara and Bin Yu
Advances and Applications of DSmT for Information Fusion Distributed Robotic Sensor Systems
 
Éric Grégoire  
Logic-based Approaches to Information and Knowledge Fusion  

Multitarget Tracking and Multisensor Fusion

Yaakov Bar-Shalom

Distinguished IEEE AESS Lecturer, Univ. of Connecticut, USA

Box U-2157

Storrs, CT 06269-2157

USA

Phone: 860-486-4823

Fax: 860-486-5585

Email: ybs@ee.uconn.edu

Objectives:
To provide to the participants the latest state-of-the art techniques to estimate the states and classifications of multiple targets with multisensor information fusion. Tools for algorithm selection, design and evaluation will be presented. These form the basis of automated decision systems for advanced surveillance and targeting. The various information processing configurations for fusion are described. A number of practical problems in multisensor tracking/fusion are also discussed.

Eligibility:
Engineers/scientists with prior knowledge of basic probability and state estimation. This is an intensive course in order to cover several important recent advances.

OUTLINE

Part I

Introduction
(OV) Overview of the course.

Review of the Basic Techniques for Tracking
[mttvf07: 1.5.1-1.5.3] The Kalman, the Alpha-Beta(-Gamma) and the Extended Kalman filters: their capabilities and limitations.

Tracking Targets with Multiple Behavior Modes
[mttvf07: 1.5.4] The Interacting Multiple Model (IMM) estimation algorithm - a real-time implementable, self-adjusting variable-bandwidth, tracking filter.

Multiple Hypothesis Tracker (MHT) and Multidimensional Assignment (MDA)
[290v] The score function in the MHT and its use with MDA.

Air Traffic Control Tracking
[200C] IMM vs. KF on real data (800 targets, from 5 FAA/JSS radars). How to evaluate estimation improvement without knowing the ground truth. Why multisensor tracking is cheaper computationally than single sensor tracking.

Multisensor Data Fusion
[mttvf07: 8.2] Information Processing Configurations in Multisensor Tracking.
Type I: Single sensor or reporting responsibility.
Type II: Single sensor tracking followed by track-to-track fusion.
Type III: Measurement-to-measurement association followed by central dynamic association and tracking.
Type IV: Centralized association and tracking.
A Hybrid Configuration: hierarchical sensor/platform/center setup.

Part II

Multisensor Data Fusion (Cont'd)
[mttvf07: 8.3,8.4] Common origin testing and fusion of local tracks. [286v] Multisensor track-to-track association for tracks with dependent errors.

Use of Classification Information in Tracking
[267] Tracking and data association with kinematic and classification information.

Tracking and Radar Management
[mttvf07: 1.8] Agile beam radar allocation: adaptive revisit time for minimum radar energy with the IMM.
[mttvf07: 3.4.11] Tracking in Clutter: the Probabilistic Data Association filter (PDAF).
[186A] Benchmark Problem for high-g targets in the presence of ECM (RGPO and jamming). Detection threshold, waveform, and revisit time selection, target RCS and jammer power estimation and tracking with the IMMPDAF. Comparison with the MHT (Multiple Hypothesis Tracker). The real-time experiment with an Aegis SPY-1 and F-14s at Wallops.

The course is based on the book Multitarget-Multisensor Tracking: Principles & Techniques by Y. Bar-Shalom and X.R. Li (YBS Publishing, 1995) and additional notes.

Background text: Y. Bar-Shalom, X. R. Li and T. Kirubarajan, Estimation with Applications to Tracking and Navigation: Algorithms and Software for Information Extraction, Wiley, 2001.

Biography:
Yaakov Bar­-Shalom received the B.S. and M.S. degrees from the Technion, Israel Institute of Technology, in 1963 and 1967 and the Ph.D. degree from Princeton University in 1970, all in electrical engineering. Currently he is Board of Trustees Distinguished Professor in the Dept. of Electrical and Computer Engineering and Marianne E. Klewin Endowed Professor in Engineering at the University of Connecticut and Fellow of IEEE. His current research interests are in estimation theory and target tracking and has published 7 books, over 350 papers and book chapters in these areas. He has been consulting to numerous companies and government agencies, and originated the series of Multitarget­ Multisensor Tracking short courses. He served as General Chairman of FUSION 2000 and President of ISIF in 2000 and 2002. He is co­recipient of the M. Barry Carlton Award for the best paper in the IEEE Transactions on Aerospace and Electronic Systems in 1995 and 2000 and in 2002 he received the J. Mignona Data Fusion Award from the DoD JDL Data Fusion Group.

Return to schedule

Evaluation of Information Fusion Systems

Erik Blasch, PhD, MBA

AFRL/SNA, 2241 Avionics Cir,

Wright-Patterson Air Force Base, Ohio 45433

Adjunct Professor, Wright State University,

Department of Electrical Engineering, Dayton, Ohio

Phone: (937)-904-9077,

Fax: (937)-255-1122

Email: erik.blasch@wpafb.af.mil

This course provides attendees with a basic working knowledge of integrating Data, Sensor, and Information Fusion systems into implementable architectures by confirming performance. The course concentrates on Fusion algorithmic designs and performance evaluation techniques for contribution in a multi-sensory systems. Many practical and useful examples are included throughout with metrics, experiments, and evaluation considerations. You will become knowledgeable with how to effectively design Fusion systems for many varied applications and users. The course price includes the documentation presented by the instructor.

This course will enable you to:

Intended audience:
This material is intended for anyone who needs to learn how to integrate fusion designs in future information fusion architectures which will rely on multi-sensory inputs and useful outputs. Those who either design their own Fusion subsystems or who work with information fusion designers will find this course valuable.

Course level: Intermediate

OUTLINE

Lesson 1: Introduction - Importance of Sensor Fusion

Lesson 2: Sensors and Architectures - JDL Levels 1 - 5, Decision making

Lesson 3: Reducing the error /uncertainty, compensating for unknown

Lesson 4: Fusion System Evaluation - Fusion Gain, ROC Curve, Uncertainty Analysis

The tutorial is intended to cover the general topics concerning data and information fusion use with emphasis on pragmatic fusion techniques. By highlighting the advantages and disadvantages of fusion, one can assess the gain of using multiple sources of information. Many users are excited about the prospects of the concept called "fusion"; however, if applied inappropriately, could result in detrimental effects. To mitigate the embellished opportunities of fusion, it is important to correctly evaluate fusion systems and appropriately quantify the "fusion gain".

By introducing fusionists to metrics and evaluation techniques, one can determine if combining multiple sources of information is justifiable. The first session introduces the importance of fusion, as in using different sensors to capture observable quantities about the environment. The second session details the larger organization of fusion systems as machines that process large quantities of data and reduce the dimensionality for a user to reason and make decisions over. The third topic highlights the fact that any information uncertainty must be reduced for the user. In the third session, mathematical techniques to reduce uncertainty are explained and developed within complex system design. If the user is better able to make decisions with appropriate mathematical models for information fusion, data collection, and timeliness of results; the fusion system is doing its job.

The final session details the process by which a fusionist can determine a "fusion gain" for user decision making and verifiable situational awareness. A fusion evaluation process will be explained which includes

After attending the tutorial, the student would better appreciate the complexity surrounding defending a realistic efficacy in fusion system performance, knowledge of where to get information in developing a fusion test plan, and practical examples and performance evaluation studies that include end users and field tests of such systems in complex environments.

Biography:
Dr. Erik Blasch (PhD, MBA) is a Fusion Evaluation engineer at Air Force Research Lab as the Government Fusion Technical Evaluation Lead for the COMprehensive Performance Assessment of Sensor Exploitation (COMPASE) which comprises about 60 government and contractors supporting various projects. Dr. Blasch is a co-founder of the International Society of Information Fusion (ISIF), adjunct professor in the Electrical Engineering at Wright State University, and a Major in the Air Force Reserves supporting the Air Force Office of Scientific Research (AFOSR). He has been involved in information fusion design and engineering for over 15 years. He has authored over 180 papers and been associated with many fusion programs within extended international community. Further detailed resume information can be found on line or upon request.

Return to schedule

Computational Approaches to Situation Assessment and Decision Support

Dr. Subrata Das

Chief Scientist

Charles River Analytics, Inc.,

625 Mt Auburn St., Cambridge, MA 02138,

USA

Phone: 617 491 3474 x547,

Fax: 617 868 0780

Email: sdas@cra.com

Abstract:
This tutorial is intended to provide a detailed understanding of both the cutting-edge and the most commonly used computational approaches to situation assessment (a.k.a. level 2 fusion) and the associated generation of appropriate response recommendations for decision making under uncertainty.

A diverse range of level 2 fusion techniques, including Bayesian belief networks, fuzzy logic, and the theory of belief function, will be covered in the tutorial. These techniques are suitable for modeling uncertain knowledge under different situations, and their suitability will be discussed in each case. A number of temporal modeling techniques, such as dynamic belief networks and hidden Markov models will also be presented. The tutorial will discuss interactions between level 1 and level 2 fusion processes. Special emphasis will be given to a discussion on particle filtering techniques as unifying methods for both filtering under level 1 fusion and inferencing in dynamic Bayesian networks for Level 2 fusion. Finally, distributed level 2 fusion techniques, as appropriate within network centric warfare environments, will be discussed.

For the response recommendations part of the tutorial, traditional expected utility theory, rule-based expert systems, and influence diagram based decision-making processes will be described. Then a symbolic argumentation technique using first-order and non-classical modal logics will be presented. Various techniques for aggregating arguments including probability, possibility, and Dempster-Shafer theories will be covered. The argumentation technique and probabilistic aggregation are the major focus of the speaker’s recent book on symbolic decision-making and a forthcoming book on the foundations of decision-making agents.

As for software tools, an in-house 5th generation application development platform (Prolog and Lisp) and argumentation building engine (Reason), and an in-house belief network engine (BNet@Builder, with its temporal extension) will be used for illustrating response recommendations and situation assessment respectively. The commercial-off-the-shelf tools Matlab and Hugin will be used for illustrating Kalman/particle filtering for level 1 fusion, fuzzy inferencing, and influence diagrams for decision-making. Military examples and prototype demos involving tasks of determining relationships among entities and events, target classification, and target identification will be provided throughout the tutorial.

Prerequisite:
Background in the basics of probability and statistics, mathematics, and a basic understanding of artificial intelligence.

OUTLINE

Architectures – JDL Model and other architectures

Military and Homeland Security Scenarios – Conventional, MOUT, OOTW, Bioterrorism

Foundational Technologies – Probability and Statistics, First-Order and Modal Logics

Brief Introduction to Level 1 Fusion – Data Association, Single and Multi-target Tracking, Kalman Filtering and Extensions, Particle Filtering, Rao-Blackwellised Filtering

Situation Assessment – Bayesian Belief Networks, Message Passing and Junction Tree Algorithms, Theory of Belief Function, Fuzzy Logic, Hidden Markov Model, Dynamic Belief Networks, Approximate Inferencing via Particle Filtering

Decision Making – Expected Utility Theory, Rule-based Expert Systems, Influence Diagrams, Dempster-Shafer Theory, Certainty Factor, Symbolic Argumentation

Foundational Tools – Bayesian Belief Network Engine, 5th Generation Application Development Environment (Prolog and Lisp), Argumentation Building Engine (Reason)

Applications – Problem Modeling using Foundational Tools

Biography:
Dr. Subrata Das is the Chief Scientist at Charles River Analytics, Inc. (www.cra.com)in Cambridge, MA. Subrata leads research projects in the areas of high-level and distributed information fusion, decision-making under uncertainty, intelligent agents, planning and scheduling, and machine learning. His technical expertise includes mathematical logics, probabilistic reasoning including Bayesian belief networks, symbolic argumentation, particle filtering, and a broad range of computational artificial intelligence techniques. Subrata held research positions at Imperial College and Queen Mary and Westfield College, both part of the University of London. He received his PhD in Computer Science from Heriot-Watt University in Scotland and a M.Tech. from the Indian Statistical Institute.
Subrata has published many journal and conference articles. He is the author of the book entitled “Deductive Databases and Logic Programming,” published by Addison-Wesley, and has co-authored the book entitled “Safe and Sound: Artificial Intelligence in Hazardous Applications,” published by the MIT Press. His forthcoming book entitled, “Foundations of Decision Making Agents: Logic, Modality, and Probability,” is due shortly for publication by the World Scientific/Imperial College Press.

Subrata is a member of the editorial board of the journal, “Information Fusion”, published by Elsevier Science. He is in the process of editing a special issue for the journal relating to agent-based information fusion. Subrata has been a regular contributor, a technical committee member, and a tutorial lecturer at each of the last three International Conferences on Information Fusion. He has been invited to join the technical committee at the forthcoming fusion conference to be held in Quebec City this year, and to be part of the guest panel that will discuss lessons learned from level 2 fusion system implementations. He has also been giving a series of tutorials on multi-sensor data fusion on behalf of the Technology Training Corporation.

Return to schedule

Advances and Applications of DSmT for Information Fusion

Dr. Jean Dezert

ONERA

(French National Establishment for Aerospace Research)

France

Email: jean.dezert@onera.fr; jdezert@yahoo.com


Dr. Florentin Smarandache

University of New Mexico

USA

France

Email: smarand@unm.edu; fsmarandache@yahoo.com

http://www.gallup.unm.edu/~smarandache/DSmT.htm

Abstract:
The combination of information is a hot topic of research specially in the development of complex systems involving imprecise, uncertain and potentially highly conflicting information/data with usually (but not necessarily) human interaction at some higher fusion level for efficient decision-making. Modern multisensor systems for tracking, classification, diagnosis, situation assessment, etc need solid theoretical tools to combine efficiently information in order to reduce as best as possible ignorances and contradictions in a coherent way to help to take proper decision. This task is very difficult and many theories (probability theory, possibility theory, Dempster-Shafer theory (DST), etc) have been proposed to deal with different kinds of uncertainties (randomness, fuzziness, epistemic nature, etc). After a brief reminder of classical combination rules based on belief functions used up to now in most of (non Bayesian) multisensor/expert systems, a detailed presentation of foundations and advances obtained in the development of Dezert-Smarandache theory (DSmT) for the combination of uncertain, imprecise and potentially highly contradicting sources of information will be given. DSmT appears as a natural extension of DST because DSmT takes into consideration any kind of model (free, hybrid DSm models and also the classical Shafer's model) according to the integrity constraints of the fusion problem. DSmT proposes a new mathematical framework and rules for information fusion that potentially allows some intersections of elements of the frame (i.e. some degree of consensus between elements). Fusion rules developed in DSmT framework overcome limitations of Dempster's rule and its alternatives as it will be showed in very simple examples. DSmT appears well adapted to static or dynamic fusion applications represented in terms of belief functions based on the same unified general mathematical formalism. The mathematical level of this tutorial and didactic examples will be kept as simple as possible to show the advantages of this new approach over previous ones. Aside basis of DSmT, we will present the recent Proportional Conflict Redistribution (PCR) rules and show their performances on several examples and will present also a new general arithmetic for the fusion of qualitative beliefs. An introduction to new quantitative belief conditioning rules will be also presented. A direct extension of the quantitative/numerical information fusion and conditioning rules to their quantitative counterparts in order to deal with qualitative information drawn from human sources and expressed in natural language will complete this tutorial.

Bonus material:
The attendees of this tutorial will received a printed copy of the presentation and as bonus material a printed copy of the last book published by the authors entitled « Advances and Applications of DSmT for Information Fusion », Collected Works, Vol.2, Rehoboth, July 2006.

DSmT Web page : http://www.gallup.unm.edu/~smarandache/DSmT.htm

Biography:
Jean Dezert was born in l'Hay les Roses, France, on August 25, 1962. He received the electrical engineering degree from the Ecole Française de Radioélectricité Electronique and Informatique (EFREI), Paris, in 1985, the D.E.A. degree in 1986 from the University Paris VII (Jussieu), and his Ph.D from the University Paris XI, Orsay, in 1990, all in Automatic Control and Signal Processing. During 1986-1990 he was with the Systems Department at the French Arerospace Research Lab (ONERA), Châtillon, France, and did research in multisensor multitarget tracking (MS-MTT). During 1991-1992, he visited the Department of Electrical and Systems Engineering, University of Connecticut, Storrs, U.S.A. as an European Space Agency (ESA) Postdoctoral Research Fellow. During 1992-1993 he was teaching assistant in Electrical Engineering at the University of Orléans, France. Since 1993, he is senior research scientist in the Information Modelling and Processing Department (DTIM) at ONERA. His current research interests include autonomous navigation, estimation theory, stochastic systems theory and its applications to MS-MTT, information fusion, plausible reasoning and non-standard Logics. Dr. Jean Dezert is developing since 2001 with Professor Smarandache a new theory of plausible and paradoxical reasoning for information fusion (DSmT) and has edited two textbooks (collected works) devoted to this new emerging research field published by American Research Press, Rehoboth in 2004 and 2006 respectively. He owns one international patent in the autonomous navigation field and has published several papers in international conferences and journals. He coauthored a chapter in Multitarget-Multisensor Tracking: Applications and Advances, Vol.2 (Y. Bar-Shalom Editor). He is member of IEEE and of Eta Kappa Nu, serves as reviewer for different International Journals, taught courses on MS-MTT and Data Fusion at the French ENSTA Engineering School, collaborates for the development of the International Society of Information Fusion (ISIF) since 1998, and has served as Local Arrangements Organizer for Fusion 2000 Conference in Paris. He has been involved in the Technical Program Committees of Fusion 2001-2007 International Conferences. Since 2001, he is a member of the board of the International Society of Information Fusion (http://www.isif.org) and serves in ISIF executive board. He served as executive vice-president of ISIF in 2004. In 2003, he organized with Professor Smarandache, the first special session devoted to plausible and paradoxical reasoning at Fusion 2003, Cairns, Australia and also a panel discussion and a special session on DSmT at Fusion 2004, Fusion 2006. Dr. Dezert gave several invited seminars and lectures on Data Fusion and Tracking during recent past years – the last recent one being Marcus Evans Sensor Fusion Europe, Brussels, Jan 29, 2007. He also participates as member of Technical Committee of last Fuzzy Set and Technology Conferences. He is also Associate Editor of Journal of Advances in Information Fusion (JAIF). Recent advances on DSmT can be found on DSmT web page.

Florentin Smarandache was born in Balcesti, Romania, in 1954. He got a M. Sc. Degree in both Mathematics and Computer Science from the University of Craiova in 1979, received a Ph.D. in Mathematics from the State University of Kishinev in 1997, and continued postdoctoral studies at various American Universities (New Mexico State University in Las Cruces, Los Alamos National Laboratory) after emigration. In 1988 he escaped from his country, pasted two years in a political refugee camp in Turkey, and in 1990 emigrated to USA. In 1996 he became an American citizen. Dr. Smarandache worked as a professor of mathematics for many years in Romania, Morocco, and United States, and between 1990-1995 as a software engineer for Honeywell, Inc., in Phoenix, Arizona. In present, he teaches mathematics at the University of New Mexico, Gallup Campus. Very prolific, he is the author, co-author, and editor of 75 books, over 100 scientific notes and articles, and contributed to about 50 scientific and 100 literary journals from around the world (in mathematics, informatics, physics, philosophy, rebus, literature, and arts). He wrote in Romanian, French, and English. Some of his work was translated into Spanish, German, Portuguese, Italian, Dutch, Arabic, Esperanto, Swedish, Farsi, Arabic, Chinese. He was so attracted by contradictions that, in 1980s, he set up the "Paradoxism" avant-garde movement in literature, philosophy, art, even science, which made many advocates in the world, and it's based on excessive use of antitheses, antinomies, paradoxes in creation - making an interesting connection between mathematics, engineering, philosophy, and literature and led him to coining the neutrosophic logic, a logic generalizing the intuitionistic fuzzy logic that is able to deal with paradoxes. In mathematics there are several entries named Smarandache Functions, Sequences, Constants, and especially Paradoxes in international journals and encyclopedias. He organized the 'First International Conference on Neutrosophics' at the University of New Mexico, 1-3 December 2001. Small contributions he had in physics and psychology too. Much of his work is held in "The Florentin Smarandache Papers" Special Collections at the Arizona State University, Tempe, and Texas State University, Austin (USA), also in the National Archives (Rm. Vâlcea) and Romanian Literary Museum (Bucharest), and in the Musée de Bergerac (France). In 2003, he organized with Dr. Jean Dezert, the first special session devoted to plausible and paradoxical reasoning for information fusion at the Fusion 2003 in Cairns, Australia and has participated to several international workshop and seminar on Information Fusion since 2003. He published two books devoted to DSmT-based Information Fusion in 2004 and 2006 respectively. The complete list of references and past seminars and workshop on DSmT can be found on DSmT web page for convenience.

Return to schedule

Logic-based approaches to information and knowledge fusion

Dr. Éric Grégoire

CRIL CNRS

Université d’Artois

Rue Jean Souvraz SP18

F-62307 Lens Cedex

France

Phone/Fax: (+33) 321791785

Email: gregoire@cril.univ-artois.fr

http://www.cril.univ-artois.fr/~gregoire

Keywords:
Knowledge fusion, beliefs fusion, logic, semantics, knowledge representation and reasoning, knowledge dynamics

Objective of the tutorial and target audience:
The objective of this tutorial is to help Information Fusion 2007 attendees to understand and assess the current state or research about logic-based approaches to information and knowledge fusion. There are no specific prerequisites for this tutorial.

Abstract:
The focus is on qualitative approaches to fuse decision-making knowledge and beliefs. Both syntax-based and semantic-oriented families of techniques are presented. The major concern of handling inconsistency in logic-based knowledge fusion is addressed. Other important issues that are covered concern the rationality analysis of the fusion operators, the fusion of decision rules and the handling of logically weaker but more informative pieces of knowledge.

Biography:
É. Grégoire holds a Master of Science of Engineering in Computing Science (1984) and a Ph.D. (1989) from the University of Louvain in Belgium. He was there awarded the IBM Belgium Prize for the best Ph.D. thesis. He spent one year at the University of Maryland Institute for Advanced Computer Studies (UMIACS) before joining IRISA-INRIA in Rennes in France and holding a permanent senior CNRS position at LORIA in Nancy. Since 1993, É. Grégoire has been a full professor at the Université d’Artois at Lens in France. He is the founding-director of CRIL CNRS, a 50-strong members research laboratory focusing on computer-based knowledge representation and reasoning issues. É. Grégoire has published more than 100 referred papers, many of them about information and knowledge fusion. He has served on various program committees about knowledge fusion. As a member of the editorial board of the Information Fusion Journal, he co-edited in 2006 a special issue on « Logic-based information and knowledge fusion ». É. Grégoire is also Adjunct Scientific Director in charge of the computing science domain at the French Research Ministry.

Return to schedule

Ontology Based Situation Awareness and High Level Fusion: Methods and Tools

Mieczyslaw M. Kokar

Department of Electrical and Computer Engineering,

Northeastern University, Boston, MA

USA

Email: mkokar@ece.neu.edu

http://www.ece.neu.edu/groups/scs/kokar

Intended audience:
This tutorial is directed towards both researchers and practitioners in various areas of information fusion.

Prerequisites:
Basic knowledge of logic. Familiarity with declararive programming is desirable but not required./P>

Description:
In this tutorial the participants will learn:

Higher-level fusion involves estimation of abstract entities - sometimes called “situations” - that can be represented as relations among objects, both physical and conceptual. Unlike features of physical objects, features of relations are not directly measured by sensors. Instead, the existence of a relation is derived from a domain theory relevant to a specific scenario.

This tutorial will cover both theoretical and practical aspects of situation awareness and high-level information fusion. First, a motivational example will be given to demonstrate the importance of relations and to introduce the concept of situation. This will be followed by a presentation of some methodological techniques and some technologies that are needed for establishing an ontological approach to higher level information processing. The notion of ontology will be introduced in theoretical, computational and practical terms. Examples of specific ontologies will be discussed using both a graphical representation and an evolving standard language used for communicating ontologies and annotations as well as for processing and fusion of semantic annotations (OWL – Web Ontology Language). An overview of OWL constructs will be provided using Protégé, the most popular tool for editing ontologies, and some graphical plugins. Other tools will also be demonstrated in the context of a methodology for ontology engineering. Situation awareness and high-level information fusion will be discussed using an illustrative example.

OUTLINE

Hour 1: From Level 1 to High Level fusion. The notions of “situation” and “situation awareness.” A motivational example.

Hour 2: Ontologies and Web Ontology Language (OWL). Ontology engineering and ontology tools.

Hour 3: Situation awareness scenario, demo and analysis. Research directions in situation awareness and higher level information fusion.

Biography:
Professor Kokar is with the Department of Electrical and Computer Engineering at Northeastern University in Boston. His technical research interests include Information Fusion, Ontology-Based Information Processing, Self-Controlling Software and Modeling Languages. In particular, he is interested in higher-level information fusion and situation awareness, ontology-based software radios, the specification and design of self-controlling software using the control theory metaphor, ontology development, ontological annotation of information, logical reasoning about OWL annotated information, consistency checking, formalization of the UML language, consistency checking of UML models vs. UML Metamodel and of UML Metamodel vs. MOF. Dr. Kokar teaches various graduate courses in software engineering, formal methods and artificial intelligence. He has an M.S. and a Ph.D. in computer systems engineering from Wroclaw University of Technology, Poland. He is a senior member of the IEEE and member of the ACM.

More information about Professor Kokar can be found at his web site: http://www.ece.neu.edu/groups/scs/kokar

Return to schedule

Possibilistic Information Fusion

Henri Prade

Institut de Recherche en Informatique de Toulouse (IRIT) - C.N.R.S

Université Paul Sabatier

118 route de Narbonne

31062 Toulouse Cedex 4

France

Email: prade@irit.fr

Abstract:
Possibility theory and the body of aggregation operations from fuzzy set theory provide some tools to address the problem of merging information coming from several sources. Possibility theory is a representation framework that can model various kinds of information items: numbers, intervals, consonant random sets, special kind of probability families, as well as linguistic information, and uncertain formulae in logical settings. The possibilistic approach to fusion is general enough to encompass logical modes of combination (conjunctive and disjunctive) as well as fusion modes used in statistics. The choice of a fusion mode depends on assumptions on whether all sources are reliable or not, and can be based on conflict analysis. A possibilistic logic counterpart of the combination modes applicable to possibility distributions has been developed. The approach applies to sensor fusion, aggregation of expert opinions as well as the merging of databases especially in case of poor, qualitative information. This general framework allows us to import inconsistency handling methods, inherited from logic, into numerical fusion problems. Quantified, prioritized and weighted fusion rules are described, as well as fusion under a priori knowledge. It has been shown that the possibilistic logic setting is compatible with the Bayesian approach to fusion, the main difference being the presupposed existence, or not, of prior knowledge. Relations and comparisons with combination rules in other frameworks such as Shafer theory of evidence will be discussed. Finally, fusion will be also presented in bipolar possibility theory, a framework that distinguishes between positive and negative information. The approach applies to sensor fusion, aggregation of expert opinions as well as the merging of databases especially in case of poor, qualitative information. The approach proposed and the results presented have been obtained in collaboration with Didier Dubois.

Biography:
Henri Prade was born in Mulhouse in 1953. He is "Directeur de Recherche" at C.N.R.S., and works at IRIT (Institut de Recherches en Informatique de Toulouse). He received a Doctor-Engineer degree from Ecole Nationale Supérieure de l'Aéronautique et de l'Espace, in Toulouse (1977), his "Doctorat d'Etat" (1982) and the "Habilitation à Diriger des Recherches" (1986) both from Paul Sabatier University in Toulouse. He is the co-author, with Didier Dubois, of two monographs on fuzzy sets and possibility theory published by Academic Press (1980) and Plenum Press (1988) respectively. He has contributed a great number of technical papers and has edited several books including the seven volumes of the "Handbooks of Fuzzy Sets Series" (Kluwer, 1998-2000). He is a member of the editorial board organization of several technical journals including Fuzzy Sets and Systems, IEEE Transactions on Fuzzy Systems, the Artificial Intelligence J., ACM Trans. on Computational Logic, the Inter. J. of Approximate Reasoning, the Inter. J. of Intelligent Systems, the J. of Intelligent Information Systems, Fundamenta Informaticae, and Information Sciences, among others. His current research interests are in uncertainty and preference modeling, non-classical logics, approximate and plausible reasoning with applications to artificial intelligence and information systems.

Return to schedule


What is risk? What is probability?

Glenn Shafer

Department of Accounting and Information Systems

Rutgers Business School–Newark and New Brunswick

180 University Ave, Newark, New Jersey 07102, USA

Phone 973-353-1604.

Fax 973-353-1283

Secretary: Jackie Adams 973-353-1644

Email: gshafer@andromeda.rutgers.edu

Abstract:
For 170 years, people have been arguing about whether probability is objective or subjective. From the game-theoretic point of view that emerges from my 2001 book with Vladimir Vovk (Probability and Finance: It’s Only a Game, Wiley), the question is instead whether a particular decision problem is embedded in a repetitive structure. The method of defensive forecasting, developed by Vovk in more recent work, gives probabilities for decision making in problems that are sufficiently repetitive, without any a priori assumption about trials being identical. But when the focus is on the particular case rather than on long-run average performance, and when there is even contention about what long-run sequence the particular case should be compared with, we need instead methods of weighing evidence.

Biography:
Glenn Shafer spent his childhood on a farm near Caney, Kansas. He earned two degrees from Princeton University: an A.B. in mathematics in 1968 and a Ph.D. in mathematical statistics in 1973. After teaching in the Department of Statistics at Princeton, he returned to Kansas in 1976 to teach in the Department of Mathematics at the University of Kansas. In 1984, he moved from Mathematics to Business at Kansas. Since 1992, he has taught in the Rutgers Business School–Newark and New Brunswick.
In 1976 Glenn published A Mathematical Theory of Evidence, which formulated the Dempster-Shafer theory of belief functions, now widely used for handling uncertainty in expert systems. In 1996, he published a second major book, The Art of Causal Conjecture , which is concerned both with expert systems based on causal models and with the empirical investigation of causality.
His most important work is Probability and Finance: It's Only a Game! This book, co-authored by Volodya Vovk and published in 2001, shows how game theory can replace measure theory as a foundation for mathematical probability. The new foundation is elegant and powerful, and Glenn believes that it will eventually reshape the use of probability in many different fields. It has immediate applications in finance, where it points the way to pricing options and evaluating the performance of investments without assuming that markets behave randomly. It strengthens Glenn's earlier analysis of causality and provides a new way of understanding the distinction between objective or causal probability on the one hand and subjective or personal probability on the other.
Glenn has published in journals in statistics, philosophy, history, psychology, computer science, economics, engineering, accounting, and law. He has won teaching awards in both mathematics and business. He was a Guggenheim fellow in 1983-84, a fellow at the Center for Advanced Study in the Behavioral Sciences in 1988-89, and a Fulbright fellow at the Free University of Berlin in Spring 2001. He is a fellow of both the Institute of Mathematical Statistics and the American Association for Artificial Intelligence.

Return to schedule

Introduction to social network analysis and complex networks

Pontus Svenson

Swedish Defence Research Agency

Sweden

Email: ponsve@foi.se

Keywords:
Terrorist network analysis, social network analysis, complex networks.

Objective:
After the tutorial session, the attendees should be knowledgeable about network models and social network analysis and be able to use algorithms and methods from this research area in their own fusion research.

Abstract:
Social network analysis (SNA) is a key methodology for analyzing the dark networks of terrorists and opponents that face the armed forces today. Modeling the opponents as a network and combining information fusion methods and SNA is expected to improve the situation awareness of the users.
The subject has a long and rich history. This tutorial session will give a description of the historical roots and development of social network analysis and provide information about different ways of visualizing networks as well as how a network can be characterized mathematically using various measures. Exact as well as approximate algorithms for social network analysis will be described, and a survey of the state of the art in the field given. Mathematical models for graphs and networks will be described and used. Connections between SNA and data mining and the semantic web will be mentioned, and some of the dangers and pitfalls when using SNA will be described.
Exercises and demonstrations using open-source software and publicly available data such as the Enron dataset will be included.

Tentative outline of course:

(due to time constraints, not all of these subjects will be covered in detail)

Biography:
Dr Pontus Svenson is a senior scientist and information fusion research manager at the Swedish Defence Research Agency (FOI). His research interests include network analysis, data and text mining, random sets, situation and threat assessment, and information operations. He currently leads research that aims at providing situation and threat assessment tools to the future Swedish battle groups, and is involved in an EU collaborative project that applies social network analysis to anti-terrorism. He has a PhD in theoretical physics from Chalmers University of Technology. He was introduced to the subject of complex networks and social network analysis while doing his PhD-research, parts of which were spent studying spin models on complex networks. He lectures regularly on information fusion and decision support systems at the Swedish National Defence College and has been guest-lecturer for the past two years in the information fusion course given at University of Skövde. At the 2006 ELSNET summer school on Information fusion for natural language processing, he gave a course on Information fusion which included some lectures about social network analysis.
He is the author or co-author of more than 20 papers and unclassified technical reports on information fusion.

Selected publications:

Svenson, P., Svensson, P. and Tullberg, H.
Social Network Analysis And Information Fusion For Anti-Terrorism
In Proceedings of the Conference on Civil and Military Readiness 2006 (CIMI 2006), Enköping, Sweden, 16-18 May 2006, Paper S3.1. Försvarets Materielverk, Stockholm, 2006.

Svenson, P.
Complex networks and social network analysis in information fusion
In Proceedings of the Ninth International Conference on Information Fusion (FUSION 2006), Florence, Italy, 10-13 July 2006. IEEE, Piscataway, NJ, 2006, Paper 80, pp. 1-7.

Ahlberg, S., Hörling, P., Johansson, K., Jöred, K., Kjellström, H., Mårtenson, C., Neider, G., Schubert, J., Svenson, P., Svensson, P. and Walter, J.
An information fusion demonstrator for tactical intelligence processing in network-based defense
Information Fusion 8(1) (2007) 84-107.


Return to schedule

Distributed Robotic Sensor Systems

Katia Sycara

Carnegie Mellon University

USA

Email: katia@cs.cmu.edu


Bin Yu

Quantum Leap Innovations, Inc.

USA

Email: byu@quantumleap.us

Keywords:
Distributed sensor systems, data fusion, task allocation, path planning.

Target audience:
This tutorial is intended for any delegates who have some basic knowledge of AI and algorithms. Familiarity with basic concepts of sensor systems is desirable but not essential.

The topic of sensor systems is of general interest for Fusion 2007 attendees. A similar tutorial at Fusion 2006 was deemed to be highly successful, attracting one of the largest number of attendees in any tutorial at Fusion 2006.

Abstract:
Robotic sensor systems of the near future are envisioned to consist of hundreds of unmanned vehicles such as UAVs and UGVs. These networked autonomous and geographically distributed sensors play strong roles in military and civilian operations, e.g., battlefield surveillance and disaster rescue. At the same time, sensor systems offer many exciting research challenges due to their real-world constraints such as imperfect sensor data, real-time execution, and scarce wireless communication bandwidth. Many algorithms have been developed in the context of sensor systems, however, most of them are centralized and not scalable due to their optimistic assumption of unlimited communication bandwidth.

In this tutorial we will survey the state-of-the-art of distributed algorithms for collaborative sensor systems, where sensors need to be coordinated in an effective way. We will summarize major advances of distributed coordination theories and algorithms in the areas of AI and multiagent systems. Specifically, we will focus on the design and analysis of distributed coordination algorithms for data fusion, task allocation, and cooperative path planning in a large-scale autonomous sensor system, where large numbers of autonomous mobile sensors may enter and leave the system dynamically. We will conclude by discussing some open research issues in distributed coordination algorithms and their relation to other fields, such as control theory and operation research (OR).

OUTLINE

1. Overview
1.1. Autonomous sensors
1.2. Current coordination techniques
1.3. Why distributed coordination

2. Distributed coordination techniques
2.1. Coordinated data fusion
2.2. Task allocation and execution
2.3. Cooperative path planning

3. Conclusion


Biography:
Dr. Katia Sycara is a Professor in the School of Computer Science at Carnegie Mellon University. She holds BS in Applied Math from Brown University, MS Electrical Engineering from University of Wisconsin, and a Ph.D. in Computer Science from Georgia Institute of Technology. Her research work lies in the intersection of Operations Research, Artificial Intelligence and Software Engineering. Dr. Sycara has authored over 300 technical papers and book chapters in multiagent/multirobot systems, negotiation, auctions, agent teams, and human-agent interaction. Prof. Sycara is a Fellow of the AAAI, Fellow of the IEEE and the recipient of the 2002 ACM/SIGART Agents Research Award. She is a founding Editor-in-Chief of the International Journal of Autonomous Agents and Multi-Agent Systems and on the editorial board of 5 additional journals. She has served as the Program Chair for the Second International Semantic Web Conference (ISWC-03) and the General Chair of the Second International Conference on Autonomous Agents (Agents-98).

Dr. Bin Yu is a Senior Research Scientist at Quantum Leap Innovations. Before joining Quantum Leap, Dr. Yu was a Postdoctoral Fellow at the School of Computer Science at Carnegie Mellon University (CMU), Pittsburgh. He received his Ph.D. in Computer Science from North Caroli