Networked ATR Systems Design Considerations: System Performance and Resource Constraints

Transcription

1 Networked ATR Systems Design Considerations: System Performance and Resource Constraints Joseph A. O Sullivan, O WU ESSRL Michael D. DeVore,, UVA Alan Van Nevel,, NAWC

2 Outline Introduction: Principled ATR System Design Navy/Marine Corps Operational Issue System Performance System Resource Tradeoffs Successively Refinable Models Optimal Performance-Time Tradeoff Rate-Recognition Recognition Theory Computation Communication Communication Tradeoffs 2

3 Who are we and why are we here? Problem-Driven Research Draw results from relevant fields Experimental/computational results motivate modifications and contributions to those fields Problem: automatic target recognition Relevant Fields: information theory, signal processing, imaging, data modeling, computational modeling 3

4 Contributing Fields Information theory Rate-distortion theory data compression Channel coding theory communication systems Large deviations error rates Fisher information estimation error bounds Successive refinement Signal processing Detection, estimation, multiple hypothesis testing Multiresolution analysis Imaging (information-theoretic theoretic imaging) System design and analysis Image processing, computer vision Data modeling Underlying physical phenomena Sensor statistics Computational modeling Implementation design, constraints, modeling 4

5 Foundations ONR MURI on multisensor fusion ARO Center for Imaging Science DARPA Sponsorship Collaborators: D. L. Snyder, M. I. Miller, U. Grenander,, P. Moulin, R. E. Blahut, A. Srivastava,, A. Lanterman,, D. R. Fuhrmann,, S. Soatto,, R. Pless, R. S. Indeck,, J. Shapiro, T. Green, B. Rimoldi CIS: D. Mumford,, A. Yuille,, P. Belhumer,, J. Aggarwal,, H. Krim, A. Willsky,, S. Mitter Other ONR-funded research NAWC, system of systems, etc. I. Daubechies,, R. DeVore,, D. Donoho Colleagues D. Dudgeon, L. Novak, F. Robey,, F. Garber, E. Zelnio,, M. Bryant Literature 5

6 CNO SSG XX FORCENet G ENERARE U U F T RUM Vision Multi-Tiered Sensor Field Scalable - space to sea bed National/Theater/Organic assets The Warrior Networks Fully-netted force Processing and collaboration Real-time shared knowledge Enhanced Capabilities Shared battlefield awareness Speed of decision making Effects-based operations Dispersed force protection Mission flexibility Global reach back Sustainability Theater Ballistic Missile Defense Undersea Warfare Naval Fires Theater Air Defense Sustainment The architecture for a fully netted maritime force Command & Control

7 Traditional Approaches to ATR Platform-centric, stovepiped Subsystems (sensor, communications, data processing) have independent design and operation Difficult to prove overall system is operating optimally Often brittle point solutions Fixed System architecture Sensors, computational hardware, communications, geometry, signal processing blocks (compression, etc.) Time constraints Target model set 7

8 Need Navy Need/Impact Robust networked processing for automatic target recognition (ATR) Ability to leverage real-time sharing of data as envisioned in FORCEnet» Rapidly evolving system architecture» Evolving time constraints» Evolving set of targets Impact Successively refinable ATR: near optimal performance subject to dynamic time constraints ATR System performance subject to dynamic resource constraints dynamically reconfigurable ATR Development of system design guidelines based on known processor/timeline/bandwidth limitations Extract information content from imagery, maximizing bandwidth usage 8

9 ATR Problem Scope Scene (source) Imaging Sensor Platform Imaging Sensor Platform Interconnection Networks ATR Processor Inferred Target and Pose Imaging Sensor Platform Model Database Multiple imaging sensors: LADAR, IR, EO, Multispectral, Hyperspectral,, SAR Network of sensors, processors, databases Constraints imposed by system architecture, time constraints, and target model set Successive refinability to dynamically tradeoff accuracy and resources (time, bandwidth, computations, etc.) 9

10 ATR Performance and Complexity Comparison in terms of: Performance achievable at a given complexity Complexity required to achieve a given performance Information Theory Basis: Rate-Distortion Theory Rate-Recognition Theory Theoretical curves Operational curves Actual performance Model errors Varying clutter, background Nonideal implementations 10

11 Conclusions Navy/Marine Corps Impact Successively refinable ATR: near optimal performance subject to dynamic time constraints ATR System performance subject to dynamic resource constraints dynamically reconfigurable ATR Development of system design guidelines based on known processor/timeline/bandwidth limitations Extract information content from imagery, maximize bandwidth usage 11