ppt课件-kickoff meeting presentation- migma systems, inc（开球会议演示-混合岩浆系统公司）.pptVIP

Presentation Outline Problem Statement Problem in Motion Based Detection Basic System Architecture Random Field for Undersea Sensor Network Random Field Estimation Our Approach for Random Field Estimation FBMM Technology (Patented) A Simulated Network Difficult to Find Targets Visually Random Field Estimation Mathematical Morphological Operation Backward Mapping STLD Technology (Patented) Gap Statistics Based Clustering Individual Clusters in Temporal Patterns Temporal Spatial Mitigation False Reports Reduction Using STLD Combine FBMM STLD Target Synthetic Tracks A Simulation Tool Conclusion * * A Sparse Undersea Sensor Network Decision Support System Based on Spatial and Temporal Random Field April 10, 2007 Defense and Security Symposium 2007 Dr. Bo Ling Migma Systems, Inc. Dr. Mike Traweek Office of Naval Research Dr. Tom Wettergren Naval Undersea Warfare Center - Problem Statement - Forward Backward Mapping Mitigation - Spatial Temporal Layering Discrimination - A Simulation Tool for Target Detection and Tracking - Random Field Estimation of Undersea Sensor Network - Conclusion A Large Surveillance Region In a passive submarine detection system, we need to consider: - Number of sensors required in the sparse sensor network - Probability of submarine detection when a submarine might be detected by only one sensor at one particular time instance - False alarm mitigation No Overlapping In a sparse sensor network, targets can be detected and tracked over a period of time. As a target moves, although it may not be detected by any sensors at certain time instance, it will be detected by a number of sensors over a period of time. When positive and negative reports co-exist, it is difficult to determine if targets actually exist! Although targets in a surveillance region may not always be detected, as they move, sensors can collectively detect, classify and track them. A typical random field model can be described by the following difference equation: In random