Master cutting-edge Level 2 fusion techniques that help you develop powerful situation assessment services with eye-popping capabilities and performance with this trail-blazing resource. The book explores object and situation fusion processes with an appropriate handling of uncertainties, and applies cutting-edge artificial intelligence and emerging technologies like particle filtering, spatiotemporal clustering, net-centricity, agent formalism, and distributed fusion together with essential Level 1 techniques and Level 1/2 interactions. Moreover, it includes all the tools you need to design high-level fusion services, select algorithms and software, simulate performance, and evaluate systems with never-before effectiveness.
The book explains the Bayesian, fuzzy, and belief function formalisms of data fusion and a review of Level 1 techniques, including essential target tracking methods. Further, it covers Level 2 fusion methods for applications such as target classification and identification, unit aggregation and ambush detection, threat assessment, and relationships among entities and events, and assessing their suitability and capabilities in each case. The book’s detailed discussion of Level 1/2 interactions emphasizes particle filtering techniques as unifying methods for both filtering under Level 1 fusion and inferencing in models for Level 2 fusion. The book also describes various temporal modeling techniques including dynamic Bayesian networks and hidden Markov models, distributed fusion for emerging network centric warfare environments, and the adaptation of fusion processes via machine learning techniques. Packed with real-world examples at every step, this peerless volume serves as an invaluable reference for your research and development of next-generation data fusion tools and services.

Contents of the book :

Background and Concepts—JDL Architecture and Situation Assessment. Increasing Level of Abstraction of Knowledge. Assessment vs. Awareness. Example Application Domains. Sensors and Data Sources. OODA and Other Architectures. Agent-Based Situation Assessment.
Approaches to Handling Uncertainty—Classification of Uncertainties. Probability Theory: Bayesian Probability. Mathematical Logics: Modal Logics. Neo-Probabilists: Bayesian Belief Networks. Neo-Logicist: Fuzzy Logic. Neo-Calculist: Dempster-Shafer, Certainty Factor. Handling of Noisy and Unstructured Text Data
Target Tracking—Single-Sensor Single-Target Tracking. Multi-Sensor Single-Target Tracking (in Clutter). Multi-Sensor Multi-Target Tracking (in Clutter). Interacting Multiple Models.
Target Classification and Identification—Rule-Based Approaches. Expert Systems: Bayesian and Certainty Factor Formalisms. Symbolic Argumentation: D-S Theory of Belief Function. Learning Rules for Classification. Example Applications: AOC Time-Sensitive-Targets.
Unit Aggregation—Spatiotemporal Clustering Concept. Directivity and Displacement-Based Unconstrained Clustering. Singular Value Decomposition-Based Clustering. Preprocessing Through Entropy Measure. Example Applications: Ambush Detection.
Model-Based Situation Abstraction—Graphical Models for Situation Assessment. Bayesian Belief Network Technology. Algorithms for Inferencing in Belief Network Models. Handling of Continuous Variables. Adaptation of Belief Network Models. Example Applications.
Modeling of Time for Situation Assessment—State Space Models. Hidden Markov Models. Dynamic Bayesian Networks. Particle Filtering (Monte Carlo and Rao-Blackwellized). Example Applications.
Performance Enhancement and Evaluation— Level 2 Feedback for Enhanced Level 1 Fusion. Subjective Evaluation. Quantitative Assessment via ROC. Cramer-Rao Lower Bound. Example Applications.
Decision Support—Actions and Utility. Rule-based Decision Support. Expected Utility Theory and Decision Tree. Influence Diagrams. Example Applications
Distributed Situation Assessment—Network Centric Warfare. Distribution of Models. Formalism Incompatibilities. Handling of Information Pedigree. Example Applications.
Learning of Fusion Models—Rule Learning. Belief Network Learning. Sequence Learning. Hidden Markov Models.
References