Development of High Level Decision Making and Decision Fusion for Layered Sensing Systems
Abstract
The goal of dissertation was to develop and integrate intelligent self-reliant software system to eliminate or minimize the involvement of human operators in existing multi-layer sensing surveillance systems. In order to achieve this, a cross layered-sensing software system for decision-making and decision-fusion was developed, implemented, verified and validated. The system was designed and developed following a top-down systems engineering approach. The system of interest contains three subsystems: the decision-making (DM) subsystem, the decision-fusion (DF) subsystem, and the request handler and control subsystem (RHC). The DM subsystem was developed for each sensing layer based on grey relational analysis (GRA) method. The DF was developed to combine decisions across layers to provide a more accurate actionable system decision. DF was developed using grey mixed center point triangular whitening functions. The DF subsystem takes in decisions made by the DM subsystem. The RHC contains two components; control component and the request handler component. The control component was developed using logical control and SDK commands which accompanied the sensors. The request handler was developed using TCP-IP communication protocol. The different layers of the layered sensing system evaluate information about a given object in terms of its behavior and make a decision. The decision-fusion subsystem fuses decisions across layers. The integrated decision-making and fusion software system was tested on real-time scenarios obtained by a test-bed consisted of three layers. The average layers decision-making performance is 81.8% correct decision-making. The overall system decision-fusion, after the decision enhancement, outperformed the three layers' average performance rate of 92.8% correct decision-making. The obtained results from DM, RHC, and DF verified the system requirements were met. The subsystems development, implementation, testing and the system integration of the Decision-Making, Decision-Fusion, and Request Handler and Control subsystems are all presented in this dissertation report.
Subject Area
Electrical engineering|Computer Engineering
Recommended Citation
Bakhita Salman,
"Development of High Level Decision Making and Decision Fusion for Layered Sensing Systems"
(2019).
ETD Collection for Tennessee State University.
Paper AAI22587506.
https://digitalscholarship.tnstate.edu/dissertations/AAI22587506