Steganographic sonar

Author:

Park, Joonho Daniel

Graduate Program:

Electrical Engineering

Degree:

Doctor of Philosophy

Document Type:

Dissertation

Date of Defense:

October 13, 2016

Committee Members:

John F. Doherty, Dissertation Advisor/Co-Advisor
John F. Doherty, Committee Chair/Co-Chair
Vishal Monga, Committee Member
David M. Jenkins, Committee Member
Lee Culver, Outside Member

Keywords:

sonar
signal processing
tracking
steganography
information theory
detection
estimation
motion modeling
sound modeling
waveform design

Abstract:

Steganographic transmission waveforms are desirable for many underwater applications such as underwater acoustic communications, surveillance, detection, and tracking. In the context of active sonar tracking, an overall approach is presented that covers many topics including background sound modeling, steganographic security assessment of transmission waveform, target motion modeling, and batch track detection. Under shallow water environment background sound dominated by snapping shrimps, the presented approach shows feasibility of steganographic sonar for target track detection. A background sound modeling approach is presented that utilizes symbolic time series analysis with phase space representation and clustering of time series segments. The time evolving characteristics are captured with the hidden Markov model (HMM). The steganographic security of the transmission waveform can be assessed by measuring the Kullback-Leibler divergence (KLD) between the cover model and the stego model. Conventional tracking approaches fail to perform reliably, and existing track-before-detect approaches also fail due to overwhelming search computation when they use simplistic short timescale motion models. A long timescale target motion model with considerations for plausibility is presented and utilized in a batch track detection approach. A Monte Carlo-based simulation was performed to find the scenarios that demonstrate feasibility of steganographic sonar in a shallow water environment with snapping shrimp sound.