Integrated perception, modeling, and control paradigm for bistatic sonar tracking by autonomous underwater vehicles

Abstract

In this thesis, a fully autonomous and persistent bistatic anti-submarine warfare (ASW) surveillance solution is developed using the autonomous underwater vehicles (AUVs). The passive receivers are carried by these AUVs, and are physically separated from the cooperative active sources. These sources are assumed to be transmitting both the frequency-modulated (FM) and continuous wave (CW) sonar pulse signals. The thesis then focuses on providing novel methods for the AUVs/receivers to enhance the bistatic sonar tracking performance. Firstly, the surveillance procedure, called the Automated Perception, is developed to automatically abstract the sensed acoustical data from the passive receiver to the track report that represents the situation awareness. The procedure is executed sequentially by two algorithms: (i) the Sonar Signal Processing algorithm - built with a new dual-waveform fusion of the FM and CW signals to achieve reliable stream of contacts for improved tracking; and (ii) the Target Tracking algorithm - implemented by exploiting information and environmental adaptations to optimize tracking performance. Next, a vehicular control strategy, called the Perception-Driven Control, is devised to move the AUV in reaction to the track report provided by the Automated Perception. The thesis develops a new non-myopic and adaptive control for the vehicle. This is achieved by exploiting the predictive information and environmental rewards to optimize the future tracking performance. The formulation eventually leads to a new information-theoretic and environmental-based control. The main challenge of the surveillance solution then rests upon formulating a model that allows tracking performance to be enhanced via adaptive processing in the Automated Perception, and adaptive mobility by the Perception-Driven Control. A Unified Model is formulated in this thesis that amalgamates two models: (i) the Information-Theoretic Model - developed to define the manner at which the FM and CW acoustical, the navigational, and the environmental measurement uncertainties are propagated to the bistatic measurement uncertainties in the contacts; and (ii) the Environmental-Acoustic Model - built to predict the signal-to-noise power ratios (SNRs) of the FM and CW contacts. Explicit relationships are derived in this thesis using information theory to amalgamate these two models. Finally, an Integrated System is developed onboard each AUV that brings together all the above technologies to enhance the bistatic sonar tracking performance. The system is formulated as a closed-loop control system. This formulation provides a new Integrated Perception, Modeling, and Control Paradigm for an autonomous bistatic ASW surveillance solution using AUVs. The system is validated using the simulated data, and the real data collected from the Generic Littoral Interoperable Network Technology (GLINT) 2009 and 2010 experiments. The experiments were conducted jointly with the NATO Undersea Research Centre (NURC).