Search

This chapter focuses on the foundation of the system identification problem for state‐space systems that leads to subspace identification techniques. The original basis has evolved from systems theory and the work of Kalman for control system design. The fundamental problem is called the “realization problem". The basic paper by Ho and Kalman has been cited as the seminal publication showing how to extract the state‐space model. The chapter discusses this problem for two distinct data sets: impulse sequences of an infinite length (number of samples) data leading to the realization problem and input/output data leading to the subspace identification problem. Subspace identification techniques enable the capability to extract a large number of parameters compared to the prediction error methods that are quite time‐consuming and not practical for large parametric problems coupled with the need for potential real‐time applications.

This chapter discusses the stochastic realization problem from both the classical and the subspace perspectives. It starts with the classical problem that mimics the deterministic approach with covariance matrices replacing impulse response matrices. The underlying Hankel matrix, now populated with covariance rather than impulse response matrices, admits a factorization leading to the fundamental rank condition. The subspace approach also follows in a development similar to the deterministic case. Starting with the multivariable output error state‐space formulation using orthogonal projection theory, both the “past input multivariable output error state‐space” and “past input/output multivariable output error state‐space” techniques evolve. Next, the numerical algorithms for state‐space system identification approach follow based on oblique projection theory. The chapter concentrates primarily on a solution to the “combined problem” that is the identification of both the deterministic and stochastic systems directly (without separation) through the estimated state vector embedded in the innovation representation of the system.

In this chapter, we discuss the representation of a random signal, first as a stochastic process with accompanying statistics and then as a discrete random signal with certain inherent properties. We show how random signals can be characterized by their spectral representations and then consider systems excited by random inputs that employ these properties. With these concepts developed, we then show how processes can be modeled simply by a parametric model of the ARMAX class that is used quite heavily in modern spectral techniques. Finally, we discuss a set of spectral estimation algorithms to analyze the spectral content of random signals and illustrate their performance in a case study of bandpass sinusoids in noise.

In this chapter, we develop a suite of case studies applying model‐based identification (MBID) techniques to extract models for processing using both subspace and parametrically adaptive schemes. We start with a complex mechanical (structural) system applying subspace techniques followed by the development of a physics‐based model for Kalman filtering in a scintillation system applying both identification approaches. Two MBID schemes using Bayesian particle filters (PF) are developed from the underlying phenomenology in order to identify/detect fission processes as well as modal functions propagating in a shallow ocean environment. Finally, we extract (estimate) chirp and frequency‐shift key (FSK) signals from noisy data using the parametrically adaptive, unscented Kalman filter (UKF) approach.

When uncertain acoustic processes can no longer be characterized by Gaussian or for that matter unimodal (single peak) distributions along with the fact that the underlying phenomenology is nonstationary (time-varying) and nonlinear, then more general Bayesian processors must be applied to solve the underlying signal enhancement/extraction problem. A particle filter provides a solution to this multimodal (multiple peaks) posterior distribution estimation problem in noisy acoustic environments. A particle filter is a sequential Markov chain Monte Carlo processor capable of providing reasonable performance for data evolving from a multimodal distribution by estimating a nonparametric representation of the posterior distribution from which a multitude of meaningful statistics can be retrieved. However, the question of evaluating its performance can be challenging even for the simplest of processes. For instance, it is well-known that Kalman filter optimality can be obtained only when the resulting error residuals (innovations) are zero-mean and white. It is not that simple for the particle filter. Once characterized, the performance of the particle filter must be analyzed for it to be of practical value. Here a set of design and analysis criteria is discussed and applied to demonstrate their ability to quantify particle filtering performance.

When a broadband source propagates sound in a shallow ocean the received data can become quite complicated due to temperature-related sound-speed variations and therefore a highly dispersive environment. Noise and uncertainties disrupt this already chaotic environment even further because disturbances propagate through the same inherent acoustic channel. The broadband (signal) estimation/detection problem can be decomposed into a set of narrowband solutions that are processed separately and then combined to achieve more enhancement of signal levels than that available from a single frequency, thereby allowing more information to be extracted leading to a more reliable source detection. A Bayesian solution to the broadband modal function tracking, pressure-field enhancement, and source detection problem is developed that leads to nonparametric estimates of desired posterior distributions enabling the estimation of useful statistics and an improved processor/detector. To investigate the processor capabilities, we synthesize an ensemble of noisy, broadband, shallow-ocean measurements to evaluate its overall performance using an information theoretical metric for the preprocessor and the receiver operating characteristic curve for the detector.

The Silver Medal is presented to individuals, without age limitation, for contributions to the advancement of science, engineering, or human welfare through the application of acoustic principles, or through research accomplishment in acoustics.

Mechanical devices operating in noisy environments lead to low signal-to-noise ratios creating a challenging signal processing problem to monitor the vibrational signature of the device in real-time. To detect/classify a particular type of device from noisy vibration data, it is necessary to identify signatures that make it unique. Resonant (modal) frequencies emitted offer a signature characterizing its operation. The monitoring of structural modes to determine the condition of a device under investigation is essential, especially if it is a critical entity of an operational system. The development of a model-based scheme capable of the on-line tracking of structural modal frequencies by applying both system identification methods to extract a modal model and state estimation methods to track their evolution is discussed along with the development of an on-line monitor capable of detecting anomalies in real-time. An application of this approach to an unknown structural device is discussed illustrating the approach and evaluating its performance.

Dr. Edmund J. Sullivan developed synthetic aperture towed-array processing. Confronted by many nay-sayers Dr. Sullivan persisted with his work and developed the first approach as the signal processing group leader (SPGL) at the SACLANT ASW Center (now CMRE) in La Spezia, Italy. The first notional passive synthetic aperture processor using an overlapped correlation method was jointly developed [“Extended Towed Array Processing by Overlapped Correlator,” (JASA, 1989)]. This idea was to blossom even further in his collaborative works (Stergiopolous et al.) extending these ideas to a fully, recursive (in-time) model-based passive-synthetic aperture processor [“Space-time array processing: A model-based approach” (JASA, 1997)]. He also began collaborations with the Swedish Navy where he performed joint experiments in model-based passive synthetic aperture evaluating its performance in the ocean and demonstrating its effectiveness. Dr. Sullivan began mentoring more researchers and teaching at the University of Rhode Island where he advised students in underwater acoustics and processing leading researchers to the model-based approach (Cousins), synthetic aperture processing (Edelson) [“On the performance of the overlap-correlator synthetic aperture technique” (JASA, 1991)] and broadband processing (Holmes) [“Broadband passive synthetic aperture” (JASA, 2006) ]. His ideas were summarized in his recent text, [Model-Based Processing for Underwater Acoustic Arrays (Springer, 2015)].Dr. Edmund J. Sullivan developed synthetic aperture towed-array processing. Confronted by many nay-sayers Dr. Sullivan persisted with his work and developed the first approach as the signal processing group leader (SPGL) at the SACLANT ASW Center (now CMRE) in La Spezia, Italy. The first notional passive synthetic aperture processor using an overlapped correlation method was jointly developed [“Extended Towed Array Processing by Overlapped Correlator,” (JASA, 1989)]. This idea was to blossom even further in his collaborative works (Stergiopolous et al.) extending these ideas to a fully, recursive (in-time) model-based passive-synthetic aperture processor [“Space-time array processing: A model-based approach” (JASA, 1997)]. He also began collaborations with the Swedish Navy where he performed joint experiments in model-based passive synthetic aperture evaluating its performance in the ocean and demonstrating its effectiveness. Dr. Sullivan began mentoring more researchers and teaching at the University of ...

This chapter introduces the concept of discrete hidden Markov model (HMM) and illustrates their internal characteristics through a state‐space representation. It develops the concepts of Markov and hidden Markov chains and shows how they were related. Next, the chapter investigates properties of the HMM illustrating how the Bayesian concepts easily transfer over to this discrete representation. It investigates the three fundamental problems along with some variations: (1) the evaluation (simulation) problem; (2) the state estimation problem; and (3) the parameter estimation problem. A careful analysis of each led us to the popular Viterbi decoding technique and the specialized expectation‐maximization (EM) algorithm popularly called the Baum‐Welch technique. The chapter concludes with a case study to decode a transmitted coded sequence from data enhanced by a time‐reversal (T/R) processor and also considers applying the Viterbi algorithm to decode a message transmitted through a hostile environment with reverberations along with the processor and decoding algorithm.

This chapter develops sequential Bayesian detection techniques primarily aimed at the binary decision problem. It enables the extension of these estimation methods to an entire class of problems especially when a physical model is available that can be incorporated into the algorithm leading to model‐based detection. The chapter develops the Bayesian approach to decision theory primarily aimed at a coupling of sequential Bayesian estimation to sequential decision‐making. It discusses the basic theory required to comprehend the receiver operating characteristic (ROC) and its inherent features that enable us to ascertain the performance of detection algorithms. The chapter develops the theory and from it develops a variety of metrics that can be extracted from the ROC to increase understanding of the embedded information. It applies the analysis to examples and uses the ROC to compare performance of a variety of detection techniques. Finally, a case study followed showing the applicability of the model‐based designs to anomaly detection.

The detection of radioactive contraband is a critical problem in maintaining national security for any country. Gamma-ray emissions from threat materials challenge both detection and measurement technologies significantly. The development of a sequential, model-based Bayesian processor that captures both the underlying transport physics of gamma-ray emissions including Compton scattering and the measurement of photon energies offers a physics-based approach to attack this challenging problem. The inclusion of a basic radionuclide representation of absorbed/scattered photons at a given energy along with interarrival times is used to extract the physics information available from noisy measurements. It is shown that this representation leads to an “extended” physics-based structure that can be used to develop an effective sequential detection technique. The resulting model-based processor is applied to data obtained from a controlled experiment in order to assess its feasibility.

The timely and accurate detection of nuclear contraband is an extremely important problem of national security. The development of a prototype sequential Bayesian processor that incorporates the underlying physics of ?-ray emissions and the measurement of photon energies and their interarrival times that offers a physics-based approach to attack this challenging problem is described. A basic radionuclide representation in terms of its ?-ray energies along with photon interarrival times is used to extract the physics information available from the uncertain measurements. It is shown that not only does this approach lead to a physics-based structure that can be used to develop an effective threat detection technique, but also motivates the implementation of this approach using advanced sequential Monte Carlo processors or particle filters to extract the required information. The resulting processor is applied to experimental data to demonstrate its feasibility.

In the context of the recent Maritime Rapid Environmental Assessment/Blue Planet 2007 sea trial (MREA/BP07), this paper presents a range-resolving tomography method based on ensemble Kalman filtering of full-field acoustic measurements, dedicated to the monitoring of environmental parameters in coastal waters. The inverse problem is formulated in a state-space form wherein the time-varying sound-speed field (SSF) is assumed to follow a random walk with known statistics and the acoustic measurements are a nonlinear function of the SSF and the bottom properties. The state-space form enables a straightforward implementation of a nonlinear Kalman filter, leading to a data assimilation problem. Surface measurements augment the measurement vector to constrain the range-dependent structure of the SSF. Realistic scenarios of vertical slice shallow-water tomography experiments are simulated with an oceanic model, based on the MREA/BP07 experiment. Prior geoacoustic inversion on the same location gives the bottom acoustic properties that are input to the propagation model. Simulation results show that the proposed scheme enables the continuous tracking of the range-dependent SSF parameters and their associated uncertainties assimilating new measurements each hour. It is shown that ensemble methods are required to properly manage the nonlinearity of the model. The problem of the sensitivity to the vertical array (VA) configuration is also addressed.

The development of multichannel time-reversal (T/R) processing techniques continues to progress rapidly especially when the need to communicate in a reverberant environment is critical. The underlying T/R concept is based on time-reversing the Green’s function characterizing the uncertain communications channel mitigating the deleterious dispersion and multipath effects. In this paper, attention is focused on two major objectives: (1) wideband communications leading to a time-reference modulation technique; and (2) multichannel acoustic communications in two waveguides: a stairwell and building corridors with many obstructions, multipath returns, severe background noise, disturbances, and long propagation paths (∼180ft) including disruptions (bends). It is shown that T/R receivers are easily extended to wideband designs. Acoustic information signals are transmitted with an eight-element array to two receivers with a significant loss in signal levels due to the propagation environment. The results of the n...

The development of time-reversal (T/R) communication systems is a recent signal processing research area dominated by applying T/R techniques to communicate in hostile environments. The fundamental concept is based on time-reversing the impulse response or Green's function characterizing the uncertain communications channel to mitigate deleterious dispersion and multipath effects. In this paper, we extend point-to-point to array-to-point communications by first establishing the basic theory to define and solve the underlying multichannel communications problem and then developing various realizations of the resulting T/R receivers. We show that not only do these receivers perform well in a hostile environment, but they also can be implemented with a "1 bit" analog-to-digital converter design structure. We validate these results by performing proof-of-principle acoustic communications experiments in air. It is shown that the resulting T/R receivers are capable of extracting the transmitted coded sequence from noisy microphone array measurements with zero-bit error.

Signal processing refers to the acquisition, storage, display, and generation of signals – also to the extraction of information from signals and the re-encoding of information. As such, signal processing in some form is an essential element in the practice of all aspects of acoustics. Signal processing algorithms enable acousticians to separate signals from noise, to perform automatic speech recognition, or to compress information for more efficient storage or transmission. Signal processing concepts are the building blocks used to construct models of speech and hearing. Now, in the 21st century, all signal processing is effectively digital signal processing. Widespread access to high-speed processing, massive memory, and inexpensive software make signal processing procedures of enormous sophistication and power available to anyone who wants to use them. Because advanced signal processing is now accessible to everybody, there is a need for primers that introduce basic mathematical concepts that underlie the digital algorithms. The present handbook chapter is intended to serve such a purpose.

Most acoustic sources found in the ocean environment are spatially complex and broadband. When propagating in a shallow ocean these source characteristics complicate the analysis of received acoustic data considerably. On the other hand, each of the narrow‐band lines composing the broadband source spectrum can be considered multiple observations which can be used to enhance signal levels. The usual approach is to process each line separately and combine the results to achieve more enhancement at the array than that which could be utilized for a single temporal frequency. The enhancement of broadband acoustic pressure‐field measurements using a vertical array is discussed. Here the model‐based approach is developed for a broadband source using a normal mode propagation model. It is well known from propagation theroy that a different modal structure evolves for each temporal frequency line; thus it is not surprising that the model‐based solution to this problem results in a scheme that requires a ‘‘bank’’ o...

The measurement of internal-wave signatures using synthetic aperture radar (SAR) or real aperture radar (RAR) techniques is an emerging technology that offers a viable means of locating and tracking surface ship wakes by their unique signatures. Under the assumption that the image measured by the radar is dominated by the underlying dynamics of the internal wave, we develop model-based techniques for enhancement based on a recently developed generic dispersive-wave processor. Using images synthesized by a sophisticated propagation model, it is shown that the processor not only offers a unique approach compared to the more traditional image-processing techniques that do not incorporate the propagation model, but is also capable of providing reasonable enhancement of the noisy measurements.

An unscented Kalman filter (UKF) for geoacoustic inversion using scalar and vector sound fields created by a passing ship is discussed in this paper. The continuous sound field emitted by a ship of opportunity is processed by the sequential filtering technique to estimate slowly changing environmental properties along the source range. The inversion problem is solved by the UKF with a random-walk parameter model, which is expected to perform well when dealing with highly nonlinear problems. Synthetic geoacoustic inversions are performed using multi-frequency pressure, vertical particle velocity and waveguide impedance (a ratio between pressure and vertical particle velocity) data for the geoacoustic model of a mud environment offshore at the mouth of the Amazon River in Brazil (CANOGA 12). For the preliminary tests, the sound source is composed of a flat spectrum. Numerical results demonstrate that the sequential filtering technique is capable of estimating the evolution of environmental properties along the source range. In practice, ship data have complex time-varying spectral characteristics that can greatly limit the accuracy of broadband or multi-frequency passive applications. Since the vertical waveguide impedance is independent of the source spectral level, it is preferred for environmental characterization by the sound field generated from a ship of opportunity. Because of this independence property, the vertical waveguide impedance is expected to yield a more reliable inversion than that of pressure or vertical particle velocity field.

The timely and accurate detection of threat contraband especially for ports-of-entry (e.g. harbors, bays, borders, airports) is an extremely critical problem of national security. The investigation of advanced techniques to reliably and accurately detect threats and reject non-threats is the major focus of this effort. The characterization of signal processing models based on xray transport physics is a crucial element in advanced sequential Bayesian processor designs. Incorporating the underlying statistics of x-ray interactions with materials offering a potentially unique signature of an object or item under investigation leads to a (stochastic) physics-based approach. State-space models, common in many application areas, are introduced into the x-ray radiation area. Here the resulting processor incorporating this construct is developed from a pragmatic perspective. A Gaussian application is discussed to illustrate feasibility of the overall physics-based approach. It is shown that the sequential Bayesian processor is capable of providing a reliable and accurate solution with high confidence in a timely manner for this problem based on a set of synthesized object intensity data.

A model-based approach is developed to solve an adaptive ocean-acoustic signal-processing problem. Model-based signal processing is a well-defined methodology enabling the inclusion of propagation models, measurement models, and noise models into sophisticated processing algorithms. Here, we investigate the design of a so-called model-based identifier (MBID) for a general nonlinear state-space structure and apply it to a shallow water ocean-acoustic problem characterized by the normal-mode model. In this problem, we assume that the structure of the model is known and we show how this parameter-adaptive processor can be configured to jointly estimate both the modal functions and the horizontal wave numbers directly from the measured pressure-field and sound speed. We first design the model-based identifier using a model developed from a shallow-water ocean experiment and then apply it to a corresponding set of experimental data demonstrating the feasibility of this approach. It is also shown that one of the benefits of this adaptive approach is a solution to the so-called "mismatch" problem in matched-field processing (MFP).

A model‐reference approach is developed to solve the signal enhancement problem of a laser ultrasonics application for nondestructive evaluation. In this problem a sophisticated laser thermoelastic propagation model is used to synthesize the surface displacement of the specimen under test. Once synthesized, this model response is used as the reference signal in an optimal (minimum error variance) signal enhancement scheme. Both fixed and adaptive processors are considered in this application where it is shown that a significant improvement in signal levels can be achieved over the usual methods to enhance noisy data acquired from a Michelson interferometric measurement system and increase its overall sensitivity.

A model-based approach is proposed to solve the oceanic internal wave signal processing problem that is based on state-space representations of the normal-mode vertical velocity and plane wave horizontal velocity propagation models. It is shown that these representations can be utilized to spatially propagate the modal (depth) vertical velocity functions given the basic parameters (wave numbers, Brunt-Vaisala frequency profile, etc.) developed from the solution of the associated boundary value problem as well as the horizontal velocity components. These models are then generalized to the stochastic case where an approximate Gauss-Markov theory applies. The resulting Gauss-Markov representation, in principle, allows the inclusion of stochastic phenomena such as noise and modeling errors in a consistent manner. Based on this framework, investigations are made of model-based solutions to the signal enhancement problem for internal waves. In particular, a processor is designed that allows in situ recursive estimation of the required velocity functions. Finally, it is shown that the associated residual or so-called innovation sequence that ensues from the recursive nature of this formulation can be employed to monitor the model's fit to the data.

It is known that focusing of an acoustic field by a time-reversal mirror (TRM) is equivalent to a spatio-temporal matched filter under conditions where the Green’s function of the field satisfies reciprocity and is time invariant, i.e., the Green’s function is independent of the choice of time origin. In this letter, it is shown that both reciprocity and time invariance can be replaced by a more general constraint on the Green’s function that allows a TRM to implement the spatio-temporal matched filter even when conditions are time varying.

People with heart problems have had their lives extended considerably with the development of the prosthetic heart valve. Great strides have been made in the development of the valves through the use of improved materials as well as efficient mechanical designs. However, since the valves operate continuously over a long period, structural failures can occur--even though they are relatively uncommon. Here the development of techniques to classify the valve either as having intact struts or as having a separated strut, commonly called single leg separation, is discussed. In this paper the signal processing techniques employed to extract the required signals/parameters are briefly reviewed and then it is shown how they can be used to simulate a synthetic heart valve database for eventual Monte Carlo testing. Next, the optimal classifier is developed under assumed conditions and its performance is compared to that of an adaptive-type classifier implemented with a probabilistic neural network. Finally, the adaptive classifier is applied to a data set and its performance is analyzed. Based on synthetic data it is shown that excellent performance of the classifiers can be achieved implying a potentially robust solution to this classification problem.

A model-based approach to solve a deep water ocean acoustic signal processing problem based on a state-space representation of the normal-mode propagation model is developed. The design of a model-based processor (MBP) for signal enhancement employing an array consisting of a large number of sensors for a deep ocean surveillance operation is discussed. The processor provides enhanced estimates of the measured pressure-field, modes, and residual (innovations) sequence indicating the performance or adequacy of the propagation model relative to the data. It is shown that due to the structure of the normal-mode model the state-space propagator is not only feasible for this large scale problem, but in fact, can be implemented by a set of decoupled parallel second-order processors, implying a real-time capability. In the paper we discuss the design and application of the processor to a realistic set of simulated pressure-field data developed from a set of experiments and sound speed parameters. >

Sequential detection is a methodology developed essentially by Wald [1] in the late 1940s providing an alternative to the classical batch methods evolving from the basic Neyman–Pearson theory of the 1930s [2, 3]. From the detection theoretical viewpoint, the risk (or error) associated with a decision typically decreases as the number of measurements increases. Sequential detection enables a decision to be made more rapidly (in most cases) employing fewer measurements while maintaining the same level of risk. Thus, the aspiration is to reduce the decision time while maintaining the risk for a fixed sample size. Its significance was truly brought to the forefront with the evolution of the digital computer and the fundamental idea of acquiring and processing data in a sequential manner. The seminal work of Middleton [2, 4, 5, 6, 7] as well as the development of sequential processing techniques [8, 9, 10, 11, 12, 13] during the 1960s provided the necessary foundation for the sequential processor/detector that is applied in a routine manner today [7, 8, 9, 10, 11, 13].

The shallow ocean is an ever changing environment primarily due to temperature variations in its upper layers (< 100 m) directly affecting sound propagation throughout. The need to develop processors capable of tracking these changes implies a stochastic as well as an environmentally adaptive design. The stochastic requirement follows directly from the multitude of variations created by uncertain parameters and noise. Some work has been accomplished in this area, but the stochastic nature was constrained to Gaussian uncertainties. It has been clear for a long time that this constraint was not particularly realistic leading to a Bayesian approach that enables the representation of any uncertainty distribution. Sequential Bayesian techniques enable a class of processors capable of performing in an uncertain, nonstationary (varying statistics), non-Gaussian, variable shallow ocean environment. A solution to this problem is addressed by developing a sequential Bayesian processor capable of providing a joint solution to the modal function tracking (estimation) and environmental adaptivity problem. The posterior distribution required is multi-modal (multiple peaks) requiring a sequential (nonstationary) Bayesian approach. Here the focus is on the development of a particle filter (PF) capable of providing reasonable performance for this problem. In our previous effort on this problem nonlinear/non-Gaussian processors were developed to operate on synthesized data based on the Hudson Canyon experiment using normal-mode representations. Here we extend the processors by applying them to the actual hydrophone measurements obtained from the 23-element vertical array. The adaptivity problem is attacked by allowing the modal coefficients to be estimated from the measurement data jointly along with tracking of the modal functions—the main objective.

This report summarizes key research, development, and technology advancements in Lawrence Livermore National Laboratory's Engineering Directorate for FY2010. These efforts exemplify Engineering's nearly 60-year history of developing and applying the technology innovations needed for the Laboratory's national security missions, and embody Engineering's mission to ''Enable program success today and ensure the Laboratory's vitality tomorrow.'' Leading off the report is a section featuring compelling engineering innovations. These innovations range from advanced hydrogen storage that enables clean vehicles, to new nuclear material detection technologies, to a landmine detection system using ultra-wideband ground-penetrating radar. Many have been recognized with RD all are examples of the forward-looking application of innovative engineering to pressing national problems and challenging customer requirements. Engineering's capability development strategy includes both fundamental research and technology development. Engineering research creates the competencies of the future where discovery-class groundwork is required. Our technology development (or reduction to practice) efforts enable many of the research breakthroughs across the Laboratory to translate from the world of basic research to the national security missions of the Laboratory. This portfolio approach produces new and advanced technological capabilities, and is a unique component of the value proposition of the Lawrence Livermore Laboratory.more » The balance of the report highlights this work in research and technology, organized into thematic technical areas: Computational Engineering; Micro/Nano-Devices and Structures; Measurement Technologies; Engineering Systems for Knowledge Discovery; and Energy Manipulation. Our investments in these areas serve not only known programmatic requirements of today and tomorrow, but also anticipate the breakthrough engineering innovations that will be needed in the future.« less

A system-theoretic approach is proposed to investigate the feasibility of reconstructing a sound velocity profile (SVP) from acoustical hydrophone measurements. A state-space representation of the normal-mode propagation model is used. It is shown that this representation can be utilized to investigate the so-called observability of the SVP from noisy measurement data. A model-based processor is developed to extract the required information, and it is shown that even in cases where limited SVP information is available, the SVP can be estimated using this approach. Based on this framework, investigations are made of model-based solutions to the sound velocity profile and related parameter estimation problems. In particular, a processor is designed that allows in situ recursive estimation of the sound velocity profile from simulated data. >

The development of model-based processing techniques in ocean acoustics is well-known evolving from the pure statistical approach of maximum likelihood parameter estimation, matched-field processing and sequential model-based processing for Gaussian uncertainties. More recent model-based techniques such as unscented Kalman filtering (UKF) and sequential Markov chain Monte Carlo (MCMC) methods using particle filters (PF) have been developed to improve both unimodal distribution estimates (UKF) as well as multimodal estimates (PF). In this paper we apply both techniques to provide enhanced signal estimates for acoustic hydrophone measurements on a vertical array and compare their performance. We use a normal-mode propagation solution to provide synthetic data in order to make the comparison and demonstrate the approach which will open the area to direct extensions such as localization, broadband processing, inversion, etc. We show how the normal-mode model can be incorporated directly into the processors along with the measurement array enabling the resulting enhancement capabilities.