Adaptive signal processing

Ultra-Short-BaseLine (USBL) is the most practical underwater acoustic positioning system for autonomous underwater vehicles because of its small space requirement. The objective of this research is to develop a USBL system capable of estimating a source location transmitting frequency-hopped tones sequences. Such sequences are characteristic of spread spectrum signaling used in underwater acoustic communication network. It must be able to provide azimuth and elevation of a modem-type source with an accuracy of 0.3 degrees; for both angles using the synchronization stage of the transmission. The acoustic antenna is composed of four transducers arranged as a tetrahedron. Using the model of Quazi and Lerro, which provides an expression for the variance of the bearing angle, azimuth and elevation of the transmitter are estimated employing maximum likelihood estimation. This system is simulated, tested and calibrated in a tank. Simulated results satisfy the requirement with a SNR of 32dB and 8 symbols. The latest experimental measurements present an accuracy of 3 degrees.

The research presented investigates the use of cumulants in conjunction with a spectral estimation technique of the signal subspace to perform the blind separation of statistically independent signals with low signal-to-noise ratios under a narrowband assumption. A new blind source separation (BSS) algorithm is developed that makes use of the generalized eigen analysis of a matrix pencil defined on two similar spatial fourth-order cumulant matrices. The algorithm works in the presence of spatially and/or temporally correlated noise and, unlike most existing higher-order BSS techniques, is based on a spectral estimation technique rather than a closed loop optimization of a contrast function, for which the convergence is often problematic. The dissertation makes several contributions to the area of blind source separation. These include: (1) Development of a robust blind source separation technique that is based on higher-order cumulant based principle component analysis that works at low signal-to-noise ratios in the presence of temporally and/or spatially correlated noise. (2) A novel definition of a spatial fourth-order cumulant matrix suited to blind source separation with non-equal gain and/or directional sensors. (3) The definition of a spatial fourth-order cumulant matrix-pencil using temporal information. (4) The concept of separation power efficiency (SPE) as a measure of the algorithm's performance. Two alternative definitions for the spatial fourth-order cumulant matrix that are found in the literature are also presented and used by the algorithm for comparison. Additionally, the research contributes the concept of wide sense equivalence between matrix-pencils to the field of matrix algebra. The algorithm's performance is verified by computer simulation using realistic digital communications signals in white noise. Random mixing matrices are generated to ensure the algorithm's performance is independent of array geometry. The computer results are promising and show that the algorithm works well down to input signal-to-noise ratios of -6 dB, and using as few as 250 x 103 samples.

A novel method of achieving stable high-speed underwater acoustic communication with a fairly low-complexity of implementation is proposed. The proposed approach is to split the space and time processing into two separate sub-optimal processes. As a result, processing complexity is significantly reduced and the instabilities associated with large tap vectors at large time-frequency spread products are reduced. The proposed space-time signal processing method utilizes a different beamformer optimization strategy compared to the time domain optimization strategy. This allows to separately adjust the adaptation parameters for the spatial and temporal characteristics of the signal, which have vastly different requirements. The time domain signal is subject to variations in phase that require rapid filter updates whereas the directional characteristics of the signal do not vary appreciably over the message length and do not require a rapid adaptation response. The proposed method allows for high-speed underwater acoustic communication in very shallow water using coherent modulation techniques, and offers a series of unique features: significant reduction of the signal-to-noise and interference ratio (SNIR), improvement of the bandwidth efficiency by reduction of the forward-error coding redundancy requirements, real-time evaluation of the time-spread by Doppler spread product (BL) and channel stability estimate. Experimental results demonstrate that stable acoustic communication can be achieved at rates of 32000 bits per second at a distance of 3 km, in 40 feet of water and in sea-state 2 conditions. Fast and slow fading properties of the channel are measured, as the BL product can vary by a decade in 116 ms, and by two decades within minutes, from 0.001 to 0.1. The real-time analysis shows a strong correlation between time spread, Doppler spread, spatial coherence of the acoustic channel and communication performance. Overall, this research provides more scientific and experimental ground to understand the limitations of multi-channel adaptive receiver techniques in terms of stability, hardware requirements and channel tracking capability.

The DUCKW-Ling is an 8.3 foot long, amphibious water plane area twin hull (SWATH) concept vehicle which is propelled by a pair of crawler tracks on land and dual propellers when water-borne. In its operational zone, the vehicle's dynamics change dramatically as it transitions from being completely water-borne and buoyancy supported to being completely land-borne and track supported. In the water environment, a cascaded, first-order sliding mode controller was used to control the surge and heading of the vehicle, and was capable of having a faster response when compared to using a proportional controller. Additionally, field trials of the DUKW-Ling show the capability of the vehicle to navigate and track predetermined waypoints in both terrestrial and aquatic terrains. In the transitional zone, the electric motor current from the tracks was used as the feedback mechanism to adequately actuate the propellers and tracks in the system as the dynamics of the vehicle change.

This thesis discusses a new approach to tracking the REMUS 100 AUV using a modified version of the Florida Atlantic University (FAU) ultrashort baseline (USBL) acoustic positioning system (APS). The REMUS 100 is designed to utilize a long baseline (LBL) acoustic positioning system to obtain positioning data in mid-mission. If the placement of one of the transponders of the LBL field is known, then tracking the position of the REMUS 100 AUV using a passive USBL array is possible. As part of the research for this thesis, the FAU USBL system was used to find a relative range between the REMUS 100 ranger and a LBL transponder. This relative range was then combined with direction of arrival information and LBL field component position information to determine an absolute position of the REMUS 100 ranger. The outcome was the demonstration of a passive USBL based tracking system.

The design and integration of an unmanned surface vehicle (USV) control system is described. A survey of related work in both USV control, and unmanned vehicle operating software is presented. The hardware subsystem comprising a modular Guidance, Navigation, and Control (GNC) package is explained. A multi-threaded software architecture is presented, utilizing a decentralized, mutex-protected shared memory inter-process communication subsystem to provide interoperability with additional software modules. A generic GNC approach is presented, with particular elaboration on a virtual rudder abstraction of differential thrust platforms. A MATLAB Simulink simulation is presented as a tool for developing an appropriate controller structure, the result of which was implemented on the target platform. Software validation is presented via a series of sea trials. The USV was tested both in open- and closed-loop control configurations, the results of which are presented here. Lastly recommendations for future development of the GNC system are enumerated.

The design, implementation, and testing of an experimental setup intended to evaluate the dynamic maneuvering performance of the Wave Adaptive Modular Vessel (WAM-V) class USV12, a 3.7 meter unmanned surface vehicle (USV) is described. A comprehensive sensor package was designed, fabricated and assembled to record the vehicle's dynamic response to various control inputs. All subsystems were fabricated and installed on a test vehicle, GUSS, and full system, open-loop maneuvering tests were conducted to show validity of data collection technique. Simulations were performed using model parameters found in the literature to create a "simulated experimental" data set, upon which system identification techniques were used to rediscover a suitable model with similar parameterization. Combined, the sensor package and the method for creating this model support future work in the design of automatic control, navigation, and guidance systems for the WAM-V USV12.

Flow over a rough surface is known to radiate sound as a dipole source that is directional. In order to better understand this source, measurements are being made in a wind tunnel using a microphone array. The measurements collected by a microphone array are beamformed to give a source image and can be deconvolved with an assumed point spread function in order to obtain the source levels. This thesis considers alternative analysis algorithms that can be used to analyze wind tunnel data. Only numerical examples of how these algorithms work will be presented and the analysis of real data will be considered in later studies. It will be shown how estimates can be made of the source directivity by comparing the measured data with a theoretical source model and minimizing the error between the model and the measurements.

Standard GPS receivers are unable to provide the rate or precision required when used on a small vessel such as an Unmanned Surface Vehicles (USVs). To overcome this, the thesis presents a low cost high rate motion measurement system for an USV with underwater and oceanographic purposes. The work integrates an Inertial Measurement Unit (IMU), a GPS receiver, a flux-gate compass, a tilt sensor and develops a software package, using real time data fusion methods, for an USV to aid in the navigation and control as well as controlling an onboard Acoustic Doppler Current Profiler (ADCP).While ADCPs non-intrusively measure water flow, they suffer from the inability to discriminate between motions in the water column and self-motion. Thus, the vessel motion contamination needs to be removed to analyze the data and the system developed in this thesis provides the motion measurements and processing to accomplish this task.

The goal of a speech enhancement algorithm is to remove noise and recover the original signal with as little distortion and residual noise as possible. Most successful real-time algorithms thereof have done in the frequency domain where the frequency amplitude of clean speech is estimated per short-time frame of the noisy signal. The state of-the-art short-time spectral amplitude estimator algorithms estimate the clean spectral amplitude in terms of the power spectral density (PSD) function of the noisy signal. The PSD has to be computed from a large ensemble of signal realizations. However, in practice, it may only be estimated from a finite-length sample of a single realization of the signal. Estimation errors introduced by these limitations deviate the solution from the optimal. Various spectral estimation techniques, many with added spectral smoothing, have been investigated for decades to reduce the estimation errors. These algorithms do not address significantly issue on quality of speech as perceived by a human. This dissertation presents analysis and techniques that offer spectral refinements toward speech enhancement. We present an analytical framework of the effect of spectral estimate variance on the performance of speech enhancement. We use the variance quality factor (VQF) as a quantitative measure of estimated spectra. We show that reducing the spectral estimator VQF reduces significantly the VQF of the enhanced speech. The Autoregressive Multitaper (ARMT) spectral estimate is proposed as a low VQF spectral estimator for use in speech enhancement algorithms. An innovative method of incorporating a speech production model using multiband excitation is also presented as a technique to emphasize the harmonic components of the glottal speech input.