A multilayer data-driven framework for robust structural health monitoring based on a comprehensive application of machine learning and signal processing techniques is introduced. This paper focuses on demonstrating the effectiveness of the framework for damage detection in a steel pipe under environmental and operational variations. The pipe was instrumented with piezoelectric wafers that can generate and sense ultrasonic waves. Damage was simulated physically by a mass scatterer grease-coupled to the surface of the pipe. Benign variations included variable internal air pressure and ambient temperature over time. Ultrasonic measurements were taken on three different days with the scatterer placed at different locations on the pipe. The wave patterns are complex and difficult to interpret, and it is even more difficult to differentiate the changes produced by the scatterer from the changes produced by benign variations. The sensed data were characterized by 365 features extracted from a variety of signal-processing techniques. Automated feature selection methods were then developed using an adaptive boosting algorithm to identify the most effective features for damage detection. With the selected features, five machine-learning classifiers were formulated based on adaptive boosting and support vector machines and achieved 98.5–99.8% average accuracy during random testing and 84.2–89% average accuracy during systematic testing. In addition, other metrics for classifier evaluation generated from a confusion matrix and from a receiver operating characteristic curve are reported.

In a low-angle target parameter estimation scenario, the backscattered signals from targets are often distorted due to clutter and multipath, which significantly degrades the performance of direction-of-arrival (DOA) estimation. In general, the backscattered multipaths are modeled as coherent signals with respect to the direct path. Decorrelation algorithms such as spatial smoothing and matrix reconstruction can mitigate the effect of multipath. However, most of these methods will perform poorly or even fail for solving parameter estimation problem of low-angle targets in complex terrain where there are rich multipath scatterings. This paper presents a novel method that combines time reversal (TR) technique, coherent signal-subspace method (CSM), and multiple signal classification algorithm to perform DOA estimation in a low-angle target scenario. The TR technique takes advantage of multipath echoes recorded by a sensor array and adaptively adjusts TR probing waveforms to increase the signal-to-noise ratio in a rich scattering environment. Moreover, the CSM method compresses the energy of a signal into a predefined subspace to exploit the full time-bandwidth product of signal sources and cope with coherent wavefronts. As a result, the proposed new DOA estimation algorithm outperforms significantly the conventional methods. The Cramer-Rao bounds analysis and numerical simulations commendably validate the accuracy and robustness of the proposed method.

In this paper, we present and demonstrate a low complexity elastic wave signaling and reception method to achieve high data rate communication on dispersive solid elastic media, such as metal pipes, using piezoelectric transducers of PZT (lead zirconate titanate). Data communication is realized using pulse position modulation (PPM) as the signaling method and the elastic medium as the communication channel. The communication system first transmits a small number of training pulses to probe the dispersive medium. The time-reversed probe signals are then utilized as the information carrying waveforms. Rapid timing acquisition of transmitted waveforms for demodulation over elastic medium is made possible by exploring the reciprocity property of guided elastic waves. The experimental tests were conducted using a National Instrument PXI system for waveform excitation and data acquisition. Data telemetry bit rates of 10 kbps, 20 kbps, 50 kbps and 100 kbps with the average bit error rates of 0, 5.75 × 10−4, 1.09 × 10−2 and 5.01 × 10−2, respectively, out of a total of 40, 000 transmitted bits were obtained when transmitting at the center frequency of 250 kHz and a 500 kHz bandwidth on steel pipe specimens. To emphasize the influence of time reversal, no complex processing techniques, such as adaptive channel equalization or error correction coding, were employed.

Time reversal (TR) is a promising technique in detecting weak targets immersed in clutter. This paper reports a series of electromagnetic (EM) domain experiments on TR detection using a single antenna pair in a controlled scattering environment. The TR likelihood ratio detector is compared with the conventional energy detector and the matched filter. The EM experiments demonstrate the superiority of TR detection over conventional detection in rich multipath scattering environments.

This paper describes a software-defined radar sensor based on successive multiple-frequency continuous-wave (S-MFCW) modulation. S-MFCW is a radar modulation technique capable of detecting static and moving targets with high accuracy while modulating the transmitted signal over a narrow frequency bandwidth. The proposed software-defined radar is implemented as a highly reconfigurable system made of a field-programmable gate array (FPGA)-based system-on-chip (SoC), a wideband homodyne radio frequency (RF) transceiver, and two antennas. The utilized programmable general-purpose SoC and transceiver front end allow for some of the radar components to be reconfigured by a software program running on the SoC. This radar system adequately follows the software-defined radar paradigm wherein most of the radar operation is configured and controlled using software programs, thus providing the ability to combine multiple radar functions on the same hardware resource. The presented radar sensor is experimentally evaluated in both outdoor and indoor measurement scenarios. The radar operates in the S-band and occupies a narrow virtual bandwidth of 34 MHz when transmitting S-MFCW tones whose frequencies are shifted by two gaps of 4 and 30 MHz, from 3.550 to 3.584 GHz. Experimental results show that the proposed radar system is capable of measuring short-range targets with a range accuracy of less than 21 cm.

This paper presents an accelerated iterative Landweber method for nonlinear ultrasonic tomographic imaging in a multiple-input multiple-output (MIMO) configuration under a sparsity constraint on the image. The proposed method introduces the emerging MIMO signal processing techniques and target sparseness constraints in the traditional computational imaging field, thus significantly improves the speed of image reconstruction compared with the conventional imaging method while producing high quality images. Using numerical examples, we demonstrate that incorporating prior knowledge about the imaging field such as target sparseness accelerates significantly the convergence of the iterative imaging method, which provides considerable benefits to real-time tomographic imaging applications.