Sanjeev Agarwal

Adaptive Shape Control of Laminated Composite Plates Using Piezoelectric Materials

Shape control of laminated composite plates with integrated piezoelectric actuators is discussed. The effectiveness of piezoelectric actuators and position sensors is investigated for shape control under the influence of quasistatically varying unknown loads. The shape control problem is divided into two parts. 1) For the desired shape function, calculate initial actuator input voltages such that a measure of the mean-squared error between the desired and the achieved shape is minimized. 2) An adaptive feedback algorithm is developed so as to minimize the distortion in the shape introduced by the quasistatically varying loads on the structure. A finite element model based on the shear deformation theory is used to verify the performance of the shape control methodologies.

Self-adaptive vibration control of smart composite beams using recurrent neural architecture

A self-adapting vibration control system is developed for damping augmentation in smart composite beams. The conventional vibration control approaches are limited by the requirement of an explicit and often accurate identification of the system dynamics and subsequent offline design of an optimal controller. In the present study a self-adapting vibration control system is developed. A hybrid system comprised of a dynamic diagonal recurrent neural network (DRNN) and an adaptable feed forward neural network is used to control the beam vibrations. Sensing and actuation are achieved using piezoelectric sensors and actuators. A finite element model based on a higher-order shear deformation theory is used to simulate the vibration response of laminated composite beams with integrated piezoelectric sensors and actuators. The dynamic effects of mass and stiffness of the piezoelectric patches are considered in the model. The performance of the DRNN controller is verified for arbitrary initial conditions and loadings. A robustness study including the effects of tip mass, structural parameter variation and partial loss of sensor output is performed. The performance with partial failure of control actuation is also examined. It is seen that the robustness and control capabilities of the hybrid control system are excellent.

Dynamic Modeling and Neural Control of Composite Shells Using Piezoelectric Devices

A modal dynamic model was developed for the active vibration control of laminated doubly curved shells with piezoelectric sensors and actuators. The dynamic effects of the mass and stiffness of the piezoelectric patches were considered in the model. Finite element equations of motion were developed based on shear deformation theory and implemented for an isoparametric shell element. The mode superposition method was used to transform the coupled finite element equations into a set of uncoupled equations in the modal coordinates. A robust controller was developed using Linear Quadratic Gaussian with Loop Transfer Recovery (LQG/LTR) design methodology to calculate the gain and actuator voltage requirements. A Neural Network controller was then designed and trained offline to emulate the performance of the LQG/LTR controller. Numerical results have been presented for a flat plate and a spherical shell showing the variation in initial conditions and structural parameters. The neural network controller was shown to effectively emulate the LQG/LTR controller.

Detection and localization of multiple R/C electronic devices using array detectors

Accurate detection and localization of unintended emissions from multiple radio-controlled electronic devices have a wide range of security applications. First, this paper introduces a cost-effective mobile array detector. Subsequently, a novel scheme is presented for detection and localization of multiple devices emitting unintended passive emissions. Peak detection is employed in detecting the devices with superheterodyne receivers (SHRs). Experimental results demonstrate that the proposed scheme outperforms the traditional methodologies with a single antenna. A 2-D array is also proposed in this paper for localization of devices with SHRs using the angle of arrival. A new, synthetic aperture radar (SAR)-based scheme called edge synthetic aperture radar (ESAR) is also introduced in this paper for localization. Finally, the results for localization and error in localization are analyzed. The ESAR method reduces the error in localization up to 75% and increases accuracy in detection of multiple devices.

Algorithms for IR-imagery-based airborne landmine and minefield detection

In this paper we revisit and enhance various algorithms for landmine detection, discrimination and recognition. Single- band and multi-band medium wave infrared (MWIR) image data from the May data collection (part of Lightweight Airborne multispectral Minefield Detection-Interim (LAMBD-I) program) is used for the analysis. In particular discrimination based on gray-scale moments is explored and its effectiveness is evaluated for surface mines under IR imaging using receiver operating characteristics (ROC) curves. The discriminatory power of gray-scale moments is compared with the RX and matched fiber based detectors for different terrain (e.g., grass,sand) and different mine types. The performance of single-band (broadband) MWIR imagery is compared with multi- band (short-pass and long-pass) MWIR images. Also direct multi-band detection is compared against fusion of multiple single-band responses. Gray-scale moment based target discrimination at potential target locations, identified by RX or matched fiber detectors, is shown to be computationally efficient and provides better performance in terms of reduced false alarms for comparable probability of detection. An evolutionary framework for minefield identification, in the presence of inevitable false targets, is also presented. Starting from the locations of individual mine targets and false alarms, the evolutionary algorithm is used to identify the underlying structure of the minefield. Issues in the detection of different minefield layouts are discussed. Preliminary implementation shows the promise of this approach in identification of a wide variety of minefields.

Lqg/Ltr-Based Robust Control of Composite Beams with Piezoelectric Devices

The performance of an LQG/LTR-based multi-input multi-output robust vibration control system for a laminated composite beam is investigated. A finite element model based on a higher order shear defor mation theory and accounting for piezoelectric effects is developed. The model is thus applicable to both thick and thin laminated composite beams. The lateral strain is also incorporated in the model by a systematic re duction of two-dimensional plate constitutive equations. The mode superposition method is used to transform the coupled finite element equations of motion in the physical coordinates into a set of reduced uncoupled equations in the modal coordinates. The state space model of the system is obtained in the reduced-order modal coordinates. An LQG/LTR-based robust controller is designed using the reduced-order state space model of the structure. The performance of the controller is verified for various arbitrary initial conditions of the beam. The effect of structural parameter variation on the closed-loop system performance is also investi gated. The performance robustness of the linear quadratic Gaussian with loop transfer recovery (LQG/LTR) controller is compared with that of the proportional feedback controller.-

Detection of Super-Regenerative Receivers Using Hurst Parameter

The accurate and reliable detection of unintended emissions from radio receivers has a broad range of commercial and security applications. This paper presents and analyzes Hurst parameter-based detection method for super-regenerative receivers (SRRs). SRRs are low cost, easily manipulated, and widely used in common remote devices including doorbells, garage door openers, and remote controlled (R/C) toys. By design the SRR is a passive device that should only receive an RF signal. However, it also emits a low power, unintended electromagnetic signal. Such unintended emissions are enhanced by the presence of a known stimulating signal. Also, the emission is referred to as a device signature since it can uniquely identify the devices. The proposed detection method exploits a self-similarity property of such emissions to distinguish it from background noise. Hurst parameter quantifies the self-similarity. It is employed to detect and identify the SRR-based devices even if the signal fades into a noise.

Detection and Localization of Multiple R/C Electronic Devices Using Array Detectors

Detection and localization of multiple radio controlled (RC) electronic devices has many security applications. This paper first proposes a cost-effective mobile array detector for such RC electronic devices. Subsequently, by using the array detector, a novel scheme is presented for detection, identification, and localization of multiple devices emitting unintended passive emissions. The Hurst parameter method is proposed for detection and identification of devices with super-regenerative receivers while peak detection is used for detection and identification of devices with super heterodyne receivers. Experimental results demonstrate that the proposed scheme outperforms traditional methodologies with single antenna.

Warfighter-in-the-loop: Mental models in airborne minefield detection

The warfighter analyst in the data processing ground control station plays an integral role in airborne minefield detection system. This warfighter-in-the-loop (WIL) is expected to reduce the minefield false alarm rate by a factor of 5. In order to achieve such a significant false alarm reduction and to facilitate the development of an efficient WIL interface, it is critical to evaluate different aspects of WIL operations for airborne minefield detection. Recently, researchers at the University of Missouri-Rolla have developed a graphical user interface (HILMFgui) application using MATLAB to evaluate minefield detection performance for the operator. We conducted a series of controlled experiments with HILMFgui using ten participants. In these experiments, we video-recorded all the experiments and conducted post-experiment interviews to learn more about the usability of the interface and the cognitive processes involved in minefield detection. The effect of various factors including the availability of automatic target recognition (ATR), availability of zoom and time constraints were considered to evaluate their influence on operator performance. Qualitative results of the factors affecting the warfighter performance in the minefield detection loop are discussed. Through the qualitative data analysis, we observed two different types of participants (classified here as aggressive and cautious). We also identified three primary types of mental models: mine centric, mine-field centric, and logical placement. Those who used a primarily mine focus had a substantially higher false alarm rate than those whose mental models were more consistent with a mine-field centric or logical placement perspective.

Determination of Aircraft Orientation for a Vision-Based System Using Artificial Neural Networks

An algorithm for real-time estimation of 3-D orientation of an aircraft, given its monocular, binary image from an arbitrary viewing direction is presented. This being an inverse problem, we attempt to provide an approximate but a fast solution using the artificial neural network technique. A set of spatial moments (scale, translation, and planar rotation invariant) is used as features to characterize different views of the aircraft, which corresponds to the feature space representation of the aircraft. A new neural network topology is suggested in order to solve the resulting functional approximation problem for the input (feature vector)-output (viewing direction) relationship. The feature space is partitioned into a number of subsets using a Kohonen clustering algorithm to express the complex relationship into a number of simpler ones. Separate multi-layer perceptrons (MLP) are then trained to capture the functional relations that exist between each class of feature vectors and the corresponding target orientation. This approach is shown to give better results when compared to those obtained with a single MLP trained for the entire feature space.