2015 Symposium Abstracts - Electrical Engineering

ECE-01  Sub-Band Detection Of Primary User Emulation Attacks In OFDM-Based Cognitive Radio Networks

Authors: Ahmed Alahmadi; Tianlong Song; Tongtong Li

Abstract: This research considers the primary user emulation attack (PUEA) problem in OFDM-based cognitive radio networks, and proposes a robust and efficient AES-based digital TV (DTV) scheme. In the proposed scheme, the existing reference sequence used to generate the pilot subcarriers in the DTV frames is encrypted using the Advanced Encryption Standard (AES) algorithm for accurate sub-band detection of the authorized primary user, as well as malicious user. For primary user detection, we investigate the cross-correlation between the received sequence and the AES-encrypted reference sequence over a specific frequency sub-band. The malicious user detection can be performed by investigating the auto-correlation of the received sequence. It is shown that, with the AES-based DTV scheme, both the primary user and malicious user can be detected accurately under primary user emulation attacks.

 

ECE-02  Skill Level Assessment For Teleoperator Based On Electroencephalography Signal

Authors: Mustaffa Alfatlawi; Yunyi Jia; Ning Xi; Shuang Liu

Abstract: Most of the researches in the field of robotic teleoperations have focused on the robots and the communication systems and paid less attention to the teleoperator as an important factor to achieve the tasks in optimal way. There are many aspects can be taken into consideration in the tele operator, the most important one is the skill level. The skill level of the human teleoperator usually has great influence on the performance of teleoperation, since different teleoperators with different skill levels may generate different commands and decisions in teleoperation. Even the same person may improve the performance after effective training. Investigation on the influence of operator skill is thus very important. In our research, a skill level detection approach is presented based on the emotional state of the teleoperator which can be detected from the EEG signal. The conceptual link between the EEG signals and the skill level is built upon many neurophysiological and psychological researches. An experiment is designed to train a neural network in order to create a skill estimator directly form the EEG signals. The preliminary results have shown a strong connection between the emotional state of the teleoperator and the skill level.

 

ECE-03  Design And Development Of An LED-Based Optical Communication System

Authors: Mohammed Al-rubaiai; Xiabo Tan

Abstract: Strong attenuation of most electromagnetic signals in water is the main reason why underwater exploration and surveillance and other applications currently rely on sub-sea cables and tethers to send data back to the user. Though acoustic modems have long been the default wireless communication method for underwater applications, they incur high cost and can only deliver low data rates. In this work, an LED-based optical communication system is presented for applications requiring high speed, lower power,low complexity, and low-to-median communication range. A transmitter with super bright blue LED and a receiver based on blue enhanced photo diode systems were developed and tested with the goal of transmitting data at high rates over distances of at least 20 meters. Test results in a swimming pool showed the successful transmission of large data over a distance of 23 meters and at transmission rates of 100 kbps.A tracking control system has also been developed for underwater Optical Communications to prevent misalignment due to relative movement between the transmitter and the receiver.

This work was supported in part by National Science Foundation

 

ECE-04  MSU Jump Robot To Cover Maximum Area In Wireless Mobile Network With Minimum Energy

Authors: Emad Alsaedi; Ning Xi

Abstract: Mobile Robots in a network are supplied with limited energy on board. Therefore utilizing the energy in a mobile robot is a crucial case. The wheeled Robot is well limited on mobility for a non ideal surface (rugged terrains). Meanwhile, the Jump Robot can easily operate in a rugged none ideal surface.The Jump Robot has the ability to sense and transmit data to the base station for processing, Therefore, the overall goal is to find the shortest path from start point to the goal point.The model of movement is formatted and solved using a population based optimization method knows as the differential evolution. .

 

ECE-05  A Tensor-Based Approach To Tracking Dynamics Of Functional Connectivity In The Brain

Authors: Selin Aviyente; David Zoltowski; Arash Mahyari; Edward M. Bernat

Abstract: With the advances in neuroimaging technology, it is now possible to collect multi-channel neurophysiological activity across different experimental conditions and subject groups. In this work, we represent these higher-order datasets using tensor representation and propose tensor tracking and compression algorithms to identify change points in network topography and to summarize the quasi-stationary network states. Using lower rank approximations to the tensor at each time point, we employ subspace distance metrics to quantify the change in the network across time and identify the change points. Once the change points are detected, each time interval is compressed to a single network state representation through tensor-matrix projection.

This work was supported in part by NSF grant 1218377

 

ECE-06  On Developing And Enhancing Plant-Level Disease Rating Systems In Real Fields

Authors: Yousef Atoum; Jamal Afridi; Xiaoming Liu; Mitchell McGrath; Linda Hanson

Abstract: Cercospora leaf spot (CLS) is one of the most serious diseases of sugar beet worldwide, and if uncontrolled, causes nearly complete defoliation and loss of revenue for beet growers. The beet sugar industry continuously seeks CLS-resistant sugar beet cultivars as one strategy to combat this disease. Normally human experts manually observe and rate the resistance of a large variety of sugar beet plants over a period of a few months. Unfortunately, this procedure is laborious and subjective from one expert to another resulting in disagreements on the level of resistance. Therefore, we propose a novel computer vision system, CLS Rater, to automatically rate plant images in the real field to the "USDA scale". Given a set of plant images captured by a tractor-mounted camera, CLS Rater extracts multi-scale superpixels, where in each scale a novel histogram of importances feature encodes both the within-superpixel local and across-superpixel global appearance variations. These features at different superpixel scales are then fused for learning a regressor that estimates the rating for each plant image. We further address the issue of the noisy labels by experts in the field, and propose a method to enhance the performance of the CLS Rater by automatically calibrating experts ratings to ensure consistency. Experimental results on real field data show that both the CLS Rater and the enhanced CLS Rater to be highly consistent with the rating errors of 0.65 and 0.59 respectively, which demonstrates a higher consistency than the rating standard deviation of 1.31 by human experts.

This work was supported in part by Michigan Sugar Company Competitive Grant and Project Greeen

 

ECE-07  Passive Joints For Robotic Fish Pectoral Fins: Design, Modeling, And Experimental Results

Authors: Sanaz Bazaz Behbahani; Xiaobo Tan

Abstract: Aside from the tail, paired pectoral fins are an important actuation mechanism in a robotic fish. There are several works adopting a rigid connection between the actuator and the pectoral fins, which requires different actuation speeds in power and recovery strokes to produce a net thrust. The latter results in low actuation efficiency and high control complexity.In this work we explore two different types of passive joints, one being the flexible rowing passive joint and the other being the flexible feathering joint, to overcome the problems mentioned above. A dynamic model is proposed for each case and subsequently evaluated through experiments. We also provide a comparison between the two joints through experimental results on forward swimming velocity, turning radius, and turning period over different fin-beat frequencies. Overall, the flexible rowing joint outperforms the flexible feathering passive joint and the rigid joint connection. Furthermore, we investigate the influence of actual pectoral fin size, joint dimension, and joint stiffness on the locomotion performance of the robotic fish.

This work was supported in part by National Science Foundation (Grant DBI-0939454, CNS-1059373, IIS-1319602, and CCF-1331852)

 

ECE-08  Investigation Of A Comprehensive Confidence Measure In NDE

Authors: Portia Banerjee; Seyed Safdarnejad; Lalita Udpa; Satish Udpa

Abstract: Quantitative assessment of reliability of classification results is critical in detection and characterization of anomalies in any non-destructive evaluation (NDE) application. In automated data-analysis systems, reliability measure enables the system to automatically flag indications where operator intervention is required, and reduces maintenance costs and risks. Classification results are affected by inherent ambiguity of defect classes, non-discriminative features, inadequate training samples and poor data quality. Although these sources of uncertainties in classification have been studied, formulating a single measure which quantifies all of them together has not been done to date. From Bayesian point of view, posterior probability is considered as a confidence measure. Posterior probability of occurrence of an event is representative of inter-class similarities and intra-class distance and thus, may be used as a measure of inherent ambiguity of classes and discriminative quality of features. However, estimation of posterior probability itself is affected by size of available training samples. In this paper, we develop a framework to incorporate these major sources of classification error in a single quantity. In lieu of the simplistic assumption, parameters of the distribution of a class are assumed to be random variables. Bootstrapping is used to find empirical distribution of the parameters based on which a distribution of confidence is found. Utilizing this distribution, different interpretations of confidence measure may be provided. Analytic results show how statistical properties of the confidence distribution depend on number of training samples and quality of features. Initial results of the approach on eddy current data will be presented.

This work was supported in part by Electric Power Research Institute (EPRI)

 

ECE-09  Trending Videos: Measurement And Analysis

Authors: Iman Barjasteh; Ying Liu; Hayder Radha

Abstract: Unlike popular videos, which would have already achieved high viewership numbers by the time they are declared popular, YouTube trending videos represent content that targets viewers attention over a relatively short time, and has the potential of becoming popular. Despite their importance and visibility, YouTube trending videos have not been studied or analyzed thoroughly. In this paper, we present our findings for measuring, analyzing, and comparing key aspects of YouTube trending videos. Our study is based on collecting and monitoring high-resolution time-series of the viewership and related statistics of more than 8,000 YouTube videos over an aggregate period of nine months. Since trending videos are declared as such just several hours after they are uploaded, we are able to analyze trending videos time-series across critical and sufficiently-long durations of their lifecycle. In addition, we analyze the profile of users who upload trending videos, to potentially identify the role that these users profile plays in getting their uploaded videos trending. Furthermore, we conduct a directional-relationship analysis among all pairs of trending videos time-series that we have monitored. We employ Granger Causality (GC) with significance testing to conduct this analysis. Unlike traditional correlation measures, our directional-relationship analysis provides a deeper insight onto the viewership pattern of different categories of trending videos. Trending videos and their channels have clear distinct statistical attributes when compared to other YouTube content that has not been labeled as trending. Our results also reveal a highly asymmetric directional-relationship among different categories of trending videos. Our directionality analysis also shows a clear pattern of viewership toward popular categories, whereas some categories tend to be isolated.

 

ECE-10  Application Of Single Crystal Diamond For Swift-Heavy Ion Beam Detector

Authors: Ayan Bhattacharya; Andreas Stolz; Timothy A. Grotjohn

Abstract: Diamond is an exceptional material in terms of mechanical, electrical and optical properties. Single crystal diamond has a large bandgap (~ 5.47 eV), large atomic displacement energy (~ 43eV) and high electric breakdown field (107 V cm-1). These superior properties make diamond inherently radiation tolerant and an excellent swift heavy ion beam detector material. In this study, detection performance of single crystal diamond is studied with the long-term objective of understanding the detector degradation process and it`s lifetime in swift heavy ion beam. Single crystal diamonds grown by microwave assisted chemical vapor deposition (MPACVD) at Michigan State University (MSU), are used to develop swift heavy ion beam detectors. Material properties of the diamonds are characterized by birefringence, UV-VIS spectroscopy and FTIR ( Fourier Transform Infrared Spectroscopy). The detectors performance were studied by irradiating the samples with swift heavy ion beams in the range of 100-150 MeV/u at National Superconducting Cyclotron Laboratory (NSCL) at MSU. Besides the MSU grown samples, commercial electronic grade samples were also tested under the same radiation environment. Post irradiation, all samples are characterized by transient current technique and leakage current measurement to understand the degradation mechanism.

This work was supported in part by Strategic Parternership Grant, MSU and MSU Foundation

 

ECE-11  Frequency Reconfigurable Patch Antenna Array

Authors: Jennifer A. Byford;Kyoung Youl Park; Premjeet Chahal

Abstract: A frequency reconfigurable antenna array is introduced. The array consists of planar radiating elements that may be interconnected through computer-controlled switches to form larger elements radiating at frequencies lower than the resonant frequency of the individual elements. Simulations using a 2 x 2 array of patch antennas connected using either hard shorts or RF shorts are used to establish the viability of the concept. Hard shorts are created by adding copper strips between elements. Capacitors are placed across gaps between the copper strips and the patches to create RF shorts as a model for the switches used in the proposed implementation. The simulated arrays were fabricated and the measured properties of the arrays are in good agreement with the predicted results. Both of the shorting methods resulted in a reduction of the resonance frequency to half that of an individual element, while the radiation pattern of the individual element was maintained.

 

ECE-12  Fully Printed, Ultra-Flexible Logic Gates With Tunable Performance Based On Carbon Nanotubes

Authors: Le Cai; Suoming Zhang; Jinshui Miao; Qinqin Wei; Chuan Wang

Abstract: Ink-jet printing is a promising route towards unconventional electronic devices that feature large area, low cost, high flexibility and non-planar structure. Here, we report fully printed logic gates, including inverters, NOR and NAND gates, based on thin-film transistors (TFTs) using semiconducting-enriched single walled-carbon nanotubes (sSWCNTs) as channel materials. Inks of silver and barium titanate nanoparticles were printed as electrodes and gate dielectrics, respectively. The TFTs exhibit better performance and stability than organic semiconductors which are widely used for printed electronics. In addition, the characteristics of the resistor-loaded logic gates can be tuned by printing progressively increasing layers of sSWCNTs. The optimized logic gates show excellent output characteristics as well as superior flexibility. This study represents the first demonstration of fully printed ultra-flexible logic gates using carbon nanotubes and an inspiring step towards high performance electronics through low cost and scalable processes.

 

ECE-13  Stabilization Of Pitch And Yaw Angles Of A Gliding Robotic Fish Using Sliding Mode Control

Authors: Maria Castano; Xiaobo Tan

Abstract: Oceanic sustainability has been a growing global concern due to the increase of potential threats to the integrity of aquatic ecosystems. As a result more attention has been payed to the monitoring of such environments, leading to the need for autonomous aquatic robots that are capable of monitoring them in an efficient and accurate manner. A gliding robotic fish is a type of underwater robot that stems from combining the energy efficient underwater glider with the high maneuverable robotic fish. Because low energy consumption and monitoring efficiency are highly desirable, stabilization of both pitch and yaw during gliding is of great importance. We thus focus on developing a combined sliding mode controller that stabilizes both yaw and pitch simultaneously during gliding. First,an individual controller for the pitch was developed, followed by a combined controller for the stabilization of pitch and yaw. To demonstrate the validity of the proposed controller design we proceeded to implement both simulations and experiments.

 

ECE-14  3-D High Speed Infrared Camera Using Spatial Light Modulator

Authors: Liangliang Chen; Zhanxin Zhou; Ning Xi; Bo Song; Yongliang Yang; Zhiyong Sun

Abstract: In this poster, it is proposing a 3D IR imaging system that will use a carbon nanotube (CNT) based detector along with compressive sensing algorithms that will be integrated together to produce a portable, non-cryogenic-cooled 3D IR camera that will provide quality 3D images at video frame rates. The proposed system will solve several issues currently affecting 3D IR imagers, including their size, computational requirements, bulky cooling systems and slow image collection speeds. These problems make most existing systems unusable in battlefield situations. Preliminary results show that compressive sensing algorithms can dramatically decrease the amount of computation necessary to provide video-rate images and detectors based on CNTs are capable of detecting IR light at much faster speeds than currently possible without large cryogenic cooling systems. These two subsystems can therefore be combined into a portable, low cost, high speed and high quality 3D IR video system that can be integrated into infrastructure as well as many other military and civilian applications.

 

ECE-15  Data Correlation Approach For Slippage Detection In Robotic Manipulations Using Tactile Array Sensor

Authors: Yu Cheng; Chengzhi Su; Yunyi Jia; Ning Xi

Abstract: Two techniques have been presented for slippage detection. They are independent of sensor signal type and are promising for general use on tactile array sensors. The first method is based on frequency analysis of the correlation coefficient sequence of sensor array data sampled as time evolves. The main idea is that a slippage causes heavier fluctuation in the sensor signal distribution and values than a static case. The second approach employs 2-D cross correlation to detect displacements of the sliding object from tactile images, which makes it possible to estimate the slippage velocity based on commercial sensors rather than custom hardware. Experiments have been implemented to evaluate the proposed approaches. It can be seen that the first method is capable of detecting both translational and rotational slippage. Also, it works well in dynamic environments. Besides, the ability of the second method to detect slippage velocity has been confirmed in the experiment.

This work was supported in part by National Science Foundation

 

ECE-16  Higher Order (In Time) Stable PWTD-Accelerated Time Domain Integral Equation Solver

Authors: Zane Crawford; Andrew J. Pray; Shanker Balasubramaniam; John Albrecht; Leo Kempel

Abstract: Integral equation based methods for transient analysis have seen a rapid growth in computational electromagnetics over the past decade. The fundamental bottlenecks are well understood and there are prescriptions to their resolution. The first bottleneck is computational complexity. The cost in degrees of freedom have been reduced through methods such as the plane wave time domain method (PWTD) and the time domain adaptive integral method (TDAIM). The second bottleneck is late time stability. Papers attempting to address this issue have been ongoing since the 1960’s. It is only recently that some resolution has been brought to this issue. However, integration of fast methods with late time stable techniques is still a challenge. In this work, we will work to integrate late time stable methods developed by our group with the PWTD method. While this was a work that begun last year, incorporating PWTD into the solution system brings unexpected results. The PWTD algorithm yields the derivative of the field as opposed to the field. Simple numerical integration yields results that have a slowly growing late time instability. We seek a resolution to this problem by developing accurate integration schemes. To do so, we explore both Runge-Kutta (explicit and implicit) methods, as well as variational based integrators. The results of these methods and their influence on stability will be presented.

 

ECE-17  Interpolation Methods For Economical Pulse-Echo Signal Synthesis

Authors: Pedro C. Nariyoshi; Robert J. McGough

Abstract: Models for pulse-echo simulations use a distribution of point scatterers to generate realistic computer phantoms. In these simulations, the contribution from an individual transducer element evaluated at a single scatterer produces the same signal on each A-line with different time delays and amplitude scale factors. Instead of re-calculating the pressure signals for each A-line, a much more efficient approach delays and scales the pressure signals before superposing the contributions from different transducer elements to obtain each simulated A-line. To delay these signals, appropriate interpolation schemes are necessary. New routines that are under development in FOCUS (http://www.egr.msu.edu/~fultras-web) implement cubic-splines using the exact arrival and edge times. Non-uniform sampling, specifically for the first and last sample, allows better representation of the start and end of the signal where the slope is discontinuous. Simulations are performed using 24 element sub-apertures in a linear array consisting of 192 rectangular elements that are 5mm high and 0.5133mm wide with 0.1mm center-to-center spacing. The simulation is evaluated for a computer phantom with 100,000 scatterers. The center frequency of the excitation is 3MHz. The intermediate signals from the transmit and receive apertures and the pulse-echo signals are shown for each simulation step and compared to a reference signal evaluated at 1GHz. The same configuration is used in Field II (http://field-ii.dk) to compare the computation time and accuracy for the two programs.

This work was supported in part by NIH Grant R01-EB012079

 

ECE-18  Soft Bending Actuators: Explorations Of Fabrication Techniques

Authors: Thassyo Pinto; Xiaobo Tan

Abstract: Soft robotics is a recent trend in engineering that seeks to create machines that are soft, compliant, and capable of withstanding damage, wear and stress. Soft actuators are a major element of soft robots, and they can be achieved through different methods such as soft lithography, photopatterning, and 3D printing. In this work, we computationally analyzed soft pneumatic bending actuators, and developed several prototypes using silicone molding and 3D printing processes, in order to evaluate different materials properties and their performance when subject to various values of pressure. In particular, we were able to fabricate a miniaturized prototype of a stingray-like soft robot for testing a tethered actuation system in an underwater environment.

This work was supported in part by CAPES (Brazil), BEACON (USA)

 

ECE-19  Representative Selection For Big Data Via Sparse Graph And Geodesic Grassmann Manifold Distance

Authors: Chinh Dang; Hayder Radha

Abstract: This paper addresses the problem of identifying a very small subset of data points that belong to a significantly larger massive dataset (i.e., Big Data). The small number of selected data points must adequately represent and faithfully characterize the massive Big Data. Such identification process is known as representative selection [19]. We propose a novel representative selection framework by generating an l1 norm sparse graph for a given Big-Data dataset. The Big Data is partitioned recursively into clusters using a spectral clustering algorithm on the generated sparse graph.We consider each cluster as one point in a Grassmann manifold, and measure the geodesic distance among these points. The distances are further analyzed using a min-max algorithm to extract an optimal subset of clusters. Finally, by considering a sparse sub-graph of each selected cluster, we detect a representative using principal component centrality [11]. We refer to the proposed representative selection framework as a Sparse Graph and Grassmann Manifold (SGGM) based approach. To validate the proposed SGGM framework, we apply it onto the problem of video summarization where only few video frames, known as key frames, are selected among a much longer video sequence. A comparison of the results obtained by the proposed algorithm with the ground truth, which is agreed by multiple human judges, and with some state-of-the-art methods clearly indicates the viability of the SGGM framework.

This work was supported in part by National Science Foundation

 

ECE-20  Application Of Computational Intelligence In Reliability Improvement Of Active Distribution Systems Through Feeder Reconfiguration

Authors: Salem Elsaiah; Joydeep Mitra

Abstract: This study describes a method for reliability improvement of active distribution systems using intelligent and heuristic methods. The work presented here is developed based on a linearized network model in the form of DC power flow and linear programming model, in which the thermal capacities of distribution feeders and real power constraints are accounted for. The optimal distribution system architectures are obtained using an intelligent binary particle swarm optimization based search method. The probabilistic reliability assessment is conducted using a method, which is developed based on event tree analysis. In order to enhance the computational burden, the probabilistic reliability assessment method is modified so that the effect of higher-order contingencies is limited using a frequency based cut-off criteria. Several case studies are carried out on a benchmark 33 buses radial distribution system. The results show that the amount of annual unserved energy, number of affected households, and number of affected in-service consumers can all be tremendously reduced using the proposed method.

 

ECE-21  A Step-Ahead On Designing And Teaching Courses According To The ABET Guidelines: Application To ECE 474 Principles Of Electronics Devices

Authors: Salem Elsaiah; Virginia Ayres

Abstract: The new ABET engineering criteria was first formulated and introduced to American education system in the middle 1990s. The new ABET accreditation system defines an engineering graduate according to Blooms Taxonomy [1], in which criterion (a) to (k) should be met. Understanding and implementing a successful ABET engineering program is somewhat complex task and can be confused for several educators and course developers. The vast majority of the work presented in the literature is concerned with the design of learning objectives, course outcomes, assessment tools, etc, however less attention has been paid thus far on how to implement the guidelines of the new ABET engineering criteria. This study presents a step-ahead on how to design and teach engineering courses to meet the ABET guidelines. The work presented here summarizes the design procedure of an ABET syllabus for ECE 474 course. Then, it introduces and implements Demos and simulation packages to help students visualize and understand important concepts, in particular Hall Effect measurements and carrier concentrations. This work finally utilizes a method, which is developed based on Adaptive Backward Curriculum design to assess student performance and to align learning objectives with course outcomes.

 

ECE-22  Cooperative Control For A Network Of Gliding Robotic Fish

Authors: Osama Ennasr; Xiaobo Tan

Abstract: Gliding robotic fish combine the strengths of both underwater gliders and robotic fish, resulting in long duration of operation and high maneuverability. Michigan State’s gliding robotic fish, Grace, has already been deployed in Kalmazoo River and Wintergreen Lake for environmental monitoring. This research focuses on the control of a network of gliding robotic fish that work together to perform specific tasks. Utilizing a network of robots, much more sophisticated missions can be carried out that would improve the quality of our water resources and the understanding of our marine environment. Applications for cooperative control include cooperative tracking of real fish, sampling lakes and rivers at a higher spatial resolution, and monitoring lager lakes and water resources. In many of these applications, it is desirable to have all robots travelling as a group, or following a leader. Dynamic feedback linearization along with distributed linear estimators are used to track the velocity states of a leader. The leader’s information, which is available to only a subset of the network, could be the output of another physical robot or a virtual leader that represents the desired motion of that network. Simulation results are presented to determine the effectiveness of the proposed approach.

This work was supported in part by National Science Foundation

 

ECE-23  Breast Cancer Detection Using Haralick Features Of Images Reconstructed From Ultra Wideband Microwave Scans

Authors: Blair Fleet; Jinyao Yan; David Knoester; Meng Yao; John Deller, Jr.; Erik Goodman

Abstract: Microwave scanning of the breast would provide a technology for cancer detection and screening that is significantly safer than current methods involving radiation. This research focuses on finding the best way for accurate characterization of cancerous signals and normal signals using clinical data collected from a previously developed ultra wideband (UWB) antenna, BRATUMASS (Breast Tumor Microwave Sensor System). BRATUMASS detects changes in dielectric constants within the breast. The signals collected from the microwave scanning procedure are reconstructed into a single, informative representation of the breast via diffraction tomography. This representation contains the information of the breast's conductivity and the change in dielectric constants. We illustrate the feasibility of using Haralick features to make distinctions among breasts with a malignant tumor present and breasts with no malignancy in data collected from Shanghai Sixth People's Hospital and Shanghai First People's Hospital.

This work was supported in part by NSF GRFP (National Science Foundation Graduate Research Fellowship Program); GEM (National Consortium for Graduate Degrees for Minorities in Engineering and Sciences) Fellowship

 

ECE-24  Stepped Waveguide Technique For The Extraction Of Electromagnetic Parameters Of Conductor-Backed Absorbers.

Authors: Jonathan Frasch; Edward Rothwell; Benjamin Crowgey

Abstract: While well-established techniques exist for finding the dielectric and magnetic constants of unknown samples, some of these techniques become unusable when a conductive material is affixed to the back, since the conductor prevents transmission measurements from being taken. One method of overcoming this problem is to use a technique requiring two reflection measurements. In this case, use of two different stepped-waveguides to obtain two independent reflection measurements is examined.


ECE-25  An Adaptive Transfer Function For Deriving The Central Blood Pressure Waveform From A Peripheral Blood Pressure Waveform: Evaluation In Patients

Authors: M. Gao; W. Rose; B. Fetics; H. Cheng; C. Chen; D. Kass; R. Mukkamala

Abstract: Central BP is clinically more relevant than peripheral BP, but peripheral BP is easier to measure. While devices are available to derive the central BP waveform from a peripheral BP waveform, they assume that a single, universal transfer function exists that can be applied to a peripheral BP waveform of any subject over any physiologic condition so as to accurately predict the central BP waveform. We previously proposed to adapt the transfer function relating peripheral BP to central BP to the arterial properties of the subject at the time of measurement by employing a physical model of the arterial tree and exploiting the fact that ascending aortic blood flow is negligible during diastole. In this study, our objective was to evaluate the ATF in patients. We studied 49 patients undergoing cardiac catheterization. We simultaneously measured a peripheral BP waveform for analysis via a radial artery tonometer (N = 39) or radial artery catheter (N = 10) and the reference central BP waveform via a high fidelity ascending aortic catheter. In some patients (N = 14), we obtained the waveforms before and after an intervention that perturbed BP. We applied the ATF to the peripheral BP waveforms and quantitatively assessed the predicted central BP against the reference measurements. We created a GTF using all of the available central and peripheral BP waveforms and another GTF that constituted an essentially perfect model of the commercial SphygmoCor device. The Table shows that the ATF achieved significantly lower central BP RMSEs than both GTFs.

 

ECE-26  Exploration Of Inverse Depth Visual SLAM With Single Camera

Authors: Jason Greenberg; Xiaobo Tan

Abstract: Visual SLAM (Simultaneous Localization And Mapping) is popular in land-based robotic applications, but is severely limited for underwater localization and navigation. This presentation explores a terrestrial-based monocular Visual SLAM algorithm known as the Inverse Depth SLAM with hopes that its unique benefits could be applied to underwater situations. The Inverse Depth SLAM is unique in that it can fully initialize a feature vector with uncertain depth by storing a Gaussian prior of the depth's inverse as well as referencing the point's location from the position of the camera when it was first observed. The feature is then able to be converted to the standard XYZ parametrization after sufficient parallax has been observed. The poster presentation details the basic structure and purpose of this SLAM implementation among with some implementation details.

 

ECE-27  Efficient Parallel Algorithms For Electromagnetic Composite-Object Integral Equations

Authors: Steve Hughey; Shanker Balasubramaniam; Leo C. Kempel

Abstract: Modern engineering challenges require electromagnetic analysis of complex systems comprising composite objects and multiple length scales. Among modern computational methods used to solve these problems, integral equations (IEs) are held in high regard for their accuracy due to their exact imposition of radiation boundary conditions. Fast methods for IEs, such as the multilevel fast multipole algorithm (MLFMA), have been extensively studied, and reduce the cost of solutions from O(N^2) to O(NlogN), where N is the number of degrees of freedom in the problem. Even so, it is still infeasible to solve problems with very large N in a reasonable amount of time without further modifications to the solution algorithm to reduce the computational cost, limiting the utility of these methods for large, multiscale engineering problems. Parallelization of these algorithms has been the thrust of much recent work. However, much of the work on development of efficient parallel algorithms has been devoted to analysis of perfectly conducting objects with more limited work on composite objects. A recently proposed parallel algorithm which we will leverage in this work was shown to be highly scalable up to 512 processors. To this end, we will augment domain partitioning by an intelligent pre-processing step to effect optimal distribution and alleviate communication costs. Our earlier incarnation of this work used a brute force partitioning of the space filling curve, the drawback of which is loss of scalability for high-contrast inclusions. Furthermore, we will address development of efficient preconditioning strategies within the parallel framework.

This work was supported in part by DARPA

 

ECE-28  Analysis Of Instantinous Reactive Power Theory For Active Filtering Of Matrix Converter

Authors: Ameer Jnabai; Bingsen Wang

Abstract: Matrix converter, is a single stage direct AC/AC converter. By employing only nine bidirectional switches it is able to create variable amplitude and frequency output voltages. In this poster, a series and shunt active filters are applied on the matrix converter to minimize the power quality impact. The reference voltage and current in the shunt and series active filters has been obtained using the instantaneous reactive power theory. The system shows a high performance in eliminating the undesired harmonics from the input current and the output voltage waveforms.

 

ECE-29  Closed-Loop Natural Language Control Of Robotic Systems For Exception Handling

Authors: Yunyi Jia; Ning Xi; Joyce Chai, Lanbo She; Yu Cheng

Abstract: Controlling robots by natural language is more intuitive and efficient than using the conventional control manners such as joysticks. Most existing methods on natural language control have designed the systems in an open-loop manner, which cannot handle exceptions in the robotic system and environment. Therefore, we develop a closed-loop control framework for natural language controlled robotic systems with the capability of handling such exceptions. The designed method is implemented on a natural language controlled mobile manipulator system and the experimental results demonstrate its effectiveness and advantages.

 

ECE-30  Metal Insulator Metal (MIM) Based Microwave Circuits On Flex Substrates

Authors: Amanpreet Kaur; Premjeet Chahal

Abstract: This paper demonstrates fabrication and characterization of thin-film Ti-TiO2-Pd based Metal-Insulator-Metal (MIM) diodes on flexible substrates. MIM diodes with different contact areas were fabricated, and a comparison is made for their DC and RF performances. The current-voltage characteristics of the fabricated diodes show strong non-linearity. The diodes are also tested for microwave circuit applications such as detection, frequency multiplication and mixing . Details of DC characteristics, RF rectification, mixing and multiplication using MIM diodes are presented.

This work was supported in part by DARPA

 

ECE-31  A Novel Tubular IPMC Sensor Capable Of Omnidirectional Sensing

Authors: Hong Lei; Montassar Aidi Sharif; Xiaobo Tan

Abstract: An ionic polymer-metal composite (IPMC) sensor typically consists of a thin ion-exchange membrane with electrodes on both surfaces. Such IPMC sensors only respond to stimuli acting perpendicular to the beam plane. In this paper, we present a novel, omnidirectional, tubular IPMC sensor that responds to all stimuli perpendicular to the tube axis. This tubular IPMC has one common inner electrode and four outer electrodes, which form four routes of common-ground current outputs. With a custom-made setup, the response of each sensor route is characterized under tip deflection in different orientations at frequencies 1-20 Hz, which verifies the sensor’s omnidirectional sensing capability and shows little mechanoelectrical coupling between neighboring sensor routes. An analytical dynamic model, in the form of a transfer function, is developed for the sensor, which captures the internal ion-transport physics. Experimental results show that the proposed model is able to capture the tubular sensor dynamics.

This work was supported in part by National Science Foundation: DBI 0939454; the Office of Naval Research: N000140810640, N000141210149;

 

ECE-32  Isogeometric Analysis For Integral Equations In Electromagnetics

Authors: Jie Li; Shanker Balasubramaniam

Abstract: Isogeometric analysis (IGA) has recently become popular in computational science during the past decade or so and it has shown great potential in simplifying analysis-based designing procedures. IGA tries to to unify both geometric and field representation; in other words, both the geometry and the fields are represented using the same underlying basis set. However, while the concept of IGA for differential equations is more common, extension to an integral equation framework is significantly more challenging. In this work, we present for the first time, the IGA as applied to three dimensional integral equations encountered in electromagnetics. The presented approach relies on the subdivision scheme for both geometry and function representation. Due to high smoothness and partition of unity nature in the representation basis, one can construct a rigorous local Helmholtz decomposition of currents defined on the tangential space of an arbitrary surface. Hence the proposed approach will lead to a stable discretization of electric field integral equation ranging from nearly static case to high frequency. Specifically, besides better accuracy, the method presented is a robust solution to the following problems altogether: (1)the low-frequency breakdown, (2) implementation of Calderon preconditioner without using dual mesh, and (3) problems with multi-scale meshes.

 

ECE-33  Ultrashort Channel Length Black Phosphorus Field-Effect Transistors

Authors: Jinshui Miao; Suoming Zhang; Le Cai; Qinqin Wei; Chuan Wang

Abstract: Few-layer black phosphorus (BP) is an emerging two-dimensional material which is of great interest for applications, mainly in high performance electronics, optoelectronics, and chemical sensors et al. BP nanoflake has a moderate bandgap of around 0.3 eV and high carrier mobility, which lead to transistors having high on-current densities and on-off ratios. Here, we demonstrate the ultrashort channel length BP field-effect transistors (FETs) with channel length of around 20 nm. The BP-FETs show respectable current saturation with an on-off ratio that exceeds 100. We achieved a current density in excess of 170 µA/µm at Vds=0.1 V. Unencapsulated BP nanoflakes are found to chemically degrade upon exposure to ambient conditions. In this study, ultrashort channel length BP-FETs with atomic layer deposited Al2O3 layer can effectively suppress ambient degradation, allowing encapsulated BP-FETs to maintain high on-off ratio and on-current density.

 

ECE-34  Affordable 3D Printed Microwave Antennas

Authors: Mohd Ifwat Mohd Ghazali; Eleazar Gutierrez; Joshua Myers; Amanpreet Kaur; Brian Wright; Premjeet Chahal

Abstract: A variety of 3D printed microwave antennas are presented including wide band, narrow band, multiband and reconfigurable designs. In particular, single layer patch, folded E-patch, a bilateral Vivaldi, Spartan logo and Lego-like assembled antennas are demonstrated. 3D printing provides significant flexibility in the design of antennas that combine the assembly of both dielectric and metal layers to achieve desired performance characteristics such as resonant frequency and radiation pattern. Also, small Lego-like blocks can be printed that allows in the design and assembly of novel antennas structures using a combination of dielectric and metal coated blocks.

 

ECE-35  Nonlinear Identification Of Total Baroreflex Arc

Authors: Mohsen Moslehpour; Toru Kawada; Kenji Sunagawa; Masaru Sugimachi; Ramakrishna Mukkamala

Abstract: A second-order nonlinear model of the total baroreflex arc –the open-loop system relating carotid sinus pressure (CSP) to arterial pressure (AP) – of rats was established using Gaussian white noise stimulation and nonparametric identification. The nonlinear model was able to predict AP 15% better than the conventional linear model.

 

ECE-36  Irradiation Of Single Crystal Diamond For Lift-Off Processes Using Proton And Carbon Beams

Authors: M. Muehle; A. Kayani; M.F. Becker; S.J. Wickramarachchi; T. Schuelke; J. Asmussen

Abstract: A challenge for homoepitaxially growing larger single crystalline diamonds (SCD) is the separation of seed crystals after the microwave plasma assisted chemical vapor deposition (MPACVD) process. Conventional laser cutting leads to substantial material loss with increasing lateral sizes due to the cutting wedge profile. A possible solution to minimize material loss is offered by lift-off processes that create a thin subsurface damage layer through ion implantation. Proton and carbon ion beam interactions with SCD were modeled with Monte Carlo simulations. Proton energies of 300 – 700 keV result in practical penetration depths of a few μm. Similar results are achieved with 3 MeV carbon beams. The minimum doses to exceed this damage threshold were estimated to be 6.3 × 1016 cm-2 for 700 keV protons at 4.7 µm and 1.1 × 1015 cm-2 for 3 MeV carbon ions at 1.6 µm penetration depth. Ion beam irradiation experiments were carried out using a 6 MeV tandem Van De Graaff accelerator. SCDs have been irradiated with protons. Doses varied from 5 × 1015 cm-2 to above 3 × 1017 cm-2. Crystals exposed to doses exceeding 1 × 1016 cm-2 showed a grey coloration after the ion beam exposure. MPACVD deposition was performed on these samples with substrate temperatures exceeding 900 °C. After deposition the damage layer appears as a thin but incomplete line. The results show that proton irradiation can be used to create graphitization deep inside of SCD crystals though the required minimum dose of 3 × 1017 cm-2 is high.

 

ECE-37  Differential Geometry Based Co-Ordinate Tranformation For Eddy Current Modelling

Authors: Saptarshi Mukherjee; Antonello Tamburrino; Satish Udpa; Lalita Udpa

Abstract: U-bend sections are common and critical part of the tubes in steam generators of nuclear power plants. It is difficult to numerically compute the physical fields on such complicated geometries since the finite element solution is extremely sensitive to the numerical errors. It is also difficult to experimentally check for defects in these sections due to the requirement of complicated motion of the eddy current coil. Hence, reliable crack detection largely depends upon accurate numerical modeling of the U-bend region.This poster presents a numerical model of simulating eddy current distribution within such a U-bend region using coordinate transformation. A differential geometry based co-ordinate transformation has been employed to investigate eddy current testing on a complicated shaped geometry (U-Bend). The geometries are considered to be invariant along the normal axis, and hence can b e interpreted as a 2-D problem. The standard A-V formulation has been derived for the transformed basis, and the new metric is used to determine the new constitutive relationships. The similarity between results of the real and transformed basis confirms that the formulation is capable of predicting the signatures of real complicated situations.

 

ECE-38  Evolutionary Optimization Of Multi-Band Quasi-Optical THz Thin-Film Bandstop Filters

Authors: Joshua C. Myers; Amanpreet Kaur; Premjeet Chahal

Abstract: Utilization of the terahertz (THz) spectral region is growing in importance for a host of applications. These applications include broadband communications, spectroscopy, non-destructive evaluation, biomedical imaging, and security [1]. In order to successfully implement THz systems, low-cost passive elements are necessary. Compared with components in the microwave and optics spectral regions, THz passive elements are not readily available and are difficult to design. Among the many THz passive components, quasi-optical bandpass and bandstop filters are necessary in a variety of THz applications. A variety of methods have been used to design these filter, but recently interest has grown in using metamaterial structures in the THz spectrum because of inherent design benefits. Metamaterial based filters can be designed to achieve a large quality factor (Q) and out of band rejection, as well as being low-cost and compatible with fabrication on flex substrates. However, to achieve complex filter designs, such as wideband and multiband filters, multiple filter layers are often required [2]-[3]. Adding multiple layers of metamaterial structures increases design and fabrication complexity, as well as the size of the filter. The additional layers can also impact the Q-factor as well as the bandwidth of the filter due to the loading between elements. In this paper, single-layer metamaterial based THz bandpass and bandstop filter designs are presented. The filters are constructed with a single metamaterial unit cell, which can then be expanded into an array to be used as a filter. The geometry of each unit cell is optimized with a binary genetic algorithm (GA). The GA is implemented with a 3D full-wave EM solver, HFSS, by developing a HFSS-MATLAB interface. The objective of the GA is to maximize a selected filter parameter, such as the filter Q, bandwidth, number of pass/stop bands, or resonance locations. Multiple filter designs will be studied showing a tradeoff between the desired optimized parameters mentioned above. The filters will be fabricated on flex dielectric substrates using simple cleanroom fabrication methods. Measurement and testing of a wide arrangement of filters will also be investigated using a frequency domain THz system to characterize their individual resonance spectrums.

 

ECE-39  Growth And Characterization Of High Quality, Large Single Crystal Diamond Substrates (SCDs)

Authors: Shreya Nad; Yajun Gu; Jes Asmussen

Abstract: Exciting properties of diamond like extreme hardness, chemical inertness and wide band gap can be exploited in high power electronics and high power optical components. The main obstacle has been the synthesis of large (>1cm^2), high purity and defect-free SCDs to overcome the problem of grain boundaries arising in polycrystalline diamond (PCD) substrates. Here, we investigate methods of increasing the area of SCD during synthesis. It is important to mitigate the growth of the PCD rim which reduces the required SCD surface area. We have conducted experiments using several variations of open and pocket holders. By adjusting the dimensions of the pocket holder, the PCD rim can be reduced by a factor of 2 or more in comparison to open holders. Designing novel substrate holders, which allow the growth of diamond in both longitudinal and lateral directions, will help in obtaining thick large SCD substrates. We have grown high quality (< 300ppb [N] impurities), thick (~2.2mm) SCDs by using pocket holders. High growth rates of 25–32μm/h were achieved. These substrates have been synthesized via microwave plasma assisted CVD in a microwave cavity plasma reactor whose operating conditions have been optimized for an efficient process. Threaded dislocations arising from the HPHT seeds propagate parallel to growth direction <100>. After removing the HPHT seed and by flipping the grown CVD SCD by 90° for another synthesis step, the propagation of the dislocations can be minimized. By subsequent flipping of substrates, it is possible to obtain high quality, large type IIa SCDs.

This work was supported in part by II-VI Block Gift Program

 

ECE-40  Self-Structuring Antennas For Phased Arrays

Authors: Christopher Oakley; Matthew Hodek; Lee Harle

Abstract: A self-structuring antenna (SSA) has been developed to enable scalable and rapidly upgradable phased array systems over broad tunable bandwidths. The SSAs are of a patch antenna design, utilizing shorting pins distributed throughout the patch face to control current distribution through switch activation. This poster describes the SSA concept, the design approach for this application, and presents measured results of two prototype designs.

This work was supported in part by Raytheon Company and the Defense Advanced Research Projects Agency

 

ECE-41  Longitudinal Groove Loading Of A Half-Width Microstrip Leaky-Wave Antenna

Authors: Korede Oladimeji; Jonathan Frasch; Edward Rothwell; Leo Kempel

Abstract: Their utilization in applications requiring high directivity and broadband systems and ability to scan with frequency are some of the advantages provided by leaky-wave antennas. Half-width leaky-wave microstrip antennas have been studied extensively because of the simplicity of their feeding mechanisms and the ease of fabricating them. To date, they have found applications in communications systems and direction-finding. Beam control of such half-width leaky-wave antennas can be typically achieved by reactively loading the edge of the microstrip. Various methods have been utilized in implementing this loading including lumped capacitors, varactors, and microstrip stubs. An alternative to these loading methods is proposed here which involves the inclusion of a longitudinal groove next to the shorting wall of the antenna underneath the microstrip and within the substrate. In this case, the desired beam-scanning is achieved by modifying the dimensions of the longitudinal groove. The possibility of two scanning methods is explored. First, frequency tuning while the principle beam-angle is kept constant is investigated. Also, fixed frequency tuning of the principle beam-angle is considered. The inclusion of the longitudinal groove is ultimately cheaper than other currently used methods and potentially easier to fabricate. The performance of the antenna will be demonstrated in the two scenarios mentioned. Simulations showing the impedance and radiation properties, as well as the angular extent of the scanning of the antenna will be presented. Experimental results with fabricated antennas will also be explored for different scanning scenarios.

 

ECE-42  Graph Wavelet Transform: Application To Image Segmentation

Authors: Alp Ozdemir; Selin Aviyente

Abstract: Recently, there has been a lot of work on extending traditional signal processing methods to irregular domains such as graphs. Graph wavelet transform offers a multiscale analysis of graphs similar to traditional wavelets. Similar to wavelets which are effective at detecting transients in a signal, graph wavelets can be used to detect discontinuities of functions defined on graphs. In this paper, we use this realization to propose a graph wavelet based approach to image segmentation. The images are first transformed to the graph domain and the graph wavelet transform is used to detect the discontinuities in the pixel domain.

 

ECE-43  Sensing Of Torsion Based On A Tubular IPMC Sensor

Authors: Montassar Aidi Sharif; Hong Lei; Mohammed Al-rubaiai; Xiaobo Tan

Abstract: Typical ionic polymer-metal-composite (IPMC) sensor takes the shape of a beam and is able to measure stimuli that bend the beam. In this project, sensing of torsion based on a tubular, thin-wall IPMC sensor is presented. The fabrication process the tubular IPMC is explained in detail. An experimental setup which provides rotational movement is used to characterize the sensor response to angular acceleration. The experimental results have showed that the response of the sensor is similar in clockwise and counterclockwise rotations for a specific acceleration. Finite element analysis is conducted to understand the sensing mechanism.

 

ECE-44  Mathematical Modeling Of A Screw Propeller-Based Underwater Robot

Authors: Pratap Bhanu Solanki; Xiaobo Tan

Abstract: Accurate mathematical modeling of underwater vehicles is critical for navigation and control purposes. In this work a dynamic model is derived for a propeller-based underwater robot. A low- cost 10DOF IMU (10 Degree of Freedom, Inertial Measurement Unit) is integrated with the robot and calibrated. The IMU is later used in the robot as an inertial sensor to estimate the state of the robotic system. The parameters of the model can be further tuned by comparing the predicted states of the system with the actual trajectory the robot follows.

 

ECE-45  Topography And Mechanical Properties Measurement Using Nanorobot Working In Vibration Mode

Authors: Bo Song; Ning Xi; Zhiyong Sun; Yongliang Yang; Liangliang Chen; Zhanxin Zhou

Abstract: During recent several decades, atomic force microscopy (AFM) plays an important role of providing accurate topography image and near-filed mechanical properties measurement in the nanoscale. AFM enables the biomedical studies have a tremendous development from bull experiment to single molecule studies. Mechanical properties of single cell have been known as an important indicator or reporter to estimate or predict the state of the cells. Traditionally, the Mechanical properties are measured by using “force curve” nano-indentation method that works in the “contact mode” which may damage the sample surface. In addition, it is difficult to obtain the accurate topography information while measuring the mechanical properties, which makes less meaningful to the mechanical properties measurement. In order to address these issues, we present a method called “vibration mode” for the study of internal mechanical properties of elastic samples. The vibration mode has a combination of the advantage of conventional tapping mode imaging with nanoscale resolution, and the nanomechanical properties mapping using the additional deformation induced by vibrating near the resonance frequency of the measured sample. The images of mechanical properties map and the topography information can be obtained simultaneously. One of the significances of this technology is the non-invasive measurement. The AFM tip is not mechanically contacted with the sample surface, which assure the accuracy of the measurement by means of ruling out the affection by the tip-surface interactions which are the biggest issue for the current nano-indentation measurement.

This work was supported in part by NSF Grant 1301691, NIH Grant R43 GM084520

 

ECE-46  Time Domain Monotonicity Based Inversion Method For Eddy Current Tomography

Authors: Zhiyi Su; Antonello Tamburrino; Salvatore Ventre; Lalita Udpa; Satish Udpa

Abstract: Monotonicity based methods are a class of newly developed non-iterative inversion methods for addressing inverse problems related to partial differential equations (PDE). This poster proposes the monotonicity method in the time domain for pulsed eddy current data inversion. A primary inversion strategy as well as numerical examples are demonstrated in the poster. This method is suitable for real-time imaging.

 

ECE-47  A Computationally Efficient Measure Of Multivariate Phase Synchrony Based On Hyper-Torus Synchronization

Authors: Marisel Villafañe-Delgado, Ali Yener Mutlu; Mahmood Al-khassaweneh; Jason Moser; Selin   Aviyente

Abstract: Cognitive and sensory processes rely on the coordinated neural activity of multiple regions in the brain. The integration of regions involved in such processes has been mostly assessed by means of bivariate phase synchrony. However, bivariate synchrony presents various drawbacks in the assessment of large-scale synchronization. In particular, it is an indirect measure of the global connectivity in a network, is computationally expensive for large networks and difficult to interpret for large connectivity matrices. Multivariate phase synchrony measures have been proposed as an alternative to bivariate measures. However, these metrics rely on the synchronization matrix constructed from bivariate synchrony measures, on the average of phases or cluster-based methods which are computationally expensive or are not reliable under noise. This work proposes a novel multivariate phase synchrony measure based on a hyper-torus representation of multiple oscillators. The proposed method is compared to a widely implemented multivariate technique, the S-estimator, and result to be less computationally expensive, more robust to noise and exhibits higher synchrony among the central-frontal electrodes in an Event-related negativity experiment.

 

ECE-48  Nonlinear Biomedical System Identification Using A Set-Theoretic Evolutionary Approach

Authors: Jinyao Yan; John Deller; Erik Goodman

Abstract: Biomedical systems are complex and are generally nonlinear and time-variant. The modeling of biomedical systems therefore presents significant challenges that are not overcome by the classical linear methods. In recent decades, arduous research has begun to produce methods for analyzing and modeling isolated classes of nonlinear systems, but this vast class of models still presents many challenges, especially in complex biomedical systems. This application proposes a novel method for nonlinear system identification. The approach integrates three modeling and identification strategies: linear-time-invariant-in-parameters models, set-based parameter identification, and evolutionary algorithms for optimization over fitness measures derived from the set solutions. While conventional model identification has focused on the estimation of parameters, the proposed method simultaneously addresses both selection of the model and parameter estimation. The proposed algorithm can identify biomedical system models in different unknown noise scenarios, especially in correlated or multiplicative noise. An medical application is currently being analyzed using this algorithm: epileptic seizure prediction and detection. The proposed evolutionary model identification has significant potential for further theoretical development. Furthermore, the method significantly generalizes model possibilities, with wide applicability to nonlinear filtering, detection, and estimation, and to pathological state feature extraction and identification, for an array of biomedical modeling problems.

This work was supported in part by U.S. National Science Foundation under Cooperative Agreement DBI-0939454

 

ECE-49  Diamond Schottky Barrier Diodes For High Power, High Temperature Electronics

Authors: Stephen Zajac; Timothy Grotjohn

Abstract: Diamond semiconductors have the potential to revolutionize electronics. Compared to silicon based semiconductors, the breakdown electric field strength is 30 times higher, and the thermal conductivity is 10 times higher. These two properties of diamond enable high power, high temperature applications that are impossible with silicon based semiconductors. Schottky barrier diodes are fabricated using synthetic diamond grown by chemical vapor deposition. Measurements of current/voltage characteristics are taken between room temperature and 300 degrees Celsius. Reverse bias breakdown voltages of 700 Volts, and forward bias currents of 1 Amp are achieved.

This work was supported in part by ARPA-E

 

ECE-50  Compressive Sensing-Based Preisach Hysteresis Model Identification

Authors: Jun Zhang; David Torres; Nelson Sepulveda; Xiaobo Tan

Abstract: The Preisach hysteresis operator has been adopted extensively in the modeling of magnetic and smart material-based systems. In this poster we present a novel compressive sensing-based approach to identify the Preisach operator that requires much fewer measurements. Fidelity of the Preisach operator hinges on accurate identification of the Preisach density function on the Preisach plane. Existing work on the identification of density function usually involves applying an input that contains sufficient excitation and measuring a large set of output data. The proposed compressive sensing-based Preisach model identification approach would require much fewer measurements. Compressive sensing is an alternative to Nyquist-Shannon sampling theory for acquisition and reconstruction of sparse signals. According to the compressive sensing theory, any length-N signal that can be well approximated using K coefficients can be faithfully recovered from M =O(Klog(N/K)) random linear projections of the signal. The density function is transformed into the frequency domain, generating a sparse signal of discrete cosine transform (DCT) coefficients, which can be efficiently reconstructed using compressive sensing algorithms. The root-mean-square error (RMSE) and the maximum absolute error are adopted to examine the density function reconstruction capability and the model estimation performance. The effectiveness of the proposed scheme is first illustrated through simulation results. Being an interesting class of smart materials, vanadium dioxide (VO2) undergoes a thermally induced solid-to-solid phase transition, during which the electrical, mechanical and optical properties change in a hysteretic way. The effectiveness of the proposed identification approach is further verified involving the hysteresis between the voltage input and the deflection output of a VO2 integrated microactuator.

This work was supported in part by National Science Foundation (CMMI 1301243)

 

ECE-51  Bolometric Effect Application For The TFT Based On Single Chirality Carbon Nanotubes

Authors: Suoming Zhang; Le Cai; Tongyu Wang; Jinshui Miao; Qinqin Wei; Nelson Sepúlveda; Chuan Wang

Abstract: This paper reports the bolometric effect on thin film transistor using sorted 99% semiconducting Single Walled Carbon Nanotubes (SWNTs) and sorted (9,8) SWNTs as channel layer material. We observed when the laser is illuminating the channel region of the transistor, the drain current would increase under very positive gate voltage while the drain current would decrease dramatically under very negative gate voltage. Furthermore, we found that the transistor showed different thermal response upon different laser intensity, which following the trend of the absorption spectrum of semiconducting SWNTs and (9,8) SWNTs. That’s to say, our transistor is wavelength-selective.

 

ECE-52  Super Resolution Infrared Light Field Imaging Using Single Carbon Nanotube Detector

Authors: Liangliang Chen; Ning Xi; Zhanxin Zhou; Ruiguo Yang; Bo Song; Zhiyong Sun

Abstract: Computational photography which enhances or extends the capabilities of digital imaging is one of most rapidly developing research field in computer vision, image processing and applied optics. The output of these techniques can reconstruct information of scene which is not obtained by today's traditional camera. It also opens a new approach for light field sensing. In this work, by analyzing optics of single pixel imaging system, we present a novel design for single pixel light field camera, which allows for capturing high resolution light field by carbon nanotube based infrared photodetector. The spatial light modulator reflects objects and forms a virtual image behind the plane in which the mirror lies. It consists of millions micro scale mirror which works as CCD array in camera and it is controlled separately so as to project linear combination of object image on lens array. The carbon nanotube based infrared detector, which has ultra high signal to noise ratio and fast responsibility, will sum up all image information on it, without image deformation. The proposed method will sample multiple angular images through different pin holes so that more information is captured. Then, a robust reconstruction algorithm is applied to recover high resolution image. The performance of high resolution single pixel light field camera can achieve million pixels image from angular images.