The current issue of the IETE Journal of Research (Vol.62, No.5, September–October 2016) contains 20 articles. These articles present on-going research and developments taking place in the broad areas of Signal Processing, Electromagnetics, Microelectronics, Communication and Biomedical Engineering.
The paper titled ‘Dynamic pronunciation modeling for unsupervised learning of ASR systems’ presents a method for adapting pronunciation variability found in everyday speech using unsupervised adaptation. It uses dynamic phone warping, accent and word distance estimation and an unsupervised adaptation algorithm that enrolls new accents and out of vocabulary words. The results are given in terms of specific word examples and word error rate.
The next paper on ‘Wide-scan phased antenna array based on cavity-backed ELC-slot microstrip patch antenna element’, describes the design of wide-scan phased array antenna using meta-material electrical slot and L-probe fed cavity backed patch antenna element in the X-band. The impedance matching performance, scanning and other performance of the design are described using simulation and experiments.
In the paper ‘Low SAR PIFA antenna for wideband applications’, the authors present the design of a low SAR printed PIFA antenna for wideband applications relevant for GPS, DCS, GSM1800, PCS, WLAN, Bluetooth and LTE standards. The performance of the antenna is given in terms of VSWR, gain, and efficiency based on FEM and TDM studies in simulation software.
The following paper titled ‘An effective contrast enhancement method for identification of microaneurysms at early stage’, presents a contrast enhancement method for retinal image to help improve the detection of micro-aneurysms that is useful in the early detection of diabetic retinopathy. The method is based on sub-histogram intensity equalization and brightness difference equalization between input and output images. The method is tested on datasets and improvement in performance is reported with the use of the contrast enhancement method.
The paper on ‘Efficient voting-based extractive automatic text summarization using prominent feature set’, presents the study of voting based extraction approaches for text summarization. Voting based methods are sensitive to the initial ranking process for which a reciprocal ranking approach is given. The summarization method is evaluated on standard database using evaluation matrices and compared with other voting based methods.
In ‘Relevance feedback-based query expansion model using ranks combining and Word2Vec approach’, the improvement of document retrieval performance using individual query expansion term based selection methods is studied. Two co-occurrence based measures and statistics based approaches for expansion term selections for pseudo relevance feedback based automatic query expansion are presented. A semantic similarity approach named Word2Vec for removing the irrelevant and redundant expansion terms is presented. The model is tested on two benchmark datasets and compared with other models from the literature.
The following paper on ‘Different scenarios on de-noising of signals in the intrinsic mode function selection framework’, uses the framework of empirical mode decomposition (EMD) and five different noise suppression conditions based on intrinsic mode functions (IMFs) that result from EMD. The amount of noise in and signal information in the IMFs vary. Five conditions are studied and the de-noising is evaluated with simulated and real signals in additive white Gaussian noise.
The next paper on ‘Clique-based network coding technique’, presents an algorithm for network coding that addresses the problems of connected dominating set network coding. The algorithm is compared with both connected dominating set network coding and tree network coding approaches in terms of number of transmissions and number of coding operations. The throughput, coding opportunity, coding gain and robustness towards packet loss are compared with the two network coding approaches.
The paper titled ‘OP performance and power allocation for DF relaying M2M cooperative system’, presents the outage probability performance for decode-and-forward relaying mobile-to-mobile system with transmit antenna selection (TAS) over N-Nakagami fading channels. The performance is evaluated analytically and verified through numerical simulations. Based on the study, conclusions are drawn regarding the outage probability performance of TAS scheme.
The paper on ‘Trust and location-aware routing protocol for wireless sensor networks’, gives a protocol for cooperative location aware routing in wireless sensor networks. The Trust and Location aware Routing protocol consists of two parts – trust assessment and routing. The trust module dynamically adjusts weights associated with trust metrics, while the routing decisions use consolidated trust values and geographic routing principles. The performance of the protocols is given using NS-2 simulation studies and inferences regarding detection and isolation of malicious nodes from the routing process are made.
The next paper titled ‘Impact ionization rate of electrons in monolayer graphene nano-ribbons’, gives an analytical model for evaluating the impact ionization rate of electrons in monolayer graphene nano-ribbons (GNR). The model accounts for optical phonon scattering and electron-electron collision events prior to impact ionization. The impact ionization rate of electrons in monolayer GNR is compared with analytical model from the literature. The paper gives the effects of temperature and sheet electron concentration on the ionization rate.
In the following paper on ‘Performance analysis of different tree structures in heterogeneous wireless sensor network’, the authors compare three network structures viz. 2-tree, spanning tree and complete ternary tree for sensor node placement, and analyze them in Qualnet based on different simulation parameters such as energy consumption in different modes, residual battery capacity, average jitter, average delay, average path loss, average transmission delay and signal errors. Based on these studies, inferences about the three tree structures are made.
The next paper on ‘Upgrading legacy long-haul WDM systems through unbalancing the proportion of 1s and 0s in the transmitted data’, gives BER performance results for WDM transmission using unbalanced proportion of 1s and 0s for pseudo-random binary sequence. Such unbalanced transmission can emulate practical coding systems, and the study provides insight into the error-rate performance that can be used to improve legacy long-haul WDM system performance by operating it in the nonlinear power region.
The paper titled ‘Miniaturized dual broadband hexagonal slot monopole antenna’, gives the design of a miniaturized dual broadband hexagonal slot monopole antenna for purpose of WLAN bands, WIMAX bands and partial X-band applications. The miniaturization is studied by bisecting the reference antenna and the dual broadband behavior is achieved with further miniaturization. The designed reference antennas are fabricated and their performance is measured and compared with the expected performance from simulation studies.
In the paper titled ‘Phonetic engine for continuous speech in Malayalam’, the authors present a phonetic engine for continuous speech recognition in Malayalam. The phonetic engine is speaker independent as well as domain independent. The date collected in three different speaking modes viz. read mode, lecture mode, and conversation mode are transcribed using the International Phonetic Alphabet and the phonemes are modeled using continuous HMM. The performance of the phonetic engine is evaluated and reported in the paper.
In ‘Band notched UWB-BPF based on broadside coupled microstrip/CPW transition’, the design of a band notched planar ultra-wideband bandpass filter is presented. The design is based on broadside coupling of microstrip/co-planar waveguide transition. The wide stopband provides improved isolation by the suppression of spurious harmonics produced otherwise. The performance of the design is reported from the fabricated prototype and is compared with the performance of other dual notch structures.
The next paper titled ‘A tunable leapfrog complex filter for high frequency and wide bandwidth using multi-output current differencing trans-conductance amplifiers’, gives the design for realization of a 4th-order polyphase filter based on the leapfrog structure. The first-order stage uses multi-output current differencing trans-conductance amplifiers. This is useful in simplifying the circuit of the current-mode leapfrog polyphase filter due to the absence of resistors and use of minimum active components. The theoretical results are compared with simulation in a standard 0.18 μm CMOS process.
The paper on ‘Design procedure for three-stage CMOS OTAs with AC boosting frequency compensation technique’, gives a design procedure for 3-stage cascade CMOS operational amplifier that is compensated by AC boosting frequency compensation method. The design procedure provides a first-cut design without introducing circuit constraints. The procedure has been studied in simulation environment for TSMC 0.18 μm CMOS process, and several insights are presented regarding the design procedure. The procedure is also compared with several prior designs from the literature.
In the paper ‘Miniaturization of monopole antenna by metamaterial loading technique’, a monopole antenna design is presented with left-handed metamaterial for reducing the resonant frequency in a miniaturized realization. The design is supported with numerical, analytical and experimental analyses.
The following paper on ‘A novel localized entropy-based medical image retrieval’, presents two methods for retrieval of images. These are the fuzzy connectedness image segmentation with geometric moments and localized entropy based medical image retrieval. The first method uses the shape features for improving the accuracy of image retrieval while the second method uses localized entropy for reducing the computational complexity. These methods are compared with existing methods in terms of precision and recall.