This letter considers the estimation of modulation index for binary continuous phase modulated (CPM) partial response signals. The actual value of modulation index in low-cost implementations turns out to be different than the nominal value used at the transmitter, leading to severe performance degradation. One possible solution is the deployment of an estimation algorithm at the receiver. The proposed data-aided best linear unbiased estimator (BLUE) is based upon a linear model approximation of the differential phase of pre-averaged CPM signal. Theoretical analysis of the variance of the estimator is presented to establish its improved performance. Simulation results also corroborate significant improvement of the proposed estimator at lower signal-to-noise ratio (SNR) values as compared to existing estimators, while retaining the linear bound performance at high SNRs. We show that the bit error rate (BER) performance at low SNRs is also improved when the modified BLUE estimator is deployed at the receiver.

Device-to-device (D2D) communication is a novel paradigm in cellular networks and is being considered as one of the primary technologies for the upcoming fifth generation (5G) standard for cellular communication. On a similar note, millimeter wave (mmWave) communication is also one of the very enablers for 5G. For D2D communication, neighbor or device discovery is a fundamental problem in mmWave networks because of the use of highly directional antennas. In this paper, we consider this fundamental neighbor discovery problem: How can nodes discover their neighbors quickly and efficiently for communication in decentralized networks, without any prior coordination and with heterogeneous antenna configurations. We propose a novel D2D neighbor discovery algorithm that uses the idea of necklaces to reduce the worst case discovery delay as compared to the previous approaches. The results from numerical simulations confirm that the proposed algorithm leads to faster and energy efficient neighbor discover as compared to the existing algorithms.

The outage probability of a dual relay simultaneous wireless information and power transfer (SWIPT) system is investigated in the presence of Rayleigh fading. The message forwarding at the relays is categorized in three schemes. In the first two cases, both the relays are considered to be decode-and-forward (DF) and amplify-and-forward (AF), respectively, whereas in the third case, one of the relays is considered DF and the other as AF. The relaying model considers the source-relay-destination links whereas the direct link between source-destination does not exist. The power splitters at the relaying devices provide energy to the relays by splitting the received signal power into energy harvesting and information transfer parts. The outage probability is investigated using Monte-Carlo simulations and the results for all the three cases are compared. The all-AF forwarding scheme provides the least outage probability among all the three cases. The minimum outage probability is obtained for different values of PS factors for the said case at different transmit powers.

This paper proposes a method of recognizing and classifying the basic activities such as forward and backward motions by applying a deep learning framework on passive radio frequency (RF) signals. The echoes from the moving body possess unique pattern which can be used to recognize and classify the activity. A passive RF sensing test- bed is set up with two channels where the first one is the reference channel providing the un- altered echoes of the transmitter signals and the other one is the surveillance channel providing the echoes of the transmitter signals reflecting from the moving body in the area of interest. The echoes of the transmitter signals are eliminated from the surveillance signals by performing adaptive filtering. The resultant time series signal is classified into different motions as predicted by proposed novel method of convolutional neural network (CNN). Extensive amount of training data has been collected to train the model, which serves as a reference benchmark for the later studies in this field.

This paper studies the performance of a cooperative opportunistic large array (OLA) network, in which the nodes are deployed randomly in a strip-shaped fashion. Specifically, the source and the destination nodes are separated by a large distance and a number of randomly deployed relay nodes help the source message to be delivered to the destination using multi-hop cooperation strategy. The performance of network is carried out in terms of the success rate and the average number of hops to reach the destination. The reliability of the OLA network is increased, when the same message is transmitted by a group of nodes in multiple hops by exploiting spatial diversity gains. Towards the end, energy efficiency in the networks can be achieved by limiting few nodes in each hop without compromising on the quality of service (QOS). The paper summarizes and compares the performance of various network topologies in terms of energy efficiency and outage probability.

In this paper, we consider the problem of cyclic block construction, for frequency-domain equalization (FDE) of SOQPSK, using an intrafix segment in addition to the cyclic prefix. We propose a novel procedure to compute the intrafix symbols for SOQPSK-TG waveforms used in aeronautical telemetry. The proposed method is employed to generate an exact cyclic signal, which is a requirement for implementation of FDE at the receiver. We identify that the required number of symbols to generate an exact cyclic signal are unreasonably large for SOQPSK-TG (9 symbols per block). Thus, we also propose an approximate technique that utilizes only dominant Laurent pulses to help reduce the length of intrafix upto 2 symbols. A detailed numerical evaluation of this approximate technique is performed to study the trade-off between the reduction in the intrafix length and the ideal properties of the exact signal. Bit-error rate (BER) results are presented for minimum-mean squared error (MMSE) equalizer when the cyclic part of the received signal is made by approximation. Simulation results show that MMSE equalizer shows similar level of performance with original and approximate signal.

This paper presents a method for the recognition of Glagolitic characters in degraded historical documents. The Glagolitic character recognition is based on Dense SIFT for which image restoration is proposed as a pre-processing step in order to suppress background noise in degraded documents. Two different methods for image restoration are used which are Total Variation regularization and a new restoration method. Each method performs robustly against background noise while preserving character edges and strokes in the documents defected by stain, bleed through, and faded out ink. The experimental results achieved on three datasets show that by using image restoration as a pre-processing step to Dense SIFT generates better recognition rates for Glagolitic characters in degraded documents.

Scene acquisition using RGB and Near Infra-Red (NIR) filters generates useful visual information about scene contents. But it induces significant intensity and textural changes between RGB and NIR images of the same scene. It becomes a challenging problem to perform interest point based image matching under such intensity and textural changes. To cope with this problem, a novel method for the description of interest points is proposed. The method proposed is based on Center Symmetric-Local Binary Patterns (CS-LBP) which extracts distinct image features from intensity and gradient magnitude maps of the image patches centered at interest points. Those features are then used in the SIFT algorithm to compute robust descriptors against intensity and textural changes. The experimental results show that the method proposed improves the descriptor matching between RGB and NIR images and achieves better image matching results than CS-LBP and SIFT based methods for the description of interest points.

This letter presents a novel method for the description of multispectral image keypoints. The method proposed is based on a modified SIFT algorithm. It uses normalized gradients as local image features for the description of keypoints in order to achieve robustness against non linear intensity changes between multispectral images. The experimental results show that the method proposed achieves a better matching performance and outperforms the SIFT algorithm.

Two receiver architectures are presented for equalization of multi-h continuous phase modulation (CPM). In the first approach, the receiver front end is a matched-filter bank, which is derived from Laurent decomposition of the signal. A circulant-matrix model derived for the samples of the matched filter is used to perform the equalization in frequency domain. The second technique employs a receiver with a simple front end consisting of a low-pass filter and a fractional sampler. We also develop a circulant matrix model for the latter approach, which relates the Laurent pseudo-symbols to each polyphase component at the output of the sampler. A linear frequency-domain equalizer (FDE) is also proposed using this model. In addition to providing the details of these approaches, the paper proposes several simplifications applicable to each of the two cases, and provides a detailed discussion on the performance/complexity trade-off. Bit-error rate (BER) results are presented for the IRIG-106 aeronautical telemetry dual-h CPM waveform.

Frequency-domain equalization (FDE) is a low-complexity approach to remove the effects of intersymbol interference (ISI) introduced by propagation over a frequency-selective channel. In this paper, we present a novel receiver architecture for frequency-domain equalization of multi-h continuous phase modulation (CPM). The receiver front-end is a matched-filter bank, which is derived from a Laurent decomposition of the signal. A circulant-matrix model obtained for the samples of the matched filter is used to perform the equalization in frequency domain. It is shown that receiver complexity can be reduced by allowing a slight degradation in performance. The reduction in the complexity of the equalizer is proposed by dropping the FDE branches corresponding to low-energy Laurent pulses. It is observed that the loss in performance is acceptable even when a few significant branches are considered. Simulation results are presented for the IRIG-106 aeronautical telemetry dual-h CPM waveform.

This paper considers the construction of a symbol block to generate a cyclic multi-h continuous phase modulated (CPM) signal so as to permit frequency domain equalization (FDE) at the receiver. Like single-h CPM, the memory inherent in a multi-h CPM waveform requires an intra-fix in each transmitted block. We propose a novel procedure that determines the intra-fix symbols, such that the intra-fix length is minimized. The constraints on the required intra-fix symbols are shown to form a set of Linear Diophantine Equations (LDEs). The proposed procedure is applied to a dual-h CPM scheme.

This paper considers the problem of constructing a symbol block to generate a cyclic multi-h continuous phase modulated (CPM) signal so as to permit frequency domain equalization (FDE) at the receiver. Like single-h CPM, the memory inherent in a multi-h CPM waveform requires the use of an additional segment, called an intra-fix, in each transmitted block. The intra-fix forces the transmitter state machine to regress to the state it had at the start of the block. The constraints on the symbols of the intra-fix, which ensure above condition, are shown to be a set of Linear Diophantine Equations (LDEs). Three different methods are proposed for finding the symbols of the intra-fix, such that the length of the intra-fix is minimized. The last two methods connect the problem of intra-fix insertion to the problem of trellis termination of the tilted phase. This connection enables us to state and justify that existing bounds established on the length of the terminating sequence are effective for the length of the intra-fix. We observe that multi-h CPM schemes typically require a shorter intra-fix, as compared to single-h CPM schemes having similar modulation parameters. We apply the proposed procedure to a dual-h CPM scheme.

This paper addresses the problem of cyclic prefix insertion for multi-h CPM waveforms. Like single-h CPM, the memory inherent in a multi-h CPM waveform, and its continuous nature, requires the use of an intra-fix in each transmitted block. We identify the structure of the block, the constraints on its length, and the symbols of the intra-fix. A procedure is proposed for finding the symbols of the intra-fix using the tilted phase transformation. The procedure ensures that the overhead due to the intra-fix is minimized. Bounds on the intra-fix length are also presented in terms of parameters of the modulation scheme. The proposed procedure is applied on a dual-h CPM scheme for illustration of analytical results.

This paper presents a method for terminating the phase trellis in Multi-h Continuous Phase modulation (CPM). It is shown that tilted phase trellis for a Multi-h CPM signal is periodic and enables the representation of the modulator as the cascade of a periodic recursive continuous phase encoder (CPE) and a phase modulator (PM). We then pose the problem of finding the number of symbols that will terminate the tilted phase trellis, and establish a connection of its solution to the Diophantine Frobenius Problem (FP). Results are derived for the existence of a terminating sequence (TS), and bounds are established on its length in terms of the modulation parameters.

This paper presents a method for terminating the phase trellis in multi-h continuous phase modulation (CPM). It is shown that tilted phase trellis for multi-h CPM signals is periodic and enables the representation of the multi-h CPM modulator as a cascade of a periodic recursive continuous phase encoder (CPE) and a phase modulator (PM). We then pose the problem of finding the minimum number of symbols that will terminate the phase trellis in a known state of the modulator and establish a connection of its solution to the Diophantine Frobenius Problem (FP). Analytical results are established for the existence of a terminating sequence and upper and lower bounds are derived on its length for arbitrary modulation parameters. We show that the trellis terminating symbols and their quantity depends upon the initial state of the encoder, the set of modulation indices, the CPE memory order, and the number of bits per modulated symbol. It is shown that required overhead for trellis termination can be significant for schemes with integer modulation indices.

This paper presents a novel approach for registration of aerial and satellite farmland imagery. We have used field junctions as features for registration. Geometric properties of a group of neighboring fields connected to the junction are used as junction descriptor. These properties are calculated by fitting a rectangle to the field. Junction descriptor contains a group of field patterns capable of finding its match and indexing aerial image spatial location in satellite images in spite of repetitive structure of fields. It is invariant to in-plane rotation, scale and translation. Experimental results demonstrate the robustness of our approach against the junction localization errors and the errors introduced by splitting and/or merging of fields.

A significant number of new Java-based technologies for mobile code (aka agents) have recently emerged. The 'Aglets' system, from IBM's research labs, provides an elegant mechanism for creating mobile code, but lacks a native scheduling mechanism for determining where code should be executed. We present the results of an investigation into adding sophisticated scheduling capabilities to Aglets (which we refer to as brilliant Aglets, or Baglets) from the H-SWEB project, which provides hierarchical scheduling across sets of WWW server clusters. H-SWEB uses scheduling techniques in monitoring and adapting to workload variation at distributed server clusters for supporting distributed computation. We show how the two systems can be integrated, and present several algorithms indicating a major advantage (over 350%) can be achieved through the use of dynamic scheduling information. We provide a detailed discussion of our system architecture and implementation, and briefly summarize the experimental results which have been achieved.