Aiming at traditional neural networks non-uniformity correction (NUC) algorithm's disadvantages such as slow convergence, low correction precision and difficulty to meet the real-time engineering application requirements of infrared imaging system, an improved NUC algorithm for infrared focal plane arrays (IRFPA) based on neural network is proposed. The algorithm is based on linear response of detector, and in order to realize fast and synchronization convergence of correction parameters the each original image data is normalized to a value close to one. Experimental results show the method has the faster convergence speed and better vision effect than the traditional algorithms, and it is better applied in practical projects.

This paper proposes a novel service configuration approach that can realize dynamic critical Quality of Service (QoS) adaptation to ever-changing and resource-limited ubiquitous computing environments. In the approach, service configuration is reduced to a Fuzzy Control System (FCS) which aims to achieve critical QoS variations on minimal level with less power cost. Two configuration strategies, service chain reconfiguration and QoS parameters adjustment, along with a configuration algorithm, are implemented to handle different types of QoS variations. A self-optimizing algorithm is designed to enhance the adaptation of the FCS. Simulation results validate the proposed approach.

Dimensionality reduction is one of the important preprocessing steps in practical pattern recognition. SEmi-supervised Local Fisher discriminant analysis (SELF)--which is a semi-supervised and local extension of Fisher discriminant analysis--was shown to work excellently in experiments. However, when data dimensionality is very high, a naive use of SELF is prohibitive due to high computational costs and large memory requirement. In this paper, we introduce computational tricks for making SELF applicable to large-scale problems.

A system for detecting fundus lesions caused by diabetic retinopathy from fundus images is being developed. The system can screen the images in advance in order to reduce the inspection workload on doctors. One of the difficulties that must be addressed in completing this system is how to remove false positives (which tend to arise near blood vessels) without decreasing the detection rate of lesions in other areas. To overcome this difficulty, we developed classifier selection according to the position of a candidate lesion, and we introduced new features that can distinguish true lesions from false positives. A system incorporating classifier selection and these new features was tested in experiments using 55 fundus images with some lesions and 223 images without lesions. The results of the experiments confirm the effectiveness of the proposed system, namely, degrees of sensitivity and specificity of 98% and 81%, respectively.

We have proposed a terahertz (THz) emitter utilizing two-dimensional plasmons (2DPs) in a super-grating dual-gate (SGG) high electron mobility transistor (HEMT). The plasmon under each grating gate has a unique feature that its resonant frequency is determined by the plasma-wave velocity over the gate length. Since the drain bias voltage causes a linear potential slope from the source to drain area, the sheet electron densities in periodically distributed 2DP cavities are dispersed. As a result, all the resonant frequencies are dispersed and undesirable spectral broadening occurs. A SGG structure can compensate for the sheet electron density distribution by modulating the grating dimension. The finite difference time domain simulation confirms its spectral narrowing effect. Within a wide detuning range for the gate and drain bias voltages giving a frequency shifting of 0.5 THz from an optimum condition, the SGG structure can preserve the spectral narrowing effect.

In recent years huge potential benefits from novel applications in mobile ad hoc networks (MANET) have been discussed extensively. However, without robust security mechanisms and systems to provide safety shell through the MANET infrastructure, MANET applications can be vulnerable and hammered by malicious attackers easily. In order to detect misbehaved message routing and identify malicious attackers in MANET, schemes based on reputation concept have shown their advantages in this area in terms of good scalability and simple threshold-based detection strategy. We observed that previous reputation schemes generally use predefined thresholds which do not take into account the effect of behavior dynamics between nodes in a period of time. In this paper, we propose a Threshold-Adaptive Reputation System (TARS) to overcome the shortcomings of static threshold strategy and improve the overall MANET performance under misbehaved routing attack. A fuzzy-based inference engine is introduced to evaluate the trustiness of a node's one-hop neighbors. Malicious nodes whose trust values are lower than the adaptive threshold, will be detected and filtered out by their honest neighbors during trustiness evaluation process. The results of network simulation show that the TARS outperforms other compared schemes under security attacks in most cases and at the same time reduces the decrease of total packet delivery ratio by 67% in comparison with MANET without reputation system.

In digital communication systems employing binary phase-shift keying and non-data-aided carrier phase recovery, a 180 carrier phase ambiguity is inevitable. Here, we propose a simple modification to the standard regular repeat-accumulate (RA) code structure by exploiting the differential encoding inherent to the inner encoder of RA codes resulting in codes that are 180 rotationally invariant. The proposed code structure exhibit performance virtually identical to that of standard regular RA codes with zero carrier phase offset under both zero and 180 carrier phase offsets with negligible additional hardware complexity.

Fair allocation of bandwidth and maximization of channel utilization are two important issues when designing a contention-based wireless medium access control (MAC) protocol. However, fulfilling both design goals at the same time is very difficult. Considering the problem in the IEEE 802.11 wireless local area networks (WLANs), in this work we propose a method using a p-persistent enhanced DCF, called P-IEEE 802.11 DCF, to achieve weighted fairness and efficient channel utilization among multiple priority classes in a WLAN. Its key idea is that when the back-off timer of a node reaches zero, the transmission probability is properly controlled to reflect the relative weights among data traffic flows so as to maximize the aggregate throughput and to minimize the frame delay at the same time. In particular, we obtain the optimal transmission probability based on a theoretical analysis, and also provide an approximation to this probability. The derived optimal and approximation are all evaluated numerically and simulated with different scenarios. The results show that the proposed method can fulfill our design goals under different numbers of priority classes and different numbers of nodes.

In order to improve the synchronization performance of the OFDM-based FM broadcasting system, this letter addresses the problem of delay selection when the system uses a cyclic delay diversity (CDD) scheme. By proper selection of the amount of cyclic delay, an improved fine carrier frequency offset estimator is derived. By computer simulation, the proposed estimator is shown to benefit from properly chosen delay parameter and perform robustly.

Single-carrier (SC) multiple access is a promising uplink multiple access technique because of its low peak-to-average power ratio (PAPR) property and high frequency diversity gain that is achievable through simple one-tap frequency-domain equalization (FDE) in a strong frequency-selective channel. The multiple access capability can be obtained by combining either frequency division multiple access (FDMA) or code division multiple access (CDMA) with SC transmission. In this article, we review the recent research on the SC multiple access techniques with one-tap FDE. After introducing the principle of joint FDE/antenna diversity combining, we review various SC multiple access techniques with one-tap FDE, i.e., SC-FDMA, SC-CDMA, block spread CDMA, and delay-time/CDMA.

This paper proposes a novel RBF training algorithm based on immune operations for dynamic problem solving. The algorithm takes inspiration from the dynamic nature of natural immune system and locally-tuned structure of RBF neural network. Through immune operations of vaccination and immune response, the RBF network can dynamically adapt to environments according to changes in the training set. Simulation results demonstrate that RBF equalizer based on the proposed algorithm obtains good performance in nonlinear time-varying channels.

In this paper, we propose a simple nonlinear system which consists of a chaotic spiking oscillator and a controlling circuit to stabilize unknown periodic orbits. Our proposed system generates various stabilized unknown Unstable Periodic Orbits which are embedded on the chaotic attractor of the original chaotic spiking oscillator. The proposed system is simple and exhibits various bifurcation phenomena. The dynamics of the system is governed by 1-D piecewise linear return map. Therefore, the rigorous analysis can be performed. We provide conditions for stability and almost complete analysis for bifurcation and co-existence phenomena by using the 1-D return map. An implementation example of the controlled chaotic spiking oscillator is provided to confirm some theoretical results.

Ant colony optimization (ACO) algorithms are a recently developed, population-based approach which has been successfully applied to optimization problems. However, in the ACO algorithms it is difficult to adjust the balance between intensification and diversification and thus the performance is not always very well. In this work, we propose an improved ACO algorithm in which some of ants can evolve by performing genetic operation, and the balance between intensification and diversification can be adjusted by numbers of ants which perform genetic operation. The proposed algorithm is tested by simulating the Traveling Salesman Problem (TSP). Experimental studies show that the proposed ACO algorithm with genetic operation has superior performance when compared to other existing ACO algorithms.

Automatic location identification is required for emergency calls in Japan and many other countries. However, promptly providing a reliable position is a problem in places where the global positioning system (GPS) signal is extremely weak. We propose an acquisition scheme for the assisted GPS (AGPS) architecture based on a timing-synchronized mobile network. System errors as well as code uncertainty range and frequency uncertainty range are discussed. With this method, the acquisition search range is significantly reduced, and a long-time coherent correlation is made possible. Simulations and experiments prove that the method is fast, efficient and power-saving for the user handset. Furthermore, the AGPS receiver can perform a short TTFF (Time To First Fix) start even with no ephemeris kept in the handset.

In this paper, we study and propose an inter-cell co-channel interference (CCI) mitigation method for pilot signals using cyclic shift Zadoff-Chu (CS-ZC) sequences for SC-FDMA-based uplink without tight scheduler coordination among cells. Firstly, we investigate the issue of severe detection performance degradation created by the lack of orthogonality among the pilot signals without alignment of the allocated frequency resource positions among cells when using the conventional CS-ZC sequences generation scheme. Secondly, we identify the primary factor causing the issue. Thirdly, we propose a frequency-dependent CS-ZC sequence generation scheme by allocating the same spectrum elements of the ZC sequence to the overlapped subcarriers among cells to mitigate the inter-cell CCI of the pilot signals without alignment of the frequency resource positions among cells. Finally, we confirm the validity of the proposal using uplink data BLER evaluation under a multipath fading condition by computer simulation compared to the conventional method, and show that the proposal achieves around 0.9 dB and 0.6 dB better performance at 10% BLER than the conventional method for 1 RB and 2 RBs frequency offsets in 3 RBs transmission bandwidth, respectively.

A novel adaptive reputation-based virtual organization formation is proposed. It restrains the bad performers effectively based on the consideration of the global experience of the evaluator and evaluates the direct trust relation between two grid nodes accurately by consulting the previous trust value rationally. It also consults and improves the reputation evaluation process in PathTrust model by taking account of the inter-organizational trust relationship and combines it with direct and recommended trust in a weighted way, which makes the algorithm more robust against collusion attacks. Additionally, the proposed algorithm considers the perspective of the VO creator and takes required VO services as one of the most important fine-grained evaluation criterion, which makes the algorithm more suitable for constructing VOs in grid environments that include autonomous organizations. Simulation results show that our algorithm restrains the bad performers and resists against fake transaction attacks and badmouth attacks effectively. It provides a clear advantage in the design of a VO infrastructure.

Time variations of wireless multipath channels can lead to severe intercarrier interference (ICI) in orthogonal frequency division multiplex (OFDM) systems, whereas large Doppler frequency spread can provide us with time diversity gain. In order to take advantage of the time diversity and to suppress the interference and noise enhancement at the same time, the receiver normally detects the data successively. In this letter, we propose an improved detection ordering based on the log-likelihood ratio (LLR) rather than the signal-to-noise ratio (SNR) for the successive detector. Using both theoretical analysis and computer simulation, it is shown that this scheme outperforms the traditional successive detection methods.

This paper presents a comparison of hierarchical and non-hierarchical synchronization channel (SCH) structures in terms of the initial cell search time and neighboring cell search time in order to establish the optimum SCH structure in the Evolved UTRA downlink. Computer simulation results show that in a 19-cell configuration, the cell search time at 90% in the cumulative distribution function (CDF) using the hierarchical SCH structure is less than half that using the non-hierarchical SCH structure in a neighboring cell search under low signal-to-interference plus noise power ratio (SINR) conditions, although both structures achieve almost the same cell search time in the initial cell search. This is due to the cross-correlation based SCH symbol timing detection in the hierarchical SCH structure, which is affected less by noise than the auto-correlation based detection in the non-hierarchical SCH structure. Thus, we conclude that the hierarchical SCH structure is superior to the non-hierarchical SCH structure based on the cell search time performance especially in the neighboring cell search.

This paper compares the turbo code and rate-compatible low-density parity-check (LDPC) codes based on the block error rate (BLER) performance and decoding complexity in order to clarify which channel coding scheme is most appropriate for the channel coding scheme in the OFDM based Evolved UTRA (E-UTRA) downlink. Simulation results and the decoding complexity analysis show that although the Rate-Compatible/Quasi-Cyclic (RC/QC)-LDPC code employing an offset layered belief propagation (BP) method can reduce the computational complexity by approximately 30% for the channel coding rate of R ≥ 1/2, the required average received signal energy per bit-to-noise power spectrum density ratio (Eb/N0) is degraded by approximately 0.2-0.3 dB for R = 1/3, 1/2 and 3/4 compared to that for the turbo code. Moreover, the decoding complexity level of the RC/QC-LDPC code with the δ-min algorithm is almost the same or higher than that for the turbo code with a slight degradation in the required received Eb/N0. Although the decoding complexity level of the ZigZag code is lower than that of the turbo code, the code brings about a distinct loss in the required average received Eb/N0 of approximately 0.4 dB. Finally, the turbo Single Parity Check (SPC) code improves the BLER performance compared to the ZigZag code, i.e., achieves almost the same BLER performance as that for the turbo code, at the cost of a two-fold increase in the decoding complexity. As a result, we conclude that the turbo code with a contention free interleaver is more promising than the LDPC codes for prioritizing the achievable performance over complexity and as the channel coding scheme for the shared data channel in the E-UTRA.

This paper proposes a software-controllable variable line-size (SC-VLS) cache architecture for low power embedded systems. High bandwidth between logic and a DRAM is realized by means of advanced integrated technology. System-in-Silicon is one of the architectural frameworks to realize the high bandwidth. An ASIC and a specific SRAM are mounted onto a silicon interposer. Each chip is connected to the silicon interposer by eutectic solder bumps. In the framework, it is important to reduce the DRAM energy consumption. The specific DRAM needs a small cache memory to improve the performance. We exploit the cache to reduce the DRAM energy consumption. During application program executions, an adequate cache line size which produces the lowest cache miss ratio is varied because the amount of spatial locality of memory references changes. If we employ a large cache line size, we can expect the effect of prefetching. However, the DRAM energy consumption is larger than a small line size because of the huge number of banks are accessed. The SC-VLS cache is able to change a line size to an adequate one at runtime with a small area and power overheads. We analyze the adequate line size and insert line size change instructions at the beginning of each function of a target program before executing the program. In our evaluation, it is observed that the SC-VLS cache reduces the DRAM energy consumption up to 88%, compared to a conventional cache with fixed 256 B lines.