Significant challenges in ball tracking of sports analysis by computer vision technology are: 1) accuracy of estimated 3D ball trajectory under difficult conditions; 2) external forces added by players lead to irregular motions of the ball; 3) unpredictable situations in the real game, i.e. the ball occluded by players and other objects, complex background and changing lighting condition. With the goal of multi-view 3D ball tracking, this paper proposes an abrupt motion adaptive system model, an anti-occlusion observation model, and a spatial density-based automatic recovery based on particle filter. The system model combines two different system noises that cover the motion of the ball both in general situation and situation subject to abrupt motion caused by external force. Combination ratio of these two noises and number of particles are adaptive to the estimated motion by weight distribution of particles. The anti-occlusion observation model evaluates image feature of each camera and eliminates influence of the camera with less confidence. The spatial density, which is calculated based on 3D ball candidates filtered out by spatial homographic relationship between cameras, is proposed for generating new set of particles to recover the tracking when tracking failure is detected. Experimental results based on HDTV video sequences (2014 Inter High School Men's Volleyball Games, Japan), which were captured by four cameras located at each corner of the court, show that the success rate achieved by the proposals of 3D ball tracking is 99.42%.

Effective communication strategies with a properly designed source precoding matrix (PM) and a properly designed relay beamforming matrix (BM) can significantly improve the spectral efficiency of multiple-input multiple-output (MIMO) relaying broadcast channels (RBCs). In the present paper, we first propose a general communication scheme with non-regenerative relay that can overcome the half-duplex relay constraint of the general MIMO-RBC. Based on the proposed scheme, the robust source PM and relay BM are designed for imperfect channel state information at the transmitter (CSIT). In contrast to the conventional non-regenerative relaying communication scheme for the MIMO-RBC, in the proposed scheme, the source can send information continuously to the relay and users during two phases. Furthermore, in conjunction with the advanced precoding strategy, the proposed scheme can achieve a full-degree-of-freedom (DoF) MIMO-RBC with that each entry in the related channel matrix is considered to an i.i.d. complex Gaussian variable. The robust source PM and relay BM designs were investigated based on both throughput and fairness criteria with imperfect CSIT. However, solving the problems associated with throughput and fairness criteria for the robust source PM and relay BM designs is computationally intractable because these criteria are non-linear and non-convex. In order to address these difficulties, we first set up equivalent optimization problems based on a tight lower bound of the achievable rate. We then decompose the equivalent throughput problem into several decoupled subproblems with tractable solutions. Finally, we obtain the suboptimal solution for the throughput problem by an alternating optimization approach. We solve the fairness problem by introducing an adjusted algorithm according to the throughput problem. Finally, we demonstrate that, in both cases of throughput and fairness criteria, the proposed relaying communication scheme with precoding algorithms outperforms existing methods.

We propose a heuristic approximation algorithm for the 1-median problem. The 1-median problem is the problem of finding a vertex with the highest closeness centrality. Starting from a randomly selected vertex, our algorithm repeats to find a vertex with higher closeness centrality by approximately calculating closeness centrality of each vertex using simpler spanning subgraphs, which are called k-neighbor dense shortest path graphs with shortcuts. According to our experimental results using real networks with more than 10,000 vertices, our algorithm is more than 100 times faster than the exhaustive search and more than 20 times faster than the state-of-the-art approximation algorithm using annotated information to the vertices while the solutions output by our algorithm have higher approximation ratio.

This paper proposes an effective continuous super-resolution (CSR) algorithm for the multipath channel estimation. By designing a preamble including up-chirp and down-chirp symbols, the Doppler shift and multipath delay are estimated jointly by using convex programming. Simulation results show that the proposed CSR can achieve better detection probability of the number of multipaths than the eigenvalue based methods. Moreover, compared with conventional super-resolution techniques, such as MUSIC and ESPRIT methods, the proposed CSR algorithm demonstrates its advantage in root mean square error of the Doppler shift and multipath delay, especially for the closely located paths within low SNR.

In this paper, we consider a coherently distributed (CD) source model. Since the CD source is characterized by four parameters: central azimuth direction-of-arrival (DOA), azimuth angular spread, central elevation DOA and elevation angular spread, the parameter estimation is normally complex. We propose an algorithm that combines the rotational invariance techniques (ESPRIT) and the generalized ESPRIT algorithm for the 2-dimensional (2D) central DOA estimation of CD sources. Using a pair of uniform circular arrays (UCAs), the proposed solution is able to obtain the central DOAs with both high accuracy and low computational complexity. The central elevation DOAs are estimated by using the rotational invariance relation between the two uniform circular sub-arrays. Based on the centrosymmetric structure of UCA, the generalized ESPRIT algorithm is then applied to estimate the central azimuth DOAs through one-dimensional searching. It is noteworthy that the central DOAs are estimated without any information of the deterministic angular distribution function (DADF). The performance of the proposed algorithm is demonstrated via computer simulations.

In general, tunnel diodes exhibit various types of oscillation mode: the sinusoidal mode or the nonsinusoidal mode which is known as the relaxation oscillation (RO) mode. We derive a condition for generating the RO in resonant tunneling diodes (RTDs) with essential components for equivalent circuit model. A conditional equation to obtain sufficient nonlinearity towards the robust RO is clarified. Moreover, its condition also can be applied in case of a bow-tie antenna integrated RTD, thus a design policy to utilize the RO region for the antenna integrated RTD is established by numerical evaluations of time-domain large-signal nonlinear analysis towards a terahertz transmitter for broadband wireless communications.

To improve the quality of waveforms and achieve a high input power factor (IPF) for matrix rectifier, a novel quasi sliding mode control (SMC) with adaptive compensation is proposed in this letter. Applying quasi-SMC can effective obviate the disturbances of time delay and spatial lag, and SMC based on continuous function is better than discontinuous function to eliminate the chattering. Furthermore, compared with conventional compensation, an adaptive quasi-SMC compensation without any accurate detection for internal parameters is easier to be implementated, which has shown a superior advance. Theoretical analysis and experiments are carried out to validate the correctness of the novel control scheme.

The biometrical identification system, introduced by Willems et al., is a system to identify individuals based on their measurable physical characteristics. Willems et al. characterized the identification capacity of a discrete memoryless biometrical identification system from information theoretic perspectives. Recently, Mori et al. have extended this scenario to list-decoding whose list size is an exponential function of the data length. However, as the data length increases, how the maximum identification error probability (IEP) behaves for a given rate has not yet been characterized for list-decoding. In this letter, we investigate the reliability function of the system under fixed-size list-decoding, which is the optimal exponential behavior of the maximum IEP. We then use Arimoto's argument to analyze a lower bound on the maximum IEP with list-decoding when the rate exceeds the capacity, which leads to the strong converse theorem. All results are derived under the condition that an unknown individual need not be uniformly distributed and the identification process is done without the knowledge of the prior distribution.

The purpose of DNA sequencing is to determine the order of nucleotides within a DNA molecule of target. The target DNA molecules are fragmented into short reads, which are short fixed-length subsequences composed of ‘A’, ‘C’, ‘G’ ‘T’, by next generation sequencing (NGS) machine. To reconstruct the target DNA from the short reads using a reference genome, which is a representative example of a species that was constructed in advance, it is necessary to determine their locations in the target DNA from where they have been extracted by aligning them onto the reference genome. This process is called short read mapping, and it is important to improve the performance of the short read mapping to realize fast DNA sequencing. We propose three types of FPGA acceleration methods based on hash table; (1) sorting and parallel comparison, (2) matching that allows one mutation to reduce the number of the candidates, (3) optimized hash function using variable masks. The first one reduces the number of accesses to off-chip memory to avoid the bottleneck by access latency. The second one enables to reduce the number of the candidates without degrading mapping sensitivity by allowing one mutation in the comparison. The last one reduces hash collisions using a table that was calculated from the reference genome in advance. We implemented the three methods on Xilinx Virtex-7 and evaluated them to show their effectiveness of them. In our experiments, our system achieves 20 fold of processing speed compared with BWA, which is one of the most popular mapping tools. Furthermore, we shows that the our system outperforms one of the fastest FPGA short read mapping systems.

In a real-time system, tasks are required to be completed before their deadlines. Under normal workload conditions, a scheduler with a proper scheduling policy can make all the tasks meet their deadlines. However, in practical environment, system workload may vary widely. Once system workload becomes too heavy, so that there does not exist a feasible schedule can make all the tasks meet their deadlines, we say the system is overloaded under which some tasks will miss their deadlines. To alleviate the degrees of system performance degradation caused by the missed deadline tasks, the design of scheduling is crucial. Many design objectives can be considered. In this paper, we first focus on maximizing the total number of tasks that can be completed before their deadlines. A scheduling method based on satisfiability modulo theories (SMT) is proposed. In the method, the problem of scheduling is treated as a satisfiability problem. The key work is to formalize the satisfiability problem using first-order language. After the formalization, a SMT solver (e.g., Z3, Yices) is employed to solve this satisfiability problem. An optimal schedule can be generated based on the solution model returned by the SMT solver. The correctness of this method and the optimality of the generated schedule can be verified in a straightforward manner. The time efficiency of the proposed method is demonstrated through various simulations. Moreover, in the proposed SMT-based scheduling method, we define the scheduling constraints as system constraints and target constraints. This means if we want to design scheduling to achieve other objectives, only the target constraints need to be modified. To demonstrate this advantage, we adapt the SMT-based scheduling method to other design objectives: maximizing effective processor utilization and maximizing obtained values of completed tasks. Only very little changes are needed in the adaption procedure, which means the proposed SMT-based scheduling method is flexible and sufficiently general.

This paper first reviews the state-of-the-art noise reduction methods and points out their vulnerability in noise reduction performance and speech quality, especially under the low signal-noise ratios (SNR) environments. Then this paper presents an improved perceptual multiband spectral subtraction (MBSS) noise reduction algorithm (NRA) and a novel robust voice activity detection (VAD) based on the amended sub-band SNR. The proposed SNR-based VAD can considerably increase the accuracy of discrimination between noise and speech frame. The simulation results show that the proposed NRA has better segmental SNR (segSNR) and perceptual evaluation of speech quality (PESQ) performance than other noise reduction algorithms especially under low SNR environments. In addition, a fully operational digital hearing aid chip is designed and fabricated in the 0.13 µm CMOS process based on the proposed NRA. The final chip implementation shows that the whole chip dissipates 1.3 mA at the 1.2 V operation. The acoustic test result shows that the maximum output sound pressure level (OSPL) is 114.6 dB SPL, the equivalent input noise is 5.9 dB SPL, and the total harmonic distortion is 2.5%. So the proposed digital hearing aid chip is a promising candidate for high performance hearing-aid systems.

As a hot topic of speech signal processing, speech emotion recognition methods have been developed rapidly in recent years. Some satisfactory results have been achieved. However, it should be noted that most of these methods are trained and evaluated on the same corpus. In reality, the training data and testing data are often collected from different corpora, and the feature distributions of different datasets often follow different distributions. These discrepancies will greatly affect the recognition performance. To tackle this problem, a novel corpus-invariant discriminant feature representation algorithm, called transfer discriminant analysis (TDA), is presented for speech emotion recognition. The basic idea of TDA is to integrate the kernel LDA algorithm and the similarity measurement of distributions into one objective function. Experimental results under the cross-corpus conditions show that our proposed method can significantly improve the recognition rates.

In this paper, we study robust Principal Component Analysis (PCA)-based anomaly detection techniques in network traffic, which can detect traffic anomalies by projecting measured traffic data onto a normal subspace and an anomalous subspace. In a PCA-based anomaly detection, outliers, anomalies with excessively large traffic volume, may contaminate the subspaces and degrade the performance of the detector. To solve this problem, robust PCA methods have been studied. In a robust PCA-based anomaly detection scheme, outliers can be removed from the measured traffic data before constructing the subspaces. Although the robust PCA methods are promising, they incure high computational cost to obtain the optimal location vector and scatter matrix for the subspace. We propose a novel anomaly detection scheme by extending the minimum covariance determinant (MCD) estimator, a robust PCA method. The proposed scheme utilizes the daily periodicity in traffic volume and attempts to detect anomalies for every period of measured traffic. In each period, before constructing the subspace, outliers are removed from the measured traffic data by using a location vector and a scatter matrix obtained in the preceding period. We validate the proposed scheme by applying it to measured traffic data in the Abiline network. Numerical results show that the proposed scheme provides robust anomaly detection with less computational cost.

A digital image is often deteriorated by impulse noise that may occur during processes such as transmission. An impulse noise converts the pixel data in the image into black (0) or white (255) values at a random frequency and is also called salt-and-pepper noise. In this paper, we identify the details of pixels that have been damaged by impulse noise by analyzing the frequency of the noisy image using non-harmonic analysis (NHA). From experimental results, we can confirm that this method shows superior performance compared to the recent PSNR denoising method. In addition, we show that the proposed method is particularly superior in eliminating impulse noise in images with high noise rates.

Variance analysis is an important research topic to assess the quality of estimators. In this paper, we analyze the performance of the least ℓp-norm estimator in the presence of mixture of generalized Gaussian (MGG) noise. In the case of known density parameters, the variance expression of the ℓp-norm minimizer is first derived, for the general complex-valued signal model. Since the formula is a function of p, the optimal value of p corresponding to the minimum variance is then investigated. Simulation results show the correctness of our study and the near-optimality of the ℓp-norm minimizer compared with Cramér-Rao lower bound.

The increase in network access devices and demand for high quality of service (QoS) by the users have led to insufficient capacity for the network operators. Moreover, the existing control equipment and mechanisms are not flexible and agile enough for the dynamically changing environment of heterogeneous cellular networks (HetNets). This non-agile control plane is hard to scale with ever increasing traffic demand and has become the performance bottleneck. Furthermore, the new HetNet architecture requires tight coordination and cooperation for the densely deployed small cell base stations, particularly for interference mitigation and dynamic frequency reuse and sharing. These issues further complicate the existing control plane and can cause serious inefficiencies in terms of users' quality of experience and network performance. This article presents an SDN control framework for energy efficient downlink/uplink scheduling in HetNets. The framework decouples the control plane from data plane by means of a logically centralized controller with distributed agents implemented in separate entities of the network (users and base stations). The scheduling problem consists of three sub-problems: (i) user association, (ii) power control, (iii) resource allocation and (iv) interference mitigation. Moreover, these sub-problems are coupled and must be solved simultaneously. We formulate the DL/UL scheduling in HetNet as an optimization problem and use the Markov approximation framework to propose a distributed economical algorithm. Then, we divide the algorithm into three sub-routines for (i) user association, (ii) power control, (iii) resource allocation and (iv) interference mitigation. These sub-routines are then implemented on different agents of the SDN framework. We run extensive simulation to validate our proposal and finally, present the performance analysis.

Attaining extremely low latency service in 5G cellular networks is an important challenge in the communication research field. A higher QoS in the next-generation network could enable several unprecedented services, such as Tactile Internet, Augmented Reality, and Virtual Reality. However, these services will all need support from powerful computational resources provided through cloud computing. Unfortunately, the geolocation of cloud data centers could be insufficient to satisfy the latency aimed for in 5G networks. The physical distance between servers and users will sometimes be too great to enable quick reaction within the service time boundary. The problem of long latency resulting from long communication distances can be solved by Mobile Edge Computing (MEC), though, which places many servers along the edges of networks. MEC can provide shorter communication latency, but total latency consists of both the transmission and the processing times. Always selecting the closest edge server will lead to a longer computing latency in many cases, especially when there is a mass of users around particular edge servers. Therefore, the research studies the effects of both latencies. The communication latency is represented by hop count, and the computation latency is modeled by processor sharing (PS). An optimization model and selection policies are also proposed. Quantitative evaluations using simulations show that selecting a server according to the lowest total latency leads to the best performance, and permitting an over-latency barrier would further improve results.

Content extraction from deep Web pages has received great attention in recent years. However, the increasingly complicated HTML structure of Web documents makes it more difficult to recognize the data records by only analyzing the HTML source code. In this paper, we propose a method named LTDE to extract data records from a deep Web page. Instead of analyzing the HTML source code, LTDE utilizes the visual features of data records in deep Web pages. A Web page is considered as a finite set of visual blocks. The data records are the visual blocks that have similar layout. We also propose a pattern recognizing method named layout tree to cluster the similar layout visual blocks. The weight of all clusters is calculated, and the visual blocks in the cluster that has the highest weight are chosen as the data records to be extracted. The experiment results show that LTDE has higher effectiveness and better robustness for Web data extraction compared to previous works.

In this paper, the differences in navigation information design, which is important for kiosk-type pedestrian navigation systems, were experimentally examined depending on presence or absence of carriable navigation information in order to acquire the knowledge to contribute design guidelines of kiosk-type pedestrian navigation systems. In particular, we used route complexity information calculated using a regression equation that contained multiple factors. In the absence of carriable navigation information, both the destination arrival rate and route deviation rate improved. Easy routes were designed as M (17 to 39 characters in Japanese), while complicated routes were denoted as L (40 or more characters in Japanese). On the contrary, in the presence of carriable navigation information, the user's memory load was found to be reduced by carrying the same navigation information as kiosk-type terminals. Thus, the reconsideration of kiosk-type pedestrian navigation systems design, e.g., the means of presenting navigation information, is required. For example, if the system attaches importance to a high destination arrival rate, L_Carrying without regard to route complexity is better. If the system attaching importance to the low route deviation rate, M_Carrying in the case of easy routes and L_Carrying in the case of complicated routes have been better. Consequently, this paper presents the differences in the designs of pedestrian navigation systems depending on whether carriable navigation information is absent or present.

In this paper, we propose a non-cooperative line-of-sight (LOS)/non-LOS channel identification algorithm with single node channel measurements based on time-of-arrival statistics. In order to improve the accuracy of channel identification, we adopt a recalibration interval in terms of measured distance to the proposed algorithm. Experimental results are presented in terms of identification probability and recalibration interval. The proposed algorithm involves a trade-off between the channel identification quality and the recalibration rate. However, depending on the recalibration interval, it is possible to greatly improve the sensitivity of the channel identification system.