Hiroko MURAKAMI Koichi SHINODA Sadaoki FURUI
We propose an active learning framework for speech recognition that reduces the amount of data required for acoustic modeling. This framework consists of two steps. We first obtain a phone-error distribution using an acoustic model estimated from transcribed speech data. Then, from a text corpus we select a sentence whose phone-occurrence distribution is close to the phone-error distribution and collect its speech data. We repeat this process to increase the amount of transcribed speech data. We applied this framework to speaker adaptation and acoustic model training. Our evaluation results showed that it significantly reduced the amount of transcribed data while maintaining the same level of accuracy.
In this paper, we propose a novel voice activity detection (VAD) algorithm using global speech absence probability (GSAP) based on Teager energy (TE) for speech enhancement. The proposed method provides a better representation of GSAP, resulting in improved decision performance for speech and noise segments by the use of a TE operator which is employed to suppress the influence of noise signals. The performance of our approach is evaluated by objective tests under various environments, and it is found that the suggested method yields better results than conventional schemes.
Saran TARNOI Wuttipong KUMWILAISAK Poompat SAENGUDOMLERT
This paper presents novel analytical results on the successful decoding probability for random linear network coding in acyclic networks. The results consist of a tight lower bound on the successful decoding probability, its convergence, and its application in constructing a practical algorithm to identify the minimum field size for random linear network coding subject to a target on the successful decoding probability. From the two characterizations of random linear network coding, namely the set of local encoding kernels and the set of global encoding kernels, we first show that choosing randomly and uniformly the coefficients of the local encoding kernels results in uniform and independent coefficients for the global encoding kernels. The set of global encoding kernels for an arbitrary destination is thus a random matrix whose invertibility is equivalent to decodability. The lower bound on the successful decoding probability is then derived in terms of the probability that this random matrix is non-singular. The derived bound is a function of the field size and the dimension of global encoding kernels. The convergence rates of the bound over these two parameters are provided. Compared to the mathematical expression of the exact probability, the derived bound provides a more compact expression and is close to the exact value. As a benefit of the bound, we construct a practical algorithm to identify the minimum field size in order to achieve a target on the successful decoding probability. Simulation and numerical results verify the validity of the derived bound as well as its higher precision than previously established bounds.
Hiroyuki ITO Hiroshi HASEGAWA Ken-ichi SATO
We investigate the possibility of reducing router power consumption through dynamic router performance control. The proposed algorithm employs a typical low pass filter and, therefore, is simple enough to implement in each related element in a router. Numerical experiments using several real Internet traffic data sets show the degree of reduction in power consumption that can be achieved by using the proposed dynamic performance control algorithm. Detailed analysis clarifies the relationships among various parameter values that include packet loss ratios and the degree of power savings. We also propose a simple method based on the leaky bucket model, which can instantaneously estimate the packet loss ratio. It is shown that this simple method yields a good approximation of the results obtained by exact packet-by-packet simulation. The simple method easily enables us to derive appropriate parameter values for the control algorithm for given traffic that may differ in different segments of the Internet.
The energy consumption is always a serious problem for mobile devices powered by battery. As the capacity and density of off-chip memory continuous to scale, its energy consumption accounts for a considerable amount of the whole system energy. There are therefore strong demands for energy efficient techniques towards memory system. Different from previous works, we explore the different power management modes of the off-chip memory by process scheduling for the multi-core mobile devices. In particular, we schedule the processes based on their memory access characteristics to maximize the number of the memory banks being in low power mode. We propose a fast approximation algorithm to solve the scheduling process problem for the dual-core mobile device. And for those equipped with more than two cores, we prove that the scheduling process problem is NP-Hard, and propose two heuristic algorithms. The proposed algorithms are evaluated through a series of experiments, for which we have encouraging results.
ChengDong WU Long CHENG YunZhou ZHANG
In this paper, two efficient redeployment strategies which are designed to balance the detection coverage rate and maintenance period are proposed. To develop these strategies, we first analyze the sensor detection coverage and energy consumption model. We then propose a network maintenance indicator that considers the coverage rate and residual energy in each node. We adopt the network maintenance indicator as the cost function. That is, the network maintenance is formulated as a cost optimization problem. Finally we propose COST_MAX_MIN and COST_MAX_AVG strategies to select the redeployed location of candidate nodes. Simulation results show that the COST_MAX_AVG prolong the repair period in comparison with the COST_MAX_MIN strategy.
Yuki SATOMI Arata KAWAMURA Youji IIGUNI
For an adaptive system identification filter with a stochastic input signal, a coefficient vector updated with an NLMS algorithm converges in the sense of ensemble average and the expected convergence vector has been revealed. When the input signal is periodic, the convergence of the adaptive filter coefficients has also been proved. However, its convergence vector has not been revealed. In this paper, we derive the convergence vector of adaptive filter coefficients updated with the NLMS algorithm in system identification for deterministic sinusoidal inputs. Firstly, we derive the convergence vector when a disturbance does not exist. We show that the derived convergence vector depends only on the initial vector and the sinusoidal frequencies, and it is independent of the step-size for adaptation, sinusoidal amplitudes, and phases. Next, we derive the expected convergence vector when the disturbance exists. Simulation results support the validity of the derived convergence vectors.
Heejun ROH Hoorin PARK Cheoulhoon JUNG Ding-Zhu DU Wonjun LEE
A price-based spectrum investment and pricing scheme in cooperative cognitive radio networks is presented to use wireless resource more efficiently in technical and economic aspects. We analyze the impact of cooperative communications and the relationship between spectrum hole cost and leasing cost in the optimal decision of SAP.
In this letter, distributed source coding with one distortion criterion and correlated messages is considered. This problem can be regarded as “Berger-Yeung problem with correlated messages”. It corresponds to the source coding part of the graph-based framework for transmission of a pair of correlated sources over the multiple-access channel where one is lossless and the other is lossy. As a result, the achievable rate-distortion region for this problem is provided. A rigorous proof of both achievability and converse part is also given.
Hyunduk KIM Boseon YU Wonik CHOI Heemin PARK
We propose a novel scheme that aims to determine the optimal number of clusters based on the field conditions and the positions of mobile sink nodes. In addition, we merge algorithms of tree-based index structures to form an energy-efficient cluster structure. A performance evaluation shows that the proposed method produces highly-balanced clusters that are energy efficient and achieves up to 1.4 times higher survival rates than the previous clustering schemes, under various operational conditions.
Jesus ESQUIVEL-GOMEZ Raul E. BALDERAS-NAVARRO Enrique STEVENS-NAVARRO Jesus ACOSTA-ELIAS
One of the most important constraints in wireless sensor networks (WSN) is that their nodes, in most of the cases, are powered by batteries, which cannot be replaced or recharged easily. In these types of networks, data transmission is one of the processes that consume a lot of energy, and therefore the embedded routing algorithm should consider this issue by establishing optimal routes in order to avoid premature death and eventually having partitioned nodes network. This paper proposes a new routing algorithm for WSN called Micro-Economic Routing Algorithm (MERA), which is based on the microeconomic model of supply-demand. In such algorithm each node comprising the network fixes a cost for relay messages according to their residual battery energy; and before sending information to the base station, the node searches for the most economical route. In order to test the performance of MERA, we varied the initial conditions of the system such as the network size and the number of defined thresholds. This was done in order to measure the time span for which the first node dies and the number of information messages received by the base station. Using the NS-2 simulator, we compared the performance of MERA against the Conditional Minimum Drain Rate (CMDR) algorithm reported in the literature. An optimal threshold value for the residual battery is estimated to be close to 20%.
Jianping WU Ming LING Yang ZHANG Chen MEI Huan WANG
This paper proposes a novel dynamic Scratch-pad Memory allocation strategy to optimize the energy consumption of the memory sub-system. Firstly, the whole program execution process is sliced into several time slots according to the temporal dimension; thereafter, a Time-Slotted Cache Conflict Graph (TSCCG) is introduced to model the behavior of Data Cache (D-Cache) conflicts within each time slot. Then, Integer Nonlinear Programming (INP) is implemented, which can avoid time-consuming linearization process, to select the most profitable data pages. Virtual Memory System (VMS) is adopted to remap those data pages, which will cause severe Cache conflicts within a time slot, to SPM. In order to minimize the swapping overhead of dynamic SPM allocation, a novel SPM controller with a tightly coupled DMA is introduced to issue the swapping operations without CPU's intervention. Last but not the least, this paper discusses the fluctuation of system energy profit based on different MMU page size as well as the Time Slot duration quantitatively. According to our design space exploration, the proposed method can optimize all of the data segments, including global data, heap and stack data in general, and reduce the total energy consumption by 27.28% on average, up to 55.22% with a marginal performance promotion. And comparing to the conventional static CCG (Cache Conflicts Graph), our approach can obtain 24.7% energy profit on average, up to 30.5% with a sight boost in performance.
Yukio OGAWA Go HASEGAWA Masayuki MURATA
When computing resources are consolidated in a few huge data centers, a massive amount of data is transferred to each data center over a wide area network (WAN). This results in increased power consumption in the WAN. A distributed computing network (DCN), such as a content delivery network, can reduce the traffic from/to the data center, thereby decreasing the power consumed in the WAN. In this paper, we focus on the energy-saving aspect of the DCN and evaluate its effectiveness, especially considering traffic locality, i.e., the amount of traffic related to the geographical vicinity. We first formulate the problem of optimizing the DCN power consumption and describe the DCN in detail. Then, numerical evaluations show that, when there is strong traffic locality and the router has ideal energy proportionality, the system's power consumption is reduced to about 50% of the power consumed in the case where a DCN is not used; moreover, this advantage becomes even larger (up to about 30%) when the data center is located farthest from the center of the network topology.
Yuta KAWAMURA Yusuke HORIE Keisuke SANO Hiroya KODAMA Naoki TSUNODA Yuki SHIBUTA Yuki KAWACHI Mitsuho YAMADA
Three-dimensional (3D) movies have become very popular in movie theaters and for home viewing, To date, there has been no report of the effects of the continual vergence eye movement that occurs when viewing 3D movies from the beginning to the end. First, we analyzed the influence of viewing a 3D movie for several hours on vergence eye movement. At the same time, we investigated the influence of long viewing on the human body, using the Simulator Sickness Questionnaire (SSQ) and critical fusion frequency (CFF). It was suggested that the vergence stable time after saccade when viewing a long movie was influenced by the viewing time and that the vergence stable time after saccade depended on the content of the movie. Also the differences were seen in the SSQ and CFF between the movie's beginning and its ending when viewing a 3D movie.
Seung-Woo HONG Euisin LEE Ho-Yong RYU Sang-Ha KIM
For monitoring of a large-scale continuous object, a large number of sensor nodes might be participated with object detection and tracking. In order to reduce huge quantities of data from the sensor nodes, previous studies focus on representative selection for data reporting to a sink. However, they simply choose representatives among a large number of candidates without consideration of node deployment environments and detection accuracy. Hence, this letter proposes a novel object tracking scheme that first makes a small set of candidates and then chooses a small number of representatives in the set. Also, since the scheme also considers object alteration for representative selection, it can provide high energy-efficiency despite reducing data reporting.
Ziming HE Yi MA Rahim TAFAZOLLI
This letter investigates the training convergence in range-based cooperative positioning with stochastic positional knowledge. Firstly, a closed-form of squared position-error bound (SPEB) is derived with error-free ranging. Using the derived closed-form, it is proved that the SPEB reaches its minimum when at least 2 out of N (> 2) agents send training sequences. Finally, numerical results are provided to elaborate the theoretical analysis with zero-mean Gaussian ranging errors.
Katsuhiro HORIBA Keiko OKAWA Jun MURAI
On the 11th of March, 2011, a massive earthquake hit the northeast region of Japan. The government of Japan needed to publish information regarding the earthquake and its influences. However, their capacity of Web services overflowed. They called the industry and academia for help for providing stable information service to the people. Industry and academia formed a team to answer the call and named themselves the “EQ project”. This paper describes how the EQ Project was organized and operated, and gives analyses of the statistics. An academic organization took the lead in the EQ Project. Ten organizations which consisted of commercial IT industry and academics specialized in Internet technology, were participating in the EQ Project and they structured the three clusters based on their relationships and technological approach. In WIDE Cluster, one of three clusters in the structure of EQ, the peak number of file accesses per day was over 90 thousand, the mobile browsers was 3.4% and foreign languages (translated contents) were referred 35%. We have also discussed the future information distribution strategies in emergency situation based on the experiences of the EQ Project, and proposed nine suggestions to the MEXT as a future strategy.
Hong Phuc NINH Masaya MIYAHARA Akira MATSUZAWA
This paper considers a simple type of Dynamic Element Matching (DEM), Clocked Averaging (CLA) method referred to as one-element-shifting (OES) and its effectiveness for the implementation of high spurious-free dynamic range (SFDR) multi-bit Delta-Sigma modulators (DSMs). Generic DEM techniques are successful at suppressing the mismatch error and increasing the SFDR of data converters. However, they will induce additional glitch energy in most cases. Some recent DEM methods achieve improvements in minimizing glitch energy but sacrificing their effects in harmonic suppression due to mismatches. OES technique discussed in this paper can suppress the effect of glitch while preserving the reduction of element mismatch effects. Hence, this approach achieves better SFDR performance over the other published DEM methods. With this OES, a 3rd order, 10 MHz bandwidth continuous-time DSM is implemented in 90 nm CMOS process. The measured SFDR attains 83 dB for a 10 MHz bandwidth. The measurement result also shows that OES improves the SFDR by higher than 10 dB.
This report focuses on a design method for gradient index (GRIN) lens antennas with controllable aperture field distributions. First, we derive differential equations representing optical paths in a gradient index medium with two optical surfaces by using geometrical optics, and then we formulate a novel design method for GRIN lens antennas based on these equations. The Levenberg-Marquardt algorithm is applied as a nonlinear least squares method to satisfy two conditions-focusing and shaping the aperture field distribution-thus realizing a prescribed radiation pattern. The conditions can be fulfilled by optimizing only the index (or permittivity) distribution, whereas the shapes of the optical surfaces remain as free parameters that can be utilized for other purposes, such as reducing reflection losses that occur on the surfaces, as illustrated in this report. A plano-concave GRIN lens is designed as an example, applying the proposed method, to realize a sidelobe level of -30 dB pseudo Taylor distribution, and a maximum sidelobe level of -29.1 dB was observed, indicating it is sufficiently accurate for practical use. In addition, we discuss the convergence of this method considering the relationship between the number of the initial conditions and the differential order of the design equations, factoring in scale invariance of the design equations.
Kazuhiko KINOSHITA Yukio ITO Hideaki KIMURA Yuji MAEDA
This letter summarizes three talks in the tutorial session of the 13th Asia-Pacific Network Operations and Management Symposium (APNOMS2011), which focused on the disaster recovery and further emergency management regarding the Great East Japan Earthquake of 2011. We present the damage and restoration of communication networks and points to a future disaster-resilient society.