Data generated from the Internet of Things (IoT) devices in smart spaces are utilized in a variety of fields such as context recognition, service recommendation, and anomaly detection. However, the missing values in the data streams of the IoT devices remain a challenging problem owing to various missing patterns and heterogeneous data types from many different data streams. In this regard, while we were analyzing the dataset collected from a smart space with multiple IoT devices, we found a continuous missing pattern that is quite different from the existing missing-value patterns. The pattern has blocks of consecutive missing values over a few seconds and up to a few hours. Therefore, the pattern is a vital factor to the availability and reliability of IoT applications; yet, it cannot be solved by the existing missing-value imputation methods. Therefore, a novel approach for missing-value imputation of the continuous missing pattern is required. We deliberate that even if the missing values of the continuous missing pattern occur in one data stream, missing-values imputation is possible through learning other data streams correlated with this data stream. To solve the missing values of the continuous missing pattern problem, we analyzed multiple IoT data streams in a smart space and figured out the correlations between them that are the interdependencies among the data streams of the IoT devices in a smart space. To impute missing values of the continuous missing pattern, we propose a deep learning-based missing-value imputation model exploiting correlation information, namely, the deep imputation network (DeepIN), in a smart space. The DeepIN uses that multiple long short-term memories are constructed according to the correlation information of each IoT data stream. We evaluated the DeepIN on a real dataset from our campus IoT testbed, and the experimental results show that our proposed approach improves the imputation performance by 57.36% over the state-of-the-art missing-value imputation algorithm. Thus, our approach can be a promising methodology that enables IoT applications and services with a reasonable missing-value imputation accuracy (80∼85%) on average, even if a long-term block of values is missing in IoT environments.

Hardware prefetching involves a sophisticated balance between accuracy, coverage, and timeliness while minimizing hardware cost. Recent prefetchers have achieved these goals, but they still require complex hardware and a significant amount of storage. In this paper, we propose an efficient Per-page Most-Offset Prefetcher (PMOP) that minimizes hardware cost and simultaneously improves accuracy while maintaining coverage and timeliness. We achieve these objectives using an enhanced offset prefetcher that performs well with a reasonable hardware cost. Our approach first addresses coverage and timeliness by allowing multiple Most-Offset predictions. To minimize offset interference between pages, the PMOP leverages a fine-grain per-page offset filter. This filter records the access history with page-IDs, which enables efficient mapping and tracking of multiple offset streams from diverse pages. Analysis results show that PMOP outperforms the state-of-the-art Signature Path Prefetcher while reducing storage overhead by a factor of 3.4.

In this paper, we propose a predictive block-constrained trellis-coded quantization (BC-TCQ) to quantize cepstral coefficients for distributed speech recognition. For prediction of the cepstral coefficients, the first order auto-regressive (AR) predictor is used. To quantize the prediction error signal effectively, we use the BC-TCQ. The quantization is compared to the split vector quantizers used in the ETSI standard, and is shown to lower cepstral distance and bit rates.

In this paper, we consider a direct sequence spread spectrum (DS-SS) microphone system in 2.4 GHz ISM band under indoor environments. It can be modeled as the asynchronous DS-CDMA system in which multiple transmitters share a common channel to transmit information to a single receiver. However, this system has the two critical defects, one is that many burst errors are occurred because of the long-term deep fade in an indoor channel, and the other is that power control is impossible due to the unidirectional structure. Since the reliable performance of DS-CDMA systems can be guaranteed under a reliable bit error rate performance and perfect power control, the solution to these problems is imperative. Hence, in order to overcome these problems, we propose a zero forcing equalizer (ZFE) based multi-user detection (MUD) receiver for this system, since an MUD receiver doesn't require rigorous power control and is able to remove multiple access interference using linear equalization. We derive the modeling and the tap coefficients of the proposed ZFE, and also analyze the performance of the proposed receiver. Based on the performance comparison with the conventional receivers, such as single user detection and post-detection combining diversity receivers, it is shown that the proposed receiver can achieve a lower bit error rate performance.

Feedforward neural networks have been successfully developed and applied in many areas because of their universal approximation capability. However, there still remains the problem of determining a suitable network structure for the given task. In this paper, we propose a novel self-organizing neural network which automatically adjusts its structure according to the task. Utilizing both the constructive and the pruning procedures, the proposed algorithm finds a near-optimal network which is compact and shows good generalization performance. One of its important features is reliability, which means the randomness of neural networks is effectively reduced. The resultant networks can have suitable numbers of hidden neurons and hidden layers according to the complexity of the given task. The simulation results for the well-known function regression problems show that our method successfully organizes near-optimal networks.

It is known that in the selected mapping (SLM) scheme for orthogonal frequency division multiplexing (OFDM), correlation (CORR) metric outperforms the peak-to-average power ratio (PAPR) metric in terms of bit error rate (BER) performance. It is also well known that four times oversampling is used for estimating the PAPR performance of continuous OFDM signal. In this paper, the oversampling effect of OFDM signal is analyzed when CORR metric is used for the SLM scheme in the presence of nonlinear high power amplifier. An analysis based on the correlation coefficients of the oversampled OFDM signals shows that CORR metric of two times oversampling in the SLM scheme is good enough to achieve the same BER performance as four times and 16 times oversampling cases. Simulation results confirm that for the SLM scheme using CORR metric, the BER performance for two times oversampling case is almost the same as that for four and 16 times oversampling cases.

The widespread use and increasing popularity of broadband service has prompted a focus on the measurement and analysis of its empirical performance in recent studies. The worldwide view of broadband performance has been examined over the short term with Speedtest.net, but research in this area has not yet provided a long-term evolutionary insight on how DSL, Cable, and Fiber access technologies have influenced on user experience. In this study, we present 6 years of measurement results, from 2006 to 2011, of broadband performance with fast developing broadband networks in Korea. With 57% Fiber penetration in 2011, our data consist of a total of 29M test records and 10M subscribers. Over the 6 years, we have observed a 2.9-fold improvement in download speed (57Mbps), 2.8-fold increase in upload speed (38Mbps), and 0.7-fold decrease in latency due to the high penetration rate of Fiber broadband service and the advanced Cable modem technology. In addition, we carried out longitudinal analysis of various aspects of services, providers, regions, and cost-performance. We believe that the evolutionary Korean broadband measurement results can shed light on how high-speed access technologies are substantially enhancing on end-to-end performance.

In this paper, we propose a partial erasure decoding scheme with erasure-decision threshold for Reed-Solomon (RS) codes and analyze its performance in frequency-hopped multiple-access communications. RS code is used to correct erasures and errors caused by other-user interference. Binary FSK is employed to transmit the channel symbol. The proposed decoder decides whether to erase the received RS code symbol based on the ersure decsion threshold. The approximated formula for optimal erasure decision threshold is derived in such a way that packet error probability can be minimized. Numerical results show that the employment of adaptive erasure decision threshold attains the higher normalized throughput in the areas of high channel traffic cases.

This letter provides a tight upper bound on the bit error rate (BER) over the Nakagami-m fading channel for the dual carrier modulation (DCM) scheme, which is adopted by the multi-band orthogonal frequency division multiplexing (MB-OFDM) ultra-wideband (UWB) system. Its tightness is verified with the existing result for Rayleigh fading channel, i.e., for m=1, which would be also valid for a more general fading environment.

Online used markets such as eBay, Yahoo Auction, and Craigslist have been popular due to the web services. Compared to the shopping mall websites like eBay or Yahoo Auction, web community-style used markets often expose the private information of sellers. In Korea, the most popular online used market is a website called “Joonggonara” with more than 13 million users, and it uses an informal posting format that does not protect the users' privacy identifiable information. In this work, we examine the privacy leakage from the online used markets in Korea, and show that 45.9% and 74.0% of sample data expose cellular phone numbers and email addresses, respectively. In addition, we demonstrate that the private information can be maliciously exploited to identify a subscriber of the social network service.

In terrain visualization, the quadtree is the most frequently used data structure for progressive mesh generation. The quadtree provides an efficient level of detail selection and view frustum culling. However, most applications using quadtrees are performed on the CPU, because the pointer and recursive operation in hierarchical data structure cannot be manipulated in a programmable rendering pipeline. We present a quadtree-based terrain rendering method for GPU (Graphics Processing Unit) execution that uses vertex splitting and triangle splitting. Vertex splitting supports a level of detail selection, and triangle splitting is used for crack removal. This method offers higher performance than previous CPU-based quadtree methods, without loss of image quality. We can then use the CPU for other computations while rendering the terrain using only the GPU.

In this paper, an effective congestion control algorithm is proposed to increase the end-to-end delivery success ratio of upstream traffic by reduction of buffer drop probabilities and their deviation in wireless sensor networks. According to the queue length of parent and child nodes, each child node chooses one of the parents as the next hop to the sink and controls the delay before transmission begins. It balances traffics among parents and mitigates congestion based on congestion level of a node. Simulation results show that the proposed algorithm reduces buffer drop probabilities and their deviation and increases the end-to-end delivery success ratio in wireless sensor networks.

Tomita's parsing method (GLR) is a practical and successful parsing method for natural language. However, one difficulty in the GLR is that interleaved constraint processing of syntax and semantics in parallel is not trivial during parsing, because it uses the precompiled table for a fast real-time parsing. In this paper, we present a method which makes the GLR adaptable to interleaved parsing while making some limitation on its generality. For interleaved parsing, the conflicts of the LR parsing table must be resolved at the parse time. The shift-reduce conflict among the above conflicts is the most serious one for interleaved parsing because of the lack of knowledge for the conflict resolution at the parse time. Therefore, we concentrate on resolving a shift-reduce conflict by introducing a grammar which is called a shift-first LR (k) grammar. Our method for this is that the conflict resolution is delayed by the shift-first strategy which makes an unconditional choice of shift actions in the case of a shift-reduce conflict. Then, a delayed resolution that resolves the conflict, is made. Depending on the decision of the resolution, the pseudo parsing, which parses symbols in the LR parser stack, proceeds. Our experiments showed that our parser is efficient while attaining the interleaved parsing at real time.

This letter introduces an Energy-Aware LED Light System (EA-LLS) that provides adequate illumination to users according to the analysis of the sun's position, the user's movement, and various environmental factors, without sun illumination detection sensors. This letter presents research using algorithms and scenarios. We propose an EA-LLS that offers not only On/Off and dimming control, but dimming control through daylight, space, and user behavior analysis.

In this paper, we discuss both effective approaches in specification process, formal specification and reuse, and focus on providing an integrated and systematic supportbased on them. Preparing the specification model which mediates an image of the designer and another representation of it in formal method, the designer can specify the target system incrementally and smoothly. As for the specification model, we employ LTS on the early step of specification process because of its understandability for the designer. Moreover, reuse of specification leads to reduction of the cost and time, defining retrieval mechanism of reusable cases from database by mathematically calculating similarity of them. For the reuse mechanism, we define a new concept of similarity on LTS as the criterion of case retrieval, which enables more flexible matching between the besigner's requirement and the existing case than any other traditional schema on LTS, and show the case retrieval algorithm. Integration of two approaches brings us the great improvement of the productivity on system development.

This paper presents an effective application of Net-theory for all the stages of the communication protocol development process. Net-theory provides a basic mathematical model and tool for development of communication protocol. The special usability of Net-theory is that 1) visual representation of the system's stadic/dynamic structure, so that users may easily understand the represented contents, 2) formal specifications based on mathematical basis of Net-theory admit automatic verification, implementation and conformance testing. We have seen that Net-theory which has the above usability can provide a systematic and advanced paradigm for effective communication protocol development.

We provide a BER analysis of the well-known rotation code, which has been applied to various transmission schemes, such as coordinate-interleaved design (CID) for space-time block codes and trans-modulation or constellation rearrangement for a cooperative relay system. An upper bound on the BER of the rotation code under a fading channel is derived. It turns out to be much tighter than the existing one, which relies on the Chernoff bound. More specifically, the proposed bound is virtually identical to simulation result when Eb/No is larger than 4 dB.

Nowadays, a user authentication is very important in network environments. For safe authentication, they came up with six essential conditions in earlier studies. And a variety of mechanisms is presented by research scientists. However, they could not achieve the PFS. Because, though all these schemes are assumed that the communication between a smart card and a host is safe, actually it is not. Therefore, in this paper, we will point out what the communication between a smart card and a host is not safe, and propose a new user authentication mechanism that can reach to the PFS. And also, an encryption algorithm is used about 45% less than earlier studies in our proposed scheme. Thus, we can say that enhance the efficiency.

A power analysis attack is a well-known side-channel attack but the efficiency of the attack is frequently degraded by the existence of power components, irrelative to the encryption included in signals used for the attack. To enhance the performance of the power analysis attack, we propose a preprocessing method based on extracting encryption-related parts from the measured power signals. Experimental results show that the attacks with the preprocessed signals detect correct keys with much fewer signals, compared to the conventional power analysis attacks.

Many simulated annealing algorithms use the Cauchy neighbors for fast convergence, and the conventional method uses the product of n one-dimensional Cauchy distributions as an approximation. However, this method slows down the search severely as the dimension gets high because of the dimension-wise neighbor generation. In this paper, we analyze the orthogonal neighbor characteristics of the conventional method and propose a method of generating symmetric neighbors from the n-dimensional Cauchy distribution. The simulation results show that the proposed method is very effective for the search in the simulated annealing and can be applied to many other stochastic optimization algorithms.