In mobile communications, MIMO-OFDM transmission performance suffers severe degradation caused by the large delay spread channel greater than guard interval (GI). This is because the excess delay... Show More

In mobile communications, MIMO-OFDM transmission performance suffers severe degradation caused by the large delay spread channel greater than guard interval (GI). This is because the excess delay results in considerable inter-symbol interference (ISI) between adjacent symbols and inter-carrier-interference (ICI) among subcarrier in the same symbol. In case of scattered pilot (SP), the interference of pilot signals causes the deterioration of channel estimation. To mitigate this problem, in this paper, we propose the interference compensation scheme using the time domain replica signals. We make the time domain replica signals from detected signals and the excess channel impulse responses over GI. After compensation of the time domain replica signals and the received signals, we recalculate the channel state information (CSI) and the CSI is updated. Finally, we carry out the channel compensation with updated CSI for obtaining accurate compensated signals.

In recent, orthogonal frequency division multiple access (OFDMA) has been used for a multiuser wireless communication. In a wireless network, the transmitted signal of each user has independent... Show More

In recent, orthogonal frequency division multiple access (OFDMA) has been used for a multiuser wireless communication. In a wireless network, the transmitted signal of each user has independent channel fluctuation characteristic. By using this characteristic, OFDMA can achieve the multiuser diversity (MUDiv). Until this time, to achieve a low complexity and performance improvement, the adaptive subcarrier block (ASB) and frequency symbol spreading (FSS) methods have been proposed. However, the system performance in a low Eb/N0 is worse than that of maximal sum capacity (MSC) and peak to average power ratio (PAPR) does not decrease greatly. To solve these problems, in this paper, we propose the subcarrier allocation with the power priority selection (PSS) and the adaptive clipping (AC) with the peak reduction signal to improve the system and PAPR performance.

This article studies the characteristics of content on video sharing websites. A better understanding on online video content can help to analyse Internet users' behaviour and improve the... Show More

This article studies the characteristics of content on video sharing websites. A better understanding on online video content can help to analyse Internet users' behaviour and improve the video-sharing service. We improved an existing graph-sampling algorithm so that it could be more adapted to sample over the video sharing websites. A newly category system is defined in this paper, which can be applied on many different video sharing websites for content analysis. We also implement machine learning to realize the content re-classification with the newly defined category system. The efficiency reaches at 90%. From the classified content analysis, we find the content category distribution is not constant, and nowadays, cultural goods content take about 70% over all the sampled videos.

We aim at creating a robust and efficient custom routing mechanism for Distributed Shortcut Networks (Nguyen et al., 2013), which address new challenging issues posed by recently advanced studies in... Show More

We aim at creating a robust and efficient custom routing mechanism for Distributed Shortcut Networks (Nguyen et al., 2013), which address new challenging issues posed by recently advanced studies in the areas of massively parallel computing and large-scale data centers. We follow the design principles of Distributed Shortcut Networks (DSN), which construct non-random topologies with the creation of long-range shortcuts inspired by observations in small-world networks. However, we focus on designing a powerful custom routing mechanism which smartly exploits some precious properties of the topology. As a result, our new DSN-a network with a carefully refined routing logic performs significantly better than the basic DSN in term of communication latency while provides strengths in fault-tolerance as well as load-balance. These help the network become robust against link failures or burst of traffic demand while topology-agnostic deadlock-free routing (e.g. the famous up*/down* algorithm) suffers a lot.