Sharif Digital Repository

Due to special constraints in peer-to-peer (P2P) networks (such as bandwidth limitation) and because of their highly dynamic characteristics, a single node cannot provide a reliable multimedia stream to the receivers. Several multi-sender algorithms are proposed to reliably deliver a media stream to the receiver through the intrinsically unreliable P2P networks. Based on upload bandwidths and availability of peers as well as the bandwidths of the links connecting the senders and the receiver, PROMISE selects a set of active senders to maximize the expected bit-rate delivered to the receiver. By careful investigation of PROMISE, in this paper, we present why and how it deviates from finding... 

In this paper, we investigate an energy harvesting cooperative network in which the cooperation is done in a random access environment. Although the cooperation provides an extra harvested energy supply for transmissions of the source, it causes probable collisions for the transmitted source packets. Thus, this kind of cooperation can improve or degrade the QoS of the source packets. We find the optimal policy in such a scenario to maximize the source throughput and derive the necessary and sufficient condition for no-cooperation policy to be throughput-optimal. Then, we prove that the maximum throughput is obtained by considering no-cooperation or full-cooperation policy depending on the... 

Network Coding (NC) is an emerging technique that helps wireless networks to have higher throughput and better energy efficiency. In this paper, a novel network coding-aware routing (S-NC) which uses NC to improve reliability and robustness in unreliable WSNs is proposed. It is based on a structure-free routing algorithm as the underlying routing engine which employs neighbor queue length to increase coding opportunity; while exploiting a spatial coding method to control the amount of redundancy. To the best of our knowledge, this is the first study on the topic of network codingaware routing in Wsns. Simulation results in Tossim demonstrate that S-NC significantly improves the performance... 

Visual attributes are great means of describing images or scenes, in a way both humans and computers understand. In order to establish a correspondence between images and to be able to compare the strength of each property between images, relative attributes were introduced. However, since their introduction, hand-crafted and engineered features were used to learn increasingly complex models for the problem of relative attributes. This limits the applicability of those methods for more realistic cases. We introduce a deep neural network architecture for the task of relative attribute prediction. A convolutional neural network (ConvNet) is adopted to learn the features by including an... 

Many forms of diversity have been exploited in communication systems to combat fading nature of wireless channels. Some examples are time diversity, frequency diversity, space diversity and multiuser diversity. However, another form of diversity inherent in any wireless network, which is usually not exploited, is the diversity due to the shared medium of propagation. When in a network one node transmits to the other node, all the other nodes may hear this transmission and some of them can decode the data as well. This data can be used later to enhance the end-to-end network reliability. Previous works mostly consider some special cases of relay nodes configurations, not the general case. In... 

This paper presents a performability estimation approach which measures the ability of a network to deliver its services in the presence of component failures. The measurement is based on two important network service parameters, the endto- end delay and the packet loss. To evaluate the approach, different network topologies are studied using the OPNET Modeler. The results of simulations for four random topologies and two routing protocols, the RIP and the OSPF are presented and compared. The OSPF protocol with different delay timer values is also evaluated. © 2007 Springer  

In this paper, a low energy yet reliable communication scheme for network-on-chip is suggested. To reduce the communication energy consumption, we invoke low-swing signals for transmitting data, as well as data encoding techniques, for minimizing both self and coupling switching capacitance activity factors. To maintain the communication reliability of communication at low-voltage swing, an error control coding (ECC) technique is exploited. The decision about end-To-end or hop-To-hop ECC schemes and the proper number of detectable errors are determined through high-level mathematical analysis on the energy and reliability characteristics of the techniques. Based on the analysis, the extended... 

An approximated quadratic programming optimization is proposed to determine a model predictive controller to guarantee end-to-end delay in the DiffServ architecture. The proposed algorithm, called the suboptimal model predictive control scheduler (SMPCS), regulates the service rates of aggregated traffic classes dynamically, such that some constraints on proportional or absolute delay can be guaranteed. This paper investigates SMPCS complexity and its implementation problems in high-speed routers. The main problem of model predictive control (and one which has limited its use to slow processes) is its complexity. Optimization is the most computationally complex part in a model predictive... 

We present our work on end-to-end training of acoustic models using the lattice-free maximum mutual information (LF-MMI) objective function in the context of hidden Markov models. By end-to-end training, we mean flat-start training of a single DNN in one stage without using any previously trained models, forced alignments, or building state-tying decision trees. We use full biphones to enable context-dependent modeling without trees, and show that our end-to-end LF-MMI approach can achieve comparable results to regular LF-MMI on well-known large vocabulary tasks. We also compare with other end-to-end methods such as CTC in character-based and lexicon-free settings and show 5 to 25 percent... 

Reliability and energy consumption are two of the main constraints in wireless sensor networks (WSNs). In this paper, a novel energy efficient and high speed error control scheme is introduced that is based on the Redundant Residue Number System (RRNS) allowing real-time application of WSNs. The proposed approach employs a new 3-moduli set {22n+1, 2 2n1-1, 2n- 1 } and an efficient reverse converter which relies on the Mixed Radix Conversion (MRC) algorithm and achieves significant improvements both in terms of conversion delay and hardware design. In order to obtain error controllability, two-redundant-moduli set 23n+1 - 1, 24n+1 - 1 is also added to the main 3-moduli set. The theoretical... 

Recently, using Peer-to-Peer (P2P) overlay architectures have become a popular approach for live video streaming over the Internet. Since the peers in these systems are geographically distributed, the communication between them imposes huge redundant traffic into the Internet. One solution to address this problem is utilizing the locality aware algorithms to optimize the neighbor selection policy. However, for real-time applications like live video streaming we need to consider the trade-off between localizing the traffic and service quality. In this paper, we propose a novel pure P2P framework through two phases; (1) Local tracker selection, and (2) Overlay mesh construction. We introduce... 

This paper presents a novel routing protocol based on the Learning Automata method for large scale Wireless Sensor Networks (WSNs) codenamed DRLR (distributed reinforcement learning routing). In this method, each node is equipped with learning automata so that it can learn the best path to transmit data toward the sink. The approach proved to be efficient, reliable, and scalable. It also prevents routing hole by considering network density and average of energy levels available. The approach also increases network lifetime by balancing energy consumption. We compared our approach to two other methods (MMSPEED and EESPEED) and the simulation results show our algorithm to better meet... 

The thermal conductivity (λ) of nanoconfined polyamide-6,6 (PA) oligomers in polymer-graphene nanocomposites has been investigated by reverse nonequilibrium molecular dynamics (RNEMD) simulations. The preferential alignment of the PA chains parallel to the graphene plane as well as their elongation implies that λ of the polymer in nanocomposites is larger than that in the neat polymer system. The ordering of the polymer phase is enhanced in an arrangement of charged graphene surfaces made of one layer with a charge deficit and one with a charge excess. The consequence of the enhanced polymer ordering as well as the denser packing is an increase in λ in the polymer network. Differences in the...