Abstract

Internet have experienced an explosive growth accompanied with sever congestion problems. TCP congestion control mechanism is crucial for efficient use of the Internet despite largely unpredictable user access patterns and despite resource bottlenecks and limitations. This paper reviews and evaluates the current Reno and Vegas TCP congestion control mechanisms. Generally, the mechanism repeatedly increase load in an effort to find the flow rate that digests the different demands of the end devices, however, each attain different utilization of Internet resource.  Further, the authors present a new controller based on fuzzy concept to overcome the limitations of the current Reno TCP protocol. The fuzzy control in this work is a static algorithm with static rule. All the evaluation data are gained from NS2 simulator.

Abstract

The main objective of this research paper is designing automatic fuzzy parameter selection based dynamic fuzzy voter for safety critical systems with limited system knowledge. In this research paper, existing fuzzy voters for controlling safety critical systems and fuzzy voters used for sensor fusion are surveyed and the major limitation identified in the existing fuzzy voters, is the static fuzzy parameter selection. Static fuzzy parameters work only for a particular set of data with the known data ranges for which optimized values are selected for fuzzy parameters. These values may not work for other sets of data with different ranges. The static fuzzy parameter selection method may not work for continuously changing different ranges of data. In this paper a dynamic or automatic fuzzy parameter selection method for fuzzy voters is proposed based upon the local set of data in each voting cycle. Fuzzy bandwidth is decided based upon the statistical parameters like mean of the local data set and standard deviation and fuzzy parameters are updated to decide the fuzzy bandwidth in each voting cycle. Safety performance is empirically evaluated by running the static and dynamic fuzzy voters on a simulated Triple Modular Redundant (TMR) system for 10000 voting cycles. Experimental results shows that proposed Dynamic fuzzy voter is giving almost 100% safety if two of the three modules of the TMR System are error free and also giving better safety performance compared to the existing static fuzzy voter for multiple error conditions. Dynamic voter is designed in such a way that it can be just plugged in and used in any safety critical system without having any knowledge regarding the data produced and their ranges, as it processes the data locally in each voting cycle using statistical parameters.

Abstract

This paper presents a comparison between the important transforms using in speech signal such as (Fast Fourier Transform, Discrete Wavelet Transform, Wavelet Packet Transform and Discrete Wavelet-Fast Fourier Transform) and try to compression the coefficients that received from any transforms by using zero's coder before using arithmetic coding  algorithm to compress speech signal. We find the DWT and WPT are the best transform in the speech signal analysis and using the zero's coder increase the compression ratio and using DWT-FFT transform make the compression bad because the coefficients received after using FFT transform are different and make the symbols that entry to arithmetic coding more. Analysis of the compression process was performed by comparing the compressed-decompressed signal against the original.

Abstract

This paper proposes a generic font description model for optical character recognition. It is based on concept of evolutionary computing architecture. This is user interface application software which works on learning and recognizing the handwritten text of the particular user. The project allows user to write his command for the computer on blank paper and control the operations of the computer via conversational creature. The purpose is to design an easy interface with the computer for computer illiterate persons. Text written by the user will be available to computer for further processing like text editing, narrating, and messaging.

Proposed system has manifold application in government offices for storing lacks of files of record, in business meetings for maintaining review of discussion, in educations for converting professors’ notes into soft copies etc. Blind people can be highly benefited from this system as it supports narrator application. We proposed the algorithm to avoid the ambiguity. The system is tested by giving the handwritten character and sentences of different user. The system recognizes all character and sentences of all users correctly. The user written command on the paper are also recognized and executed by the system. It is found that the performance of the system is approximately is equal to 92%. This system is cost effective as it requires very less hardware support like camera or scanner.

Abstract

The transmission of data over the Internet increases, the need to protect connected systems also increases. Although the field of IDSs is still developing, the systems that do exist are still not complete, in the sense that they are not able to detect all types of intrusions. Some attacks which are detected by various tools available today cannot be detected by other products, depending on the types and methods that they are built on. Data stream classification (DSC) Detection System Using Genetic Algorithm (GA) is the latest technology used for this purpose. The behavior of the genetic algorithm, a popular approach to search and optimization problems, is known to depend, among other factors, on the fitness function formula, the recombination operator, and the mutation operator. In real-world data stream classification problems, such as intrusion detection, text classification, and fault detection novel classes may appear at any time in the stream. Traditional data stream classification techniques would be unable to detect intrusion until the classification models are trained with Genetic algorithm. We applied this technique in network traffic. The network intrusion detection system should be adaptable to all type of critical situations arise in network. This is helpful for identification of complex anomalous behaviors. This work is focused on the TCP/IP network protocols. Data stream classification pose many challenges, some of which have not been addressed yet.

Abstract

In recent years, many trust and name models are proposed to reinforce the protection of mobile ad hoc networks. However, they either fail to capture proof of trustworthiness at intervals the restrictions of the network, or introduce further issues whereas capturing the proof. In this paper, we have a tendency to propose a reputation-based trust model referred to as Secure MANET with Neighbor Collaboration Routing (SMNCR). In our model, the proof of trustworthiness is captured in an economical manner and from broader views as well as direct interactions with neighbors, observing interactions of neighbors and thru recommendations. SMNCR captures proof from direct interactions with neighbors so as to spot their benign and malicious behaviors. It conjointly captures proof for misbehavior by observing the interactions of neighbors. Lastly, the proof captured from recommendations is employed to summarize the benign behavior of multi-hop nodes. In contrast to alternative models, we have a tendency to adopt a completely unique approach to capture proof from recommendations that eliminates recommender’s bias, free-riding, and honest elicitation. SMNCR utilizes the captured proof to predict whether or not a node is either benign or misbehaving. It then applies the prediction to reinforce the protection of communications looking on the choice policies, like whether or not to send a packet to or forward a packet on behalf of alternative nodes. Finally, we have a tendency to demonstrate the performance of our model through simulation results.

Abstract

In mobile ad-hoc nodes change position due to the dynamic nature. There has to be a proviso to control the performance and place on standard basis. In this paper, the importance of management plans in ad-hoc networks is studied. Beside this, mobility models are reviewed and rated by the incorporation of real-life applications. It is explored a model for the operation of an ad hoc network and the effect of mobile nodes, where the model incorporates incentives for users to act as transit nodes on multi-hop routes and be pleased with their own ability to send traffic. In this paper, it is explored the implications of the model by simulating a network and illustrates how network resources are allocated for the users based on geological position. Mobile nodes which have incentives to work together also discussed in this paper.  Mobility and traffic pattern of mobility models are generated by using AnSim Simulator.