Rate-Optimal Scheduling of DSP Applications onto Integrated Heterogeneous Multiprocessors
Abstract:
A technique for scheduling and processor allocation of digital signal processing applications on integrated heterogeneous multiprocessor, is proposed. The proposed technique achieves efficient hardware implementations at the logic-level by minimizing the number of processing units used, without compromising other optimality criteria, namely, the rate and delay optimality. The proposed technique is composed of two algorithms: the time scheduling and the processor assignment. In the time scheduling algorithm, the relative firing times of all the operations of the given application are found. The processor assignment algorithm finds a matrix for each processor type. A matrix associated with a processor type indicates the sequence of operations on each processor of this type
back to Accepted Papers Page
On The Algorithmic Issues of the Arrangements-Star Network
Abstract:
In this paper new evidences are presented to show that the arrangement-star network, which is constructed from the product of the star and arrangement graphs is attractive alternative to both star and arrangement graphs. The arrangement-star network is shown to be vertex symmetric furthermore the hierarchical structure of the arrangement-star network is discussed and proved. A major contribution of this paper is the development and evaluation of efficient frameworks for developing parallel algorithms on the proposed network. The present study shows that the algorithmic framework enables the new network to provide efficient support for an important class of problems that are based on grid and pipeline views.
Index terms ¾ Star Graph, Arrangement Graph, Product Network, Hierarchical Structure, Vertex Symmetry, and Parallel Algorithms.
back to Accepted Papers Page
Non-Linear Clustering Algorithm for Synchronous Communication with Deadlock Avoidance Strategy
Abstract:
In this work, we are interested in developing an efficient heuristic algorithm for scheduling the tasks of a parallel program based on the class of UNC (Unbounded Number of Clusters) scheduling algorithms for clusters of NOWs. The main objective of the proposed UNC algorithm is to consider synchronous communication with deadlock avoidance strategy, for intertask message passing. The proposed algorithm generates non-linear clusters by traversing the task graph (DAG) once, using the Edgo-Zeroing (EZ) technique. The objectives of the clustering algorithm is reducing the parallel time of the program, reducing the communication cost, improving the Program computation to Communication Ratio (PCCR), and avoiding deadlock situations. The algorithm achieves its objectives with a time complexity
O (|V|(log |V| + (|E|) 2)) using nonlinear clustering in order to avoid more than one pass through the task DAG and to be able to deal with task DAG’s with fine, medium, and coarse granularity.
back to Accepted Papers Page
On the Connectivity Conditions in a Parallel Heterogeneous Systems
Abstract:
In this paper, an analytical analysis of heterogeneous parallel systems is introduced. We study the benefits of connecting systems to obtain the best response time. An inductive study has been conducted to determine a mathematical condition that serves to take an optimal decision to connect two parallel heterogeneous nodes. The continuous Markov chain has been used. A state diagram has been obtained, and stability equations have been designed. The average response time is computed and treated as our performance measures. A practical experiment has been performed.
back to Accepted Papers Page
Modelling Jitter in a Mirroring System
Abstract:
Mirroring, or running a backup system parallel to the main system, offers an enhancement of the system reliability. In case of asynchronous operation of the backup the lagging behind of the back up system may arise. In this paper we focus on modeling of delays in the backup system relative to the main. The mirroring aspect of the system is modelled analytically with a difference queue. The proposed difference queue, which is defined for a network of two queues in parallel, is the queue consisting of elements present in one queue and absent from the other queue. The queuing models under investigation are M/M/1, and M/D/1 queues.
back to Accepted Papers Page
Evaluation of New Information Technology: An Empirical Study
Abstract:
An empirical study of the evaluation of new information technology is presented. The study aims at identifying and prioritizing the most critical factors in evaluating new information technology according to information technology professionals. Information technology refers to hardware, software, networks and telecommunication technology, and supplier services. A random sample of information technology professionals was surveyed, statistically analyzed, and reported.
back to Accepted Papers Page
Spatial Bit-Sliced: A Spatial Index Structure for Processing Direction Operations
Abstract:
In this paper we present an efficient index structure to access spatial data. The proposed index, called SBS (Spatial Bit-Sliced), based on Bit-Sliced indexing method. The Bit-Sliced index has proven to be an efficient data structure for indexing one-dimensional data. The Bit-Sliced is an extension of bitmap indexing used to access data elements with high cardinality. Bitmap indexing improves I/O performance as well as storage saving by using single bits instead of multiple bytes of data to indicate a specific value of data. In this paper we present algorithms using SBS for performing some of the direction operations: South and Southeast. To measure the performance of SBS for doing the above operations, we present experimental studies using TIGER data files
back to Accepted Papers Page
VERIF: An Automatic Specification and Verification Tool for Industrial and Reactive Systems
Abstract:
This paper describes an approach towards verification of industrial real-time systems and protocol communications. We have implemented a software package for developing control programs for industrial reactive systems called VERIF. The main characteristic of VERIF package that it can be used in a virtual distributed and web-based environment connected via the Internet. Experimentation using VERIF system didn’t at all need large experiences from the users during the analysis requirement and specification life cycle. The system integrates modularity and
Abstract:
ion and follows the main principles of an Object-Oriented approach inspired by Java programming language and Java Remote Control Interface. Also, the system offers some tools to easy express timed and temporal constrains directly on the finite state machine graph or by using the expression of the behavior on the rewriting rules
back to Accepted Papers Page
Mobile Report Navigation Using Fuzzy Logic
Abstract:
This paper studies the mobile robot navigation in an unknown environment. The navigation algorithm ensembles the actions of a human driver such as goal seeking, obstacle avoidance, U-turn, and getting rid of deadlock. The action taken by the robot at any time depends on the local configurations of the sensors. Except the U-turn action, all other actions are taken using a fuzzy inference system. The getting rid of deadlock action prevents the robot from being in some of the situations in which it will be deadlocked within a certain area. Simulation results prove the effectiveness of the proposed approach.
back to Accepted Papers Page
Intelligent Automotive Volume Control
Abstract:
Advances in home and automotive and entertainment system technology have made it possible to offer exceptional sound quality reproduction to consumers. In recent years, automotive audio systems’ quality has approached that of home systems. However, high and variable ambient noise levels remain difficult to overcome and impose a major limitation on the automotive sound system. Currently, some of these systems use classical controllers, which are based on the estimation of the ambient noise using the speed of the vehicle. Due to the effects of road and weather conditions and other factors such as the status of the windows or sunroof, these systems do not provide excellent noise compensation. Other systems measure the noise levels inside the vehicle and use complicated circuitry and algorithms for compensation. This paper introduces a hybrid approach for automatic volume control in automotive systems. Our system utilizes the simplicity of the speed-based systems and the effectiveness of the ambient-noise-based systems. We have conducted several experiments on real audio signals and the results prove the effectiveness of the proposed approach.
back to Accepted Papers Page
Scalable Automotive Processing Systems
Abstract:
The number of automotive functions that are controlled by computers is rapidly increasing. In the past, these functions were confined to stand-alone control units, such as the engine or ABS brake controller. In modern cars the data exchange between different controllers has grown to a level where one or several networks are necessary to meet the communication demand [1]. Currently, automotive electronics processing is distributed in variety of functions like Power train, Air bag, Climate control, Suspension, etc. The purpose of this paper is to review related work through two state of the art models, and propose a scalable design using parallel processing and fault tolerance techniques (Excluding the functions of power train and transmission control electronics). The integration is based on the logical bisection of vehicle in four symmetric regions, which is critical for hardware, software, fault tolerance and ease of serviceability design. The proposed design can be implemented on any vehicle without going through the iterative system and module design cycles.
back to Accepted Papers Page
Recent Applications in Neuro Fuzzy
Abstract:
The reasoning feature of the fuzzy logic has been incorporated into the learning feature of neural network so as to provide a NeuroFuzzy system that deals with cognitive uncertainties in a manner more like humans. Many NeuroFuzzy-based applications have been reported in areas including expert systems, medical diagnosis, control systems, pattern recognition, target detection, material processing, and system modeling. In addition, the NeuroFuzzy approach has been applied to other problems including short-term water demand forecasting and short-term load (electric power) forecasting. Moreover, a weight is generated and fed to the forecasters to take into account the various factors that might influence the load demand The preliminary results of the NeuroFuzzy approach to these important problems are encouraging and suggest that these models are capable of achieving a high degree of accuracy as will be presented in this paper.
back to Accepted Papers Page
Predicting Groundwater Quality Using AI Techniques
Abstract:
This work addresses the assessment of groundwater quality in the Sultanate of Oman based on techniques from the field of Artificial Intelligence. In this particular case, the first task is the interpretation of physico-chemical and biological data into groundwater quality terms. The second task is to develop groundwater quality models and computer systems to assess and predict the impact of pollutants on the water column. Industry experts have been arguing that the existing models and systems are not reliable for measuring groundwater contamination because they neglect the characteristics that are significant in understanding of pollution-generation processes from various sources. The purpose of this work is; therefore, to present a unified approach to the analysis and interpretation of groundwater quality data in order to determine if qualitative or quantitative standards for groundwater quality have been exceeded. A well-known Artificial Neural Networks, SOM, and a clustering technique developed for use in the groundwater quality classification. We are also developing another approach that is to use a graphic representation of a probabilistic distribution to represent the static and dynamic cause-and-effect relationships between groundwater quality constituents.
back to Accepted Papers Page
Information System for Forecasting Process Using Neural Networks
Abstract:
The electric load demand has a dynamic nature and influences by many factors, such as weather, economic and social activities and different load components (residential, industrial, commercial etc.). By analysis of only historical load data, it is difficult to obtain accurate forecasting for load demand. An Information system for forecasting processes based on unsupervised, supervised neural networks is proposed. The unsupervised learning process is performed using Kohonen’s Neural Networks (KNN) for clustering of the input space into affinity number of classes. For each class, the supervised learning process is performed using Feedforwared Neural Networks (FNN).
In this paper, KNN is used for data classification to identify the day types/classes, which are essential for forecasting processes. The unsupervised process performs the role of front-end data compression. The historical database that contains the data of the attributes of the forecasting process that cover two years is developed using the recorded actual data collected from Libya energy system. From the results, we found that KNN is very effective in day type classification and has the capability to identify new classes
back to Accepted Papers Page
Application of Neural Network for Electric Voltage Control
Abstract:
In the paper is proposed the monitoring of mathematical model of electric voltage source with using neural network for application in control systems as sensor and command signal. The monitoring system, consist of toroidal choke or transformer with high saturated ferromagnetic cores. The input information we receive from current periodic curves. The current is distributed into Fourier or Walsh series. The combination of these harmonics and their amplitude values determine monitoring voltage value directly. For increase of this system precision the mathematical model is constructed on basis of partial differential quasi – stationary electromagnetic field equations and ordinary differential electromagnetic circuit equations combination.
back to Accepted Papers Page
Neural Network Based Mathcoprocessor
Abstract:
This paper presents a proposed design of a math coprocessor. A related architecture is organized and illustrated. Concept of design is depicted in convenience with traditional bases of control unit structure. The proposed design is relayed on back propagation model of neural networks. On driving three functions in an example, the concerned learning patterns are presented. For these patterns, the whole connection scheme is defined and determined.
Linear saturating activation function is envisaged to be a proper characteristic that models the processing element of the networks. Thus, Sl-CONE model is intruded in the learning phase. Performance measurement is discussed in terms of speedy response and structure improvement.
back to Accepted Papers Page
Supervised Classification by Modular Neural Network
Abstract:
In this paper, multi-layer feedforward neural networks are designed to perform pattern recognition tasks. To approach the functionality of the human visual system, we propose a hierarchical system using highly structured artificial neural nets comprising two stages of processing for machine vision: a low-level process and a high-level one. The neural network system contains mainly two modules: a neural (sub-) network performing edge detection, and a neural classifier. This classifier uses the edge image obtained at the output of the first stage of processing. The use of two sub-networks instead of a single one allows the application of learning sets consisting of a limited number of training patterns, and consequently results in reduced learning time. In order to have good generalization, training is made by means of the backpropagation algorithm. This system performs accurately and efficiently even with degraded test images.
back to Accepted Papers Page
Representing Multimedia Languages Using Higher Order Petri Nets
Abstract:
In this paper we use the Higher-order Petri Nets (HOPN) to represent multimedia languages. The firing rule for HOPN is applied to represent complex multidimensional sentences. HOPNs are also used to construct a powerful index for multimedia applications.
back to Accepted Papers Page
Parallel Graph Reduction Using Petri Nets
Abstract:
Functional languages have a good potential of implicit parallelism due to the referential transparency. This potential has not been used efficiently in most applications because it must be specified explicitly how processes communicate and interact. Graph reduction is one of the models used to construct an implementation of functional languages. The reduction graph exhibits the parts of a program, which have to be reduced simultaneously.
This paper presents a formal method to specify communication and verify synchronization between concurrent functional processes. It separates the aspect of graphical specification of communication from the aspect of evaluation of individual processes.
back to Accepted Papers Page
Using Higher Order and Quantized Hopfield Neural Network for Vector Quantization
Abstract:
Vector Quantization (VQ) is a pattern classification method in which each set of vectors represents a particular class or category. Many algorithms and techniques were used to perform VQ. These techniques could be classical techniques (such as k-means algorithm), genetic, or neural network techniques. This work is an attempt to study the performance of using higher order and Quantized Hopfield Neural Network as an approach to perform vector quantization. A comparison was made between various quantization techniques (K-mean, Standard, Higher order, and Quantized Hopfield neural nets) to consolidate the study.
The main problem with Quantized Hopfield Net is that, its performance is affected by the codebook initialization, while the main problem with standard Hopfield nets was the choosing of the appropriate constant values (that are multiplied by the penalty terms). Such a problem leads to the appearance of blocks that are not mapped by any codebook index (odd blocks). To solve the above two problems, a new method was suggested which makes use of the benefits of both Standard Hopfield (provide automatic codebook initialization), and Quantized Hopfield (to get rid of odd vectors).
back to Accepted Papers Page
Symbolic Simulation: Toward and Approach Using Prepositional Logic Integer Programming and Neural Networks
Abstract:
This paper reports on the application of Artificial Neural Networks (ANN) to reasoning in real-world domains such as Natural Language understanding and Logic techniques. Such problems are modeled as an arbitrary system of rules, which are then transformed into a system of Boolean value constraints. In the most case, the solution of such problems is very hard to be computed, so combining new methods are required. In this topic we show that integer programming and neuronal networks techniques are well suited for Hardware/Software verification Systems
back to Accepted Papers Page
A Two- Phase Method for Selecting the Best Simulated System
Abstract:
We consider the problem of selecting the simulated system with the best-expected performance measure when the number of alternative systems is large. In this paper, we use a two-phase procedure that uses the standard clock simulation technique for improving the computational efficiency. In the first phase, we screen out non-competent alternatives and construct a confidence set that contains the best alternative with a pre-specified precision. The second phase uses ranking and selection statistical procedure to select the best-expected alternative among the survivals of the first stage. The standard clock simulation technique is used for simulating multiple points using only one sample path. We implement this algorithm for solving a practical simulation optimization problem. The numerical results indicate that this method is promising.
back to Accepted Papers Page
Statistical Measurements of Message Sizes on Parallel Computational Fluid Dynamic Programs
Abstract:
Computational Fluid Dynamics (CFD) programs are a significant consumer of processor time on high-performance computers. Even two-dimensional CFD problems require a significant amount of computing resources to solve. With the desire to move to three-dimensional CFD problems even more computing resources will be required. To improve the performance of CFD programs effort to port the program to parallel computers have been made.
Little information on the communication characteristics, e.g. message size and frequency of messages, of parallel CFD programs should be made available so that designer of computer systems can better optimize parallel computers to support parallel CFD programs. This paper describes the statistical measurements on message sizes observed in a prototypes two-dimensional parallel CFD program. One key result observed is that the average message size is about 2.2KB, considerably smaller than the messages required obtaining peak performance on most messages passing computers.
back to Accepted Papers Page
Algebraic and Numeric Invariants of the Liner Codes
Abstract:
The theory of the codes in general concerns the information theory. Its objective is centered on the study of the protection of numerical information. We consider the problem of transmitting a message over a channel which can be affected by ” noise”, that is to say, during transmission of a message, exterior disturbances can modify the contents of the message sent .To detect and to correct these errors suitable codes are worked out (error-correcting codes, ).
The most used codes are the linear codes which are vector subspaces of dimension k of the vector space , one thus takes as elementary message a vector line , the elements are called check symbols. This concept is also used in cryptography, for example McEliece, , is a public key cryptosystem based on the complexity of syndrome decoding problem and the equivalence codes. In this note, we determine, for one (n, k) linear code, four invariants algebraic (two subgroups of the symmetric group and two subgroups of the linear group ), which makes it possible to test the equivalence of two linear codes and to calculate the number of codes equivalent to a given code.
With this intention we generalize the theorem of Frobenuis on the double classes of a finite group:
If and are two subgroups of a group G: .
We apply the obtained result to the linear codes: to any (n, k) linear code we associate its generating matrix , in addition and operate on respectively by S.G = SG, P.G = GP, denote by and the relations of equivalence associated respectively with the two preceding operations, as they permute , their product is an equivalence and this relation is the equivalence of the linear codes. In this note we build also numerical invariants associated with a class. With this intention we distinguish two cases, in the first we suppose that the characteristic of the field is different from 2 and then we use the determinant . In the other case we use the quadratic forms. The two obtained invariants give a necessary condition for the equivalence of two codes.
back to Accepted Papers Page
Scalable Dynamic Tree Quara Algorithm
Abstract:
Mechanisms that ensure resilience to failures in large distributed systems can be based on quorum consensus, but effectiveness and scalability of the method applied is crusial. The aim of this paper is to present a message/time cost analysis of the algorithm based on the tree quorum, which needs no global information, and has message complexity independent of the load. Results obtained for different request’s indentifications (topological priority, timestamps, and group priority) are presented.
back to Accepted Papers Page
Web-Based Online Registration for the Student Information System at SQU
Abstract:
In this paper, we discuss the development issues of a dynamic web-site for online registration for the Student Database Information System (SDIS) at Sultan Qaboos University. The database system incorporated in the web-site that supports various activities related to the student database information system including electronic remote registration of various courses, add and drop activities, browsing/searching degree plan, providing guidelines for registration preparation, and automatic web-page generation for the currently running courses. In developing the proposed web-site, we use a number of tools including: Oracle-JDeveloper 2.0, Java Database Connectivity, SQLJ, Front-page 2000, HTML, Symantec Visual Café, Symantec Visual Page, and Oracle 8i.
back to Accepted Papers Page
Freeman Arabic Classification Tree (FACT) for Arabic Character Recognition System
Abstract:
This paper describes the design and implementation of a system that classifies and recognizes machine-printed Arabic characters with prior segmentation using a new suggested simple classification tree called Freeman Arabic classification tree (FACT) .The technique is based on describing symbols in term of shape primitives derived using Freeman chain algorithm (Fives Paired Freeman Chain FPFC)[1]. A heuristic segmentation algorithm is initially used to segment each word. This algorithm is based on describing Arabic words and sub words in term of direction primitives using Freeman chain technique to obtain an enhanced fives pair freeman chain as a feature for each isolated character [1]. At the recognition time, the FPFC primitives (for characters, holes, and dots) or features are used in matching stage. The detected FPFC, for both Arabic characters under recognition and prototypes are classified according to the suggested classification tree (FACT). In this FACT, the Arabic characters are classified into groups according to the presence of hole, dots, and FPFC features. The advantage of using FACT versus lookup table approach used in [1] is that the result of recognition is optimized with regard to the FACT approach. This approach deals with the problem of dot and holes related to Arabic characters in completely different way than that used in [1]. Results of preliminary experiments using this approach shows a recognition rate of 98% for noise free Arabic text and 88% for scanned Arabic text based on 2200 Arabic words.
back to Accepted Papers Page
A Framework for an Arabic Information Extraction System
Abstract:
The purpose of this paper is to provide an overview of the issues involved in Arabic Information Extraction. It wills primarily focus on the framework for an Arabic information extraction system. This system is built around an Arabic Natural Language Processing module where morphological analysis plays a significant role. The output of the core NLP system can be used in many applications such as database filling, finding names of people, places, and organization, and text summarization.
back to Accepted Papers Page
Arabic World Stemming Using Letter Successor and Predecessor Variety
Abstract:
This paper considers the possibility of adapting the “letter successor variety” method, as described in the literature by Hafer and Weiss [11], to the automatic stemming of Arabic words. Based on the statistical properties of letter successor varieties, this method is expected to automatically segment words into their basic constituent stems and affixes. Some modifications had to be made to the original method to conform to the special features of Arabic textual words. The results presented in this paper show that letter successor/predecessor varieties approach provides about 84% level of correctness and about 40% level of stem clustering. There are several problems associated with this method that make it inefficient, from the complexity and recall perspectives, for application in real-life IR systems that are characterized by a high rate of corpus change.
back to Accepted Papers Page
Automatic Text Summarization Approach for Arabic Newspaper Articles
Abstract:
In this paper we describe and evaluate an enhanced statistical approach for automatic text summarization for Arabic documents. In this approach, we assign scores to sentences in a document, based on part of speech and statistical information. The word weights, combined with the presence of highly important nouns, are used in assigning a measure of importance to a specific sentence. Sentences with higher importance scores are then extracted to form the document summary. The performance of our method is evaluated using the word weights approach as a base line. The results obtained are very promising.
back to Accepted Papers Page
An Object-Oriented Arabic Lexicon Suitable to Natural Language Understanding Problem
Abstract:
This paper suggests a computational model for a knowledge-based Arabic Lexicon. The proposed model depends on adapting concepts from three linguistic semantic theories: Semantic markers and selection restrictions theory (Katz and Fodor 1963), Case Grammar theory (Fillmore 1968) and Semantic fields theory (Trier 1931) to build a fully semantic lexicon entry suitable to NLU systems. Moreover, the paper exploits the facilities of object-oriented programming language, such as encapsulation, inheritance, and polymorphism to achieve economical modularized and consistent lexicon organization.
back to Accepted Papers Page
A Hybrid Approach for Arabic Literal Amounts Recognition
Abstract:
The challenge of hybrid learning systems is to use the information provided by one source of information to compensate information missing from the other source. The neuro-symbolic combination represents a promising research way. The synergy between the symbolic (theoretical) and neural (empirical) approaches makes their combination more effective than each of them used alone.
In this article, we describe an Arabic literal amount recognition system that uses a neuro-symbolic classifier.
For this purpose, we first extract structural features from the words contained in the amount vocabulary. Then, we build a symbolic knowledge base that reflects a classification of words according to their features.
In a third step, we use a translation algorithm (from rules to neural network) to determine the neural network architecture and to initialize its connections with specific values rather than random values, as is the case in classical neural networks.
This construction approach provides the network with theoretical knowledge and reduces the training stage, which remains necessary because of styles and writing conditions variability.
After this empirical training stage using real examples, the network acquires a final topology, which allows it to recognize new handwritten amounts.
back to Accepted Papers Page
Identification and Analysis of Pathological Voices by Speech Processing
Abstract:
This paper describes a speech processing interface designed for clinical observation of vocal disorders, speech assessment and pathology identification. Thanks to its accurate measurement of the vocal parameters, this PDA (Pitch Detection Algorithm) will enhance speech recognition, synthesis and prosody analysis.
By using a recorded speech sound of a patient, we extract the vocal parameters (pitch period, formant frequencies, vocal timbre, Jitter and Shimmer) which depend on the speaker physical-physiological state and the linguistic spoken language structure. The obtained results are compared with the standard values in the sake of an eventual disorder classification, a pathological prediction and prostheses assessments.
The database is constituted of sentences phonetically balanced, pronounced by several speakers (male, female, and child). After filtering and accentuation, we use a sliding Hamming window in order to obtain locally stationary signals. A compromise between the size of the window (frequency resolution) and the signal stationarity has been realized. After segmentation, we identify the voiced/unvoiced (V/UV) speech and frameworks by analysing its zero-crossing ratio and the modeling residue.
back to Accepted Papers Page
Data Image Compression Schemes for Pattern Recognition in Digital Images Using Fractals
Abstract:
Most implementations of Artificial Neural Networks (ANN) and associated algorithms when used for recognizing objects in digital images such as remote sensing, electron microscopy, medical images and others unreliable, especially if the object has been rotated or scaled. Equally, such implementations take a large amount of memory space and long CPU time to complete. This is a major problem for many computer applications such as machine vision, automatic pattern recognition and real time pattern recognition. A solution to these problems is proposed, based on the Fractal Transform Image Compression Technique [1]. The process stages are described and implementation results are provided.
back to Accepted Papers Page
Hiding Image in Image
Abstract:
A new approach introduced to embed a gray scale image into another gray or colored image. The proposed is called Image in Image Stenography System (IISS). Good results have been obtained in comparison with literature survey.
The IISS is a secret key stenography, where the algorithm is public and only the key is secret between the two parties. The main goal of stenography was fulfilled, since the resulted image did not draw any suspicion by both subjective and objective observations.
The system includes six stages: Test-change stage, Transformation stage, Best-fit search stage, Substitution stage, Position hiding stage and finally the Hiding stage.
back to Accepted Papers Page
On Comparison Results Between Low Power Design Methodologies
Abstract:
Study and comparison results of three low power design methods are presented in this paper. All the methods, based on design techniques in circuit level, achieved to reduce the power dissipation. The techniques are low-Swing voltage technique, Multithreshold voltage technique, and the Charge Recycling. Basically these are used for the same applications, such as bus architectures and long line interconnections.
back to Accepted Papers Page
Using Step-Graded Method to Decrease the Electrode Finger Reflections of SAW Filters
Abstract:
Several techniques have been used to decrease electrode finger reflections of SAW filters, like the time-gating method and the split-electrode structure IDT. In this paper we proposed a new design structure for SAW filters to decrease the electrode finger reflection effects, using of what we call “Step-Graded” method at the output or the input of SAW filter, this new method becomes more efficient as we increase the IDT electrodes, therefore, we are suggesting this new technique to replace any uniformly apodized (weighted) and normalized input or output SAW filter
back to Accepted Papers Page
VLSI Design and Implementation of RSA Encryption Algorithm
Abstract:
A VLSI design and implementation for an RSA processor is presented in this paper. This algorithm is mainly used for message authentication operations and offers high level of security strength. The proposed architecture has been implemented both for ASIC and FPGA devices. The two alternative implementations perform efficiently and are suitable for cryptography hardware implementations.
back to Accepted Papers Page
An Efficient Hardware Implementation of the SAFER+ Encryption Algorithm
Abstract:
In this paper, a VLSI implementation for the SAFER+ encryption algorithm for efficient use in portable Bluetooth telecommunication system is presented. The SAFER+ algorithm is a basic component in the authentication Bluetooth mechanism. The relation between the algorithm properties and the VLSI architecture are described. The whole design was capture by using VHDL language and the implementation was in the Xilinx FPGA (Field Programmable Gate Arrays). The VLSI architecture for the Safer+ algorithm achieves a data throughput up to 640 Mbit/sec at a clock frequency of 50 MHz.
back to Accepted Papers Page
Low Swing / Low Power CPL Logic Gate
Abstract:
A low power design circuit using low-swing voltage technique is proposed in this paper. The proposed technique could be used in order to decrease the power dissipation practically in Complementary Pass-Transistor Logic (CPL). The main idea of the proposed technique is based on the replacement of the conventional CMOS inverter at the output of the logic gate with a new Low Swing voltage Inverter based on Multithreshold voltage technology (LSIM). The inserted LSIM achieves a reduction in the static power dissipation, the dynamic power dissipation and the propagation delay time of the gate. To demonstrate the impact of the proposed technique in different applications, various types of circuits are designed for different conditions of: speed operation, load capacitance and supply voltages. In order to ensure the validity of the proposed technique in large circuit designs, an 8_bit Braun multiplier is designed in the proposed CPL of logic gate. SPICE simulation results for 3.3V supply voltage using 0.5μm multithreshold technology prove that 60%, reduction in power dissipation and 10%, reduction in delay time could be achieved.
back to Accepted Papers Page
Quantum Information Between Dream and Reality
Abstract:
This review aims to give a brief description of quantum information theory. In order to explain the relationship between information theory and quantum mechanics, the review begins with an introduction to historical background of quantum mechanics and some ideas about classical and quantum information. Finally some recent experimental methods concern the application of quantum ideas in communications, cryptography, and computation is briefly sketched. This review shows that the almost daily progress in quantum information technology is truly revolutionary innovation.
back to Accepted Papers Page
A Methodology for Error Rate Prediction for Microprocessor-Based Systems
Abstract:
This paper investigates an approach allowing evaluating the consequences of transient errors for the reliable operation of microprocessor-based digital architectures. The method is based on the injection of bit flips, randomly in time and location, by using the capabilities of typical application boards. The effects of transit bit flips on the operation of the studied architecture is investigated through fault injection experiments performed in the hardware itself by means of the asynchronous interrupt mechanism. This paper focuses in the methodology adapted to set-up such experiments on a typical microprocessor-based architecture.
back to Accepted Papers Page
Towards Another Approach for Memory Hierarchy
Abstract:
The widening CPU-memory performance gap in modern computer systems is one of the most challenging problems to overcome in today’s computer architecture designs. There is no doubt that the technological and architectural advances in the past 2 decades have caused this gap to widen. To bridge this gap, plenty of techniques have been suggested. This paper classifies these techniques into 3 categories: Techniques hiding the memory latency, Techniques redefining the role of memory, and IRAM techniques. It is proposed that L2 cache sizes are now starting to reach the point where it makes sense to manage them as the main memory of the computer, and relegate the traditional DRAM main memory to the role of a paging device. The paper details the advantages of the proposed memory hierarchy, as well as the problems that need to be solved in managing an extra level of virtual to physical translation.
back to Accepted Papers Page
Single-Bit-Difference Error Detection and Correction in Redundant Residue Number System
Abstract:
During the past few years, the residue number system (RNS) has been receiving considerable interest due to its parallel and fault-tolerant properties. In this paper, new technique for Single-Bit-difference error Detection and Correction using Redundant Residue Number System (RRNS) and base extension operation is proposed. This technique can be implemented in highly integrated processors where a concurrent fault tolerance capability is particularly important especially for real time applications. The proposed method for error detection and correction requires a very small Read Only Memory (ROM). In the paper the block diagram of the digital circuit that realizes the proposed method is introduced.
back to Accepted Papers Page
Decoding Techniques for RAC Array Codes Employed in Memory Less Channels
Abstract:
Two-dimensional array codes are easily constructed from single parity check codes, but there is a need for effective decoding algorithm to be devised, in order to use them in practical communication applications. Effectiveness of a decoding algorithm can be measured in terms of speed, memory requirement, complexity, and bit error rate performance. In this paper decoding algorithms used for row and column (RAC) two-dimensional array codes employed in memoryless channels are presented. Some traditional decoding algorithms for RAC codes, found in coding literature, are summarized and then our trellis decoding technique is presented. Comparisons between these decoding schemes are given.
back to Accepted Papers Page
Virtual Link Based Connection Admission Control in ATM Networks ( VCAC )
Abstract:
In this paper we introduced a new Call Admission Control (CAC) based on the virtual link with ultra wide bandwidth. Under the virtual link we mean that we have a reservoir that can accommodate an infinitive number of calls. So in each node we have to attach an auxiliary memory such as magnetic tapes with ultra high capacity to receive all the arriving calls. Traditional CAC will be replaced by two controllers: Virtual CAC (VCAC) and Draft CAC (DCAC). The main function of VCAC is to collect all the arriving calls without negotiations depending on its high capacity. DCAC will import the calls from VCAC, sort them into classes according to their service category, give them the appropriate VPI over which they will be carried later on the least loaded link among the outgoing links and export them to the router according to the congestion level of the physical link. The proposed scheme looks like a transistor network. The collector is represented by VCAC, the emitter is represented by DCAC and finally the base is represented by the router. The idea of the proposed strategy can be explained as follows: All the calls are admitted to the virtual memory as they arrive without any delay with respect to the extra large memory size. The virtual memory is divided into segments according to the number of the services (audio, data, video, …etc) and each segment is divided into a number of blocks size of each is a multiple of cell size (i.e. 53 octets), these blocks have the same number of VP’s (for all outgoing links) to be connected with them later (the recommended bandwidth allocation strategy is service and path separation). The block capacity is too much greater than its associated link VP. Once the block receives a new call it will spool (roll) it to its associated VP which belongs to the least loaded link, unless it becomes saturated. Once a new room appears in one of the associated VP’s in any of the outgoing links it will be occupied by a new call. It is worth noting that the calls will be pushed out by applying the famous rule FIFO: first in first out.
back to Accepted Papers Page
A TABU Search Algorithm for Scheduling Distributed Real-Time Tasks with Minimum Jitter
Abstract:
The problem of scheduling real-time tasks with minimum jitter in a distributed computing environment is important in many control applications. In this paper, we present a tabu search algorithm for nonpreemptive static scheduling of real-time tasks where tasks are periodic and have arbitrary deadlines, precedence, and exclusion constraints. The performance of the algorithm has been studied through a simulation.
back to Accepted Papers Page
A New Cost Function for Integrated Dynamic Routing in an ATM Network
Abstract:
A new efficient cost function to be used with any routing strategy and consequently to maximize the utilization of the installed resources as well as balancing the load in ATM networks is introduced. The policy is based on subdividing the traffic into a number of service classes according to their statistical and performance characteristics. At each node (ATM switch) there is a controller for computing the occupied capacity, the residual capacity and the available capacity to prevent any call rejection if there is sufficient amount of bandwidth supporting the arriving calls. When a new call request a connection, the bandwidth allocation controller assigns for it the required bandwidth, if there is enough resources, otherwise it will be rejected. After the call is finished, the assigned bandwidth will be released and returned back to the remaining amount of capacity, which we will call the channel reserve capacity. Once the call is accepted, the cost function will guide it to follow the least congested route taking into account the congestion level at the next node. Some simulation results are presented.
back to Accepted Papers Page
Domain Names: When is it Going to be in Arabic?
Abstract:
Nowadays, the Internet is considered to be the most life effecting technology. It has a huge impact on our lives. It is changing the way we do usual things. We are now using e-commerce, remote learning and telemedicine along with online entertainment. Those new concepts made our lives easier and opened new areas of knowledge and relationships. Despite all these benefits of the Internet, the majority of the world population is unable to use it. Surprisingly, economical problems are not the only reason behind that. The language barrier is the biggest obstacle in the way. Users are forced to read most of the Internet’s content in English, and they will need to use English in email and searching and other tasks. We know that Internet technology now became a worldwide issue. Users from different cultures and backgrounds are representing the majority of Internet users. Thus, localizing Internet and computer applications is becoming a strategic goal. The aim of Internet internationalization is to make the Internet globally and easily accessible for all the people of the world in their own languages.
We, as Arabs, are concerned with Arabization as a part of Internet internationalization. It is a multilevel process that takes place at more than one level. It starts with providing an Arabic content on the web, this requires an application (web design applications, email applications, etc) that supports Arabic language. With such tools, Arabic materials can be published and maintained on the web. The second step is to provide Arabic users with the needed tools (such as search engines) to locate, reach and utilize that content. The final step is to make the Arabic user able to reach the required Arabic content by using the Arabic language. It is obvious that domain names are among first steps to reach the content we want. This paper will focus on Arabizing domain names.
Using Arabic domain names is important to present Internet technology to our societies. This paper presents the needs for Arabic domain names, for example, users' rights to use their won language and not to be forced to learn other languages; and the limited nature of English language compared to the richness of Arabic language and also the limitations of current solutions. This paper tries to gather and discuss the expected problems that lay in the way of adapting Arabic domain names. This includes issues that are generated from the Arabic language grammar while the others are technical issues. Also, the paper presents the primarily result of a survey related to some Arabic linguistic issues. Finally, the paper proposes a new method of arranging Arabic top-level domain names (TLDs). This new suggestion take into accounts the nature of the Arabic language, which led to natural locking Arabic domain names.
back to Accepted Papers Page
Advanced Information Searching Strategies Using the Internet
Abstract:
This paper explores briefly some advanced techniques for information searching over the Internet. Information content on the Internet is classified broadly to help define which search methodology is suited for each division. Search engines are explored in detail and also classified. Methods for evaluating the information content on the web is presented and steps for narrowing search results are put forward. Rarely used but effective information sources are presented such as newsgroups, listservers, FTP retrieval e-mail and peer-to-peer file exchange with a brief discussion of their utility.
back to Accepted Papers Page
Multiagent System for Information-Rich Environments Application in Information Filtering from the Internet
Abstract:
This paper presents a conceptual framework of multiagent system destined to information-rich environment such as the Internet. The system has three types of agents: Interface Agent, Execution Agent and Information Agents. We particularly describe the Interface Agent “FilterAgent” that interact with the user receiving user specifications and delivering results according to user preferences (user profile). “FilterAgent” is an information filter for Internet news groups using a neural prediction model to score articles based on relevance. The filtering process uses the technique called the backward propagation of errors to adjust the connection weights of a neural network with three layers. The elements of network input vector is not simply the occurrences of terms in documents, but we use the TFIDF formula. The user can construct his profile in a particular domain, providing keywords set, and selecting a sample of documents, which he evaluates with feedback values. The trained network uses the article profile to predict expected usefulness as represented by the feedback value. Documents are sorted in descending order of the output value computed.
back to Accepted Papers Page
A Proposal for Increasing the Data Rate of the WLAN IEEE 802.11b PHY Protocols To 16.5 Mbps
Abstract:
In this paper, we propose a new method to extend the IEEE 802.11b WLAN standard physical layer protocols to achieve data rate of 16.5 Mbps. Our proposal assume the constraint that the payload modulation should continue to employ the complementary code keying (CCK) codes employed for the 5.5 and 11 Mbps rates of IEEE 802.11b. The encoding and the decoding process will be changed in minor ways, but we still use the same fast Walsh Transform (FWT) that has been used in the decoding process for 5.5 and 11Mbps of data rates. This proposal is a new attempt to extent the data rates for the IEEE 802.11b and the Es/N0 versus BER for this proposal shows a good performance.
back to Accepted Papers Page