IRAQI JOURNAL OF COMPUTERS, COMMUNICATIONS, CONTROL AND SYSTEMS ENGINEERING

Linear phase requirement can be achieved by direct design with an FIR filter atthe expense of high order and high computational burden for the given specification.Alternatively, IIR digital filters can be designed with much smaller order than their FIRcounterparts, but at the expense of the non-linear phase. In this paper, a new design oflinear phase IIR digital filter is presented by composing a zero phase efficientlydesigned FIR digital filter with maximally flat group delay. The new IIR digital filter isdesigned with a lower order satisfying the same roll-off magnitude response of thecorresponding FIR filter and preserving the approximate pass-band characteristics.

The paper attempts to provide most of the well-known security services through adesign and implementation of a client/server-based network security system. The systemis based on recommendation of the Advanced Encryption Standard (AES) together withsome secure techniques. These techniques include message digest (MD5), datacompression, data scrambling, traffic padding, novel PN sequence generator that is basedon AES and the RSA for secure key distribution.The generator is based on the standard ANSI X9.17 with triple AES instead oftriple DES. Recommended statistical tests are carried out to check the randomness of thenew generator. The results of these tests show that the new generator passes all therequired tests successfully. The latter results being even better than those achieved by thestandard generator ANSI X9.17 using AES or triple DES.Finally, the security system was implemented and tested between two PCs in alocal area network. The results clearly demonstrate the successful operation of thesecurity system through the secure transmission of data. The security of the proposedsystem is shown to be efficient since all the combined techniques are well-establishedand recommended. Furthermore, with the multi-keys required to perform the operation,the tasks of cryptanalyst with available computing power will be difficult.

The concern of the paper is to use novel approach to identify the signal. The approachrepresents the relation ship between the input and output through the systolic arraystructure, and drivers the relation directly from the transfer function of the digital filter, sothat signal identification can be effectual in real-time. The paper proposes a one-dimension(1-D) systolic array architecture of infinite and finite impulse response (IIR &FIR) digitalfilters. The proposed architecture is employed in real-time signal identification .This wasverified after the identification process of the signal in real-time by applying the proposedstructure in simulink.

There are measures that can be taken to protect users from objectionable orinappropriate Internet content and secure the LAN from theft, modification, or deletionof data. Firewall is one of many security measures that can be used to protect networks.Although firewalls are powerful, they suffer from many types of attacks.This paper concentrates on one particular aspect, which is providing firewallssecurity against five types of attacks. These attacks are: SYN Flooding, Ping of Death,IP Address Spoofing, Impersonate one-half of a Session and Session Hijacking. Thiscan be done by building a specific strategy for a firewall; a strategy that has threeprocedures, some of which could be stood for more than one attack.

A new algorithm for error concealment of a completely lost blocks using the 1-Ddiscrete wavelet transform is proposed in this paper. The transmission of block-codedimages, fading in wireless channels can cause errors in entire blocks and consequentlycauses the loss of information of the block. Instead of using common retransmissionquery protocols, the error block is substituted for pixel values of the neighborhood ofthe lost block.The proposed algorithm does not require a DC estimation method. While most of thepreviously error concealment methods assume that the DC value is available or DCestimation is required.

Artificial Neural Networks are employed for demodulation of spread spectrumsignals in a multiple – access environment. In fact, the conventional matched filterreceiver, when used in multi-user system, suffers from the interfering signals, so thatwhen reducing this problem, we need to make the receiver in optimum state. Thisreceiver is too complex for practical use. Two simple structures employing multilayernodes are produced for demodulation of spread-spectrum signals. The Neural Networksare trained for the demodulation of signals by using back propagation type algorithms.A comparative analysis of the two receivers, the conventional and that employingNeural Network, has been presented.

A neural network-based self-tuning PID controller is presented. The scheme ofthe controller is based on using a modified Elman recurrent neural network as a selftunerfor (PID) controller. The proposed method has the advantage of not necessarilyusing a combined structure of identification and decision, common in a standard selftuningcontroller, because it uses a genetic algorithm based model reference. The paperexplains the algorithm for a general case, and then presents a specific application onnon-linear dynamical plant.

The core objectives of this research are to identify, understand and describerange of workable–based methods for quantitatively measuring the quality of an imagerepresented in various file formats. Moreover, it suggests methods for measuring animage quality in varied graphic file formats. The research describes current popular andavailable graphic file formats, and the compression methods are utilized as well as colorspaces, etc. The type of image display in use also affects the visual appearance ofimages. Furthermore, the lower depth output and use of colors fail affects themagnification features.