Oguzhan Yavuz

Managing 6LoWPAN sensors with CoAP on internet

In this paper, we focus on accessing issue to IPv6 over Low Power Wireless Personal Area Networks (6LoWPAN) sensors. Constrained Application Protocol (CoAP) standard defined by IETF and standardisation works are still going on, is used to access and control 6LoWPAN sensors on internet. Simulations are implemented to show how to manage a 6LoWPAN sensor on a web page. This paper is prepared based on Bachelors Degree thesis.

Multi country support on VoIP switches: Central international switch

The VoIP switch, that Netas is responsible for the R&D, provided integration between TDM customers and new fashion IP Technologies. Moreover, the switch has “independence from location” feature, via this solution. In this work, we share our experience and talk about efficiency gained with Multi Country VoIP Switch solution that is used to manage capacity increase and IP service diversity on international switches.

FPGA implementation of Differential Evaluation Algorithm for MLP training

In this work, Differential Evolution Algorithm (DEA) is implemented on an embedded systems based on FPGA for the training of multi-layer perceptron (MLP). The classification performance of the MLP trained by DEA on FPGA has been analyzed by using a non-linear database. The MLP performance on FPGA has been compared with that on MATLAB in terms of computational performance and test accuracy. It is proved that DEA is suitable for realizing on FPGA considering simplicity of the algorithm. Simulation results of each component for DEA on FPGA are demonstrated in this paper.

Increasing trunk capacity on central international soft-switch

The High-Capacity Central VoIP switch designed by Netas, provides infrastructure and oppurtunities for integration of TDM/ATM networks and new generation IP technologies. Firstly, this new VoIP switch has been deployed for transformation of residental lines into IP networks. While the residental line network transformation is going on, it is seen that the central switches also require more trank carriers to support increased inter-office capacity. In this work, we share requirement analysis, high level design topics and experience on the developed prototype with measuring and testing activities.

IMS-TDM integration: AGCF solution

In this work, we share our experience and knowledge about integration of public service telephone network (PSTN) and IP multi-media systems (IMS). For this purpose, the enhancements and innovations realized by the Netas on access gateway control function (AGCF) solution that enables the integration of IMS with PSTN are presented.

The performance of differential evolution algorithm for training CSFNN using a pattern recognition application

In this work, Conic Section Function Neural Network (CSFNN) has been trained by differential evolution algorithm (DEA) to overcome local minimum problems. The classification performance of the CSFNN trained by DEA has been analyzed by using high-dimensional and non-linear signature recognition database. The CSFNN training performance of the DEA has been compared with that of the gradient based back-propagation algorithm (BPA). The simulation results show that the classification performance of the CSFNN trained by DEA is more stable than that of the CSFNN trained by BPA for running several trials.

Realization of the differential evolution algorithm on FPGA

In this work, a hardware implementation of differential evolution algorithm (DEA) on field-programmable gate array (FPGA) is realized. The implementation of DEA on FPGA is carried out to find minimum value of an equation with two unknowns. This study is a preparatory work for the training of neural networks with DEA on FPGA which is more complicated.

3-D spatio-temporal Gabor-type filter implementations with Time-Derivative Cellular Neural Networks

In this paper, a 3-D spatio-temporal Gabor-type filter is implemented using Time-Derivative Cellular Neural Network (TDCNN) structure. To this end, the transfer function of the spatio-temporal filter previously constructed by cascading a CNN spatial filter with a temporal filter is shown to be made equal to that of a band-pass filter obtained using a TDCNN structure having first-order temporal derivative diffusion connections only. By equating the coefficients, we obtain the discrete-space Fourier transforms of the TDCNN templates of the Gabor-type band-pass filter. Taking the inverse Fourier transform yields the TDCNN templates. Simulation results are presented.

Field Programmable Gate Array implementation of Conic Section Function Neural Network: An alternative to analog CSFNN circuitry

In this study, Field Programmable Gate Array (FPGA) implementation of Conic Section Function Neural Network (CSFNN) for a classification problem focused on iris plant is presented. This work demonstrates for the first time to our knowledge, the feed-forward computation of CSFNN implementation on FPGA. Using 16-bit floating point arithmetic and the look-up tables (LUTs) for the sigmoid function and the square root function, 83% and 72% of slices and LUTs on Spartan 3-E XC3S1600E are used for the realization of CSFNN with five neurons. The classification results obtained from the FPGA implementation and software simulation show that the accuracy error between two platforms is only 0.1%.