Murat H. Sazli

Speech recognition with artificial neural networks

In this paper, artificial neural networks were used to accomplish isolated speech recognition. The topic was investigated in two steps, consisting of the pre-processing part with Digital Signal Processing (DSP) techniques and the post-processing part with Artificial Neural Networks (ANN). These two parts were briefly explained and speech recognizers using different ANN architectures were implemented on Matlab. Three different neural network models; multi layer back propagation, Elman and probabilistic neural networks were designed. Performance comparisons with similar studies found in the related literature indicated that our proposed ANN structures yield satisfactory results.

Effective cooperative spectrum sensing in IEEE 802.22 standard with time diversity

IEEE is developing a wireless communication standard that will take place in cognitive radio class and be valid in the world. Data communication capability supporting is aimed with this standard called IEEE 802.22 to the rural area users in TV bands. In the TV bands, not to cause harmful interference to the existing TV receivers and to communicate in unused channels, effective detection of unused channels is needed. This is succeeded with effective spectrum sensing. Sensing the presence of TV signals in the channel being tracked is more possible with using effective spectrum sensing techniques. In this paper, it is demonstrated with detailed simulations that a base station and users in an IEEE 802.22 cell sense TV signals better with setting up cooperation and it is reached to the better performance results by adding time diversity to this cooperation. Besides, it is also presented in this paper that effective cooperation is done with decrease of cost, power consumption and complexity by choosing relays who have higher SNR threshold.

Neural network implementation of the BCJR algorithm

In this paper, we first show that the BCJR algorithm (or Bahl algorithm) can be implemented via some matrix manipulations. As a direct result of this, we also show that this algorithm is equivalent to a feedforward neural network structure. We verified through computer simulations that this novel neural network implementation yields identical results with the BCJR algorithm.

Neural network based face detection from pre-scanned and row-column decomposed average face image

This paper introduces a methodology for detecting human faces with minimum constraints on the properties of the photograph and appearance of faces. The proposed method uses average face model to save the computation time required for training process. The average face is decomposed into row and column sub-matrices and then presented to the neural network. To reduce the time required for scanning the images at places where the probability of face is very low, a pre-scan algorithm is applied. The algorithm searches the faces in the image at different scales for detecting faces in different sizes. Arbitration between multiple scales and heuristics improves the accuracy of the algorithm. Experimental results are presented in this paper to illustrate the performance of the algorithm including accuracy and speed in detecting faces.

S-band TCR patch antenna design for nano satellites

The small satellites offer reduced cost and low mass solutions for various mission needs such as Earth monitoring, scientific experiments, and data communication. Thus the small satellite market grows proportionally as their developing technologies replace the operations of conventional large satellites. Nowadays manufacturers consider small satellites due to their reduced time scale during design and integration phases. In this trend, light weight, reliable antennas with wide radiation patterns and adequate gains have to be employed in small satellites to accomplish telemetry, command, ranging (TCR) functions and high speed payload data transmission. In this study, an S-Band patch antenna design procedure is examined in order to meet the TCR functional requirements in nano satellites. The proposed patch antenna can be utilized for the payload transmission which includes the amateur radio applications and earth monitoring.

Cooperative spectrum sensing for mobile users

Cognitive radio is a radio communication technique, where unlicensed or cognitive users monitor the spectrum for the existence of licensed (primary) users and have a communication capability when the primary users donpsilat transmit their data. This technique needs an effective spectrum sensing because of the not disturbing the primary users. The increasing of this effectiveness is related with the higher detection probability of the existing primary users. In this paper, new cooperative spectrum sensing technique recommended for mobile users when the environment is free space line of sight (LOS). It is confirmed with this technique that the probability of detection of the primary userpsilas existence increases with the using of cooperative diversity by mobile users. It is resulted that this increasing of the detection probability is well-proportioned with relay numbers who helps target secondary user. In addition to this, systempsilas performance increases with the threshold distance of the true sensing. But it is demonstrated with the detailed simulations that the probability of detection of the primary userpsilas existence decreases with the speed of the mobile users.

Accurate indoor localization with optimized fingerprinting algorithm

Global positioning systems, that proved their success in the outdoors, cannot perform as well in enclosed environments because they suffer from absence of line of sight or bad reception quality of base stations. In this regard, methods are being developed for highest accuracy indoor locating performance with least cost. Among these methods localization with fingerprinting is far more superior to other indoor localization methods as it uses the surrounding signals in the environment for accurate positioning and is available to most common mobile devices. In this work, indoor fingerprinting algorithms for localization of mobile devices based on Correlation Database (CDB) Filtering, Genetic Algorithm (GA) and Big Bang — Big Crunch (BB-BC) are compared. Results show that using adaptive GA or BB-BC < 3 m error of %95 can be achieved with 4 Bluetooth Low Energy (BLE) beacons distributed around a 40 m2 testbed, compared to < 5.4 m for CDB Filtering.

Slot Parameter Optimization for Multiband Antenna Performance Improvement Using Intelligent Systems

This paper discusses bandwidth enhancement for multiband microstrip patch antennas (MMPAs) using symmetrical rectangular/square slots etched on the patch and the substrate properties. The slot parameters on MMPA are modeled using soft computing technique of artificial neural networks (ANN). To achieve the best ANN performance, Particle Swarm Optimization (PSO) and Differential Evolution (DE) are applied with ANN’s conventional training algorithm in optimization of the modeling performance. In this study, the slot parameters are assumed as slot distance to the radiating patch edge, slot width, and length. Bandwidth enhancement is applied to a formerly designed MMPA fed by a microstrip transmission line attached to the center pin of 50 ohm SMA connecter. The simulated antennas are fabricated and measured. Measurement results are utilized for training the artificial intelligence models. The ANN provides 98% model accuracy for rectangular slots and 97% for square slots; however, ANFIS offer 90% accuracy with lack of resonance frequency tracking.

Discrete Lissajous Figures and Applications

In this paper, an innovative method based on an algorithm utilizing discrete convolutions of discrete-time functions is developed to obtain and represent discrete Lissajous and recton functions. They are actually discrete auto- and cross-correlation functions. The theory of discrete Lissajous figures is developed. The concept of rectons is introduced. The relation between the discrete Lissajous figures and autocorrelation functions is set. Some applications are described including phase, frequency, and period determination of periodic signals, time-domain characteristics (such as damping ratio) of a control system, and abnormality and spike detection within a signal, are described. In addition, an electrocardiogram signal with an abnormality of atrial fibrillation is given for abnormality detection by means of recton functions. An epileptic activity detection within an electroencephalography signal is also given.

Cooperative spectrum sensing in IEEE 802.22 standard

IEEE is developing a wireless communication standard that will take place in cognitive radio class and be valid in the world. Data communication capability supporting is aimed with this standard called IEEE 802.22 to the rural area users in TV bands. In the TV bands, not to cause harmful interference to the existing TV receivers and to communicate in unused channels, effective detection of unused channels is needed. This is succeeded with effective spectrum sensing. Understanding of presence of TV signal in channel tracked is more possible with that reason. In this paper, it is demonstrated with detailed simulations that a base station and users in an IEEE 802.22 cell sense TV signals better with setting up cooperation and it is reached to the better performance result with adding time diversity to this cooperation.