Oguzhan Cakir

Novel composite method for determining the location of the transmitter using Particle Swarm Optimization

In this study, the Time Difference of Arrival Averaging (TDOAA) method is studied to minimize the estimation error and increase the accuracy for emitter location finding using Particle Swarm Optimization (PSO). We combined TDOAA method with both classic and improved version of PSO and found out a considerable performance increase on the location finding of the transmitter. The improved PSO is to combine different tasks of implementations of PSO for a single decision. Therefore the improved PSO also means more accuracy but price is paid for more complexity. However, the increase on complexity does not prohibit applying the technique, since coherence time for decision may well tolerate more complexity when using todays state of art microprocessors processing power.

Optically reconfigurable fractal antennas for RoF systems

This study presents the operating frequency reconfigurability for the Sierpinski gasket fractal antenna. This design may be a good candidate for RF over fiber (RoF) systems in which the built-in optical systems have already been included. Doped silicon with a suitable laser light is considered as a switching component.

Investigation the effect of the time difference of arrival sets on the positioning accuracy for source localization

Electromagnetic, acoustic, or seismic sources can be positioned using direction, time, frequency or time/frequency difference of their transmitted signals. Source coordinates are found using time difference of arrival (TDOA) method that depends on the time differences of spatially separated receivers. In these methods, two different TDOA sets are used, including independent/spherical and full. In this study, positioning with particle swarm optimization (PSO) is achieved using independent set and full sets and positioning error is compared with Cramer-Rao lower bound (CRLB). In addition, it is proven that CRLB can be exceeded using the PSO algorithm that uses the full TDOA set.

Optimization of receiver arrangements for emitter location finding using time difference of arrival

In this study, particle swarm optimization (PSO) is applied to the time difference of arrival (TDOA) method to determine the emitter location. In addition, receiver arrangements have been optimized to reduce the positioning error. Depending upon the position of the emitter, the receiver arrangements which move the Cramer-Rao lower bound (CRLB) to the lowest value is calculated by PSO.

Real-time cutting simulation based on stiffness-warped FEM

Finite element method (FEM) is a widely used method in the simulation of virtual surgery since it can model elasticity behavior of soft tissues accurately. Linear FEM uses constant stiffness matrix that achieves fast and stable simulation but objects increase unnaturally in volume under large rotational deformation. Nonlinear FEM, on the other hand, models large deformation correctly but it is computationally complex. “Stiffness Warping” method is suitable for real-time virtual surgery applications because it updates rotational parts of the large deformations rapidly and stably. It also preserves the volume. Cutting is the main activity in virtual surgery and various methods have been presented to conduct this activity. This paper presents an improved version of the prior cutting methods used in virtual surgery. The test results indicate that the proposed surgery simulation method based on Stiffness-Warped FEM could perfectly be utilized in real-time virtual surgery applications.

Accuracy comparison of time difference of arrival based source localization methods

The position of an electromagnetic, acoustic or seismic source can be obtained by spatially separated synchronize receivers. Using the time differences between the receivers, hyperbolic lines of position is defined and the target can be localized at the intersect point of them. Different algorithms based on time difference of arrival nonlinear and linear algorithms can be found in literature. In this study, nonlinear least square, maximum likelihood, linear least square and weighted linear least square methods are compared each others. Additionally, the mean square errors are compared with the Cramer-Rao lower bound for different receiver geometries and noise levels.

Investigation the effect of the distribution of receivers on the positioning error for emitter location finding

Emitter location can be found using three or more synchronized receivers. In order to determine the emitter location, different type of methods may be utilized such as received signal strength, direction/time of arrival, and time/frequency difference of arrival. In this study, the time differential of arrivals (TDOA) is selected as the basic method with particle swarm optimization (PSO) algorithm. In addition, the effect on positioning error of the various receiver distributions is examined.

Different perspective of time difference of arrival averaging

Any emitter can be positioned using three or more receivers in two dimensional planes and four or more receivers in three dimensional planes. The transmitter signal reaches at the receivers that synchronized each others in different times. Using this time differences of arrival (TDOAs) hyperbolic lines of positions (LOPs) are defined and target is positioned at the intersection of LOPs. It is well known that in the absence of noise, the sum of the TDOAs which produces a closed loop considered as zero. This rule is the basic idea of the TDOA Averaging (TDOAA) method. TDOAA significantly decreases the estimation error and improves the accuracy of positioning. In this study a different perspective has been introduced for TDOAA by proposing a simple graph traversal algorithm in order to produce coefficient matrix. The effect of TDOAA on the TDOAs estimation error is also investigated.

Variable Threshold Based Cutting Method for Virtual Surgery Simulations

We propose a variable threshold based cutting method for tetrahedralized deformable objects used in virtual surgery simulations. Our method uses cut-nodes and cut-edges scheme to determine cutting path with variable threshold improvement. Static threshold may cause ambiguity in cut-node selection when the threshold is so large that one of two vertices of the triangle to be cut can be selected as a cut-node. In addition, small threshold value results in many tetrahedrons while a smooth cut and large threshold value leads to a more jagged incision with less small tetrahedrons. Thus, both small and large threshold values have advantages and disadvantages. Simulation results show that our variable threshold based cutting method offers relatively better results than static threshold based one in terms of advantages and disadvantages mentioned above.

A Novel Approach for Blind Channel Equalization

The Constant Modulus Algorithm (CMA), while the most commonly used blind equalization technique, converges very slowly. The convergence rate of the CMA is quite sensitive to the adjustment of the step size parameter used in the update equation as in the Least Mean Squares (LMS) algorithm. A novel approach in adjusting the step size of the CMA using the fuzzy logic based outer loop controller is presented in this paper. It also presents a computational study and simulation results of this newly proposed algorithm compared to other variable step size CMA such as conventional CMA, Normalized CMA (N-CMA) [1], Modified CMA (M-CMA) [2], CMA-Soft Decision Directed (CMA-SDD) [3]. The simulation results have demonstrated that the proposed algorithm has considerably better performance than others.