Abstract—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 today’s state of art microprocessors processing power. Keywords—Time Difference of Arrival Averaging, Particle Swarm Optimization, Time Difference of Arrival I. INTRODUCTION TDOAA was presented by Ralph O. Schmidt for three receivers in 1972 and generalized in 1996 [1, 2]. The method claims that the sum of the TDOAs in a closed loop must be zero in the absence of TDOA estimation error. The transmitter can be positioned using at least three or more receivers in two dimensional planes. Increasing number of the receivers decreases the positioning error. Variety of positioning methods has been indexed in the literature so far [3-10]. These methods can be classified in two categories as geometric and stochastic. Stochastic methods are more preferred thanks to their nature of being less sensitive to noise. Determination of the source location is executed using different source of information such as Received Signal Strength (RSS), Direction of Arrival (DOA), Time of Arrival (TOA), Time Difference of Arrival (TDOA), and Frequency Difference of Arrival (FDOA). Comparing to RSS, DOA and TOA, determining the location of the transmitter is obtained with more accurate using the TDOA. Furthermore there is no necessity to know the transmitter signal. On the other hand receivers must be synchronized. PSO and TDOAA were used separately for source localization applications. In this study Manuscript received February 27, 2012. O. Cakir, A. Yazgan, O. Cakir, E. Tugcu, and I. Kaya are with the Karadeniz Technical University, The Faculty of Engineering, 61080, Trabzon, Turkey (corresponding authors to provide phone: +904623774203; fax: +904623257405; e-mail: cakir@ktu.edu.tr, ayhanyazgan@ktu.edu.tr, cakiro@ktu.edu.tr, emintugcu@ktu.edu.tr, ikaya@ktu.edu.tr). we combined both proposing a new method and obtained a notable performance increase. The paper is organized as following. Section II outlines the emitter location finding and TDOAA. Section III describes the PSO and IPSO algorithm. Section IV presents the using PSO and IPSO algorithms for source localization. Section V presents and discusses the results comparatively. Consequently, Section VI concludes the paper. II. EMITTER LOCATION FINDING AND TDOA AVERAGING A. Emitter Location Finding using TDOA The number of TDOA for n receivers can be calculated using (1) ( ) 1 2 nn m - = (1) Here m shows the number of TDOAs. Only n-1 TDOAs are independent from each other’s. m TDOAs can be calculated using n-1 independent TDOAs. Three receivers positioning system can be seen in Fig. 1. where x 1 , y 1 , x 2 , y 2 and x 3 , y 3 are the receivers’ positions and x t , y t show the transmitter location. To calculate the exact (error free) TDOAs, first of all, the distances between the transmitter and the receivers must be obtained. These distances are given l 1 , l 2 and l 3 for the first, second and third receiver respectively and calculated in (2-4). 2 2 1 1 1 ( ) ( ) t t l x x y y = - + - (2) 2 2 2 2 2 ( ) ( ) t t l x x y y = - + - (3) 2 2 3 3 3 ( ) ( ) t t l x x y y = - + - (4) 12 1 2 r l l = - (5) 13 1 3 r l l = - (6) 23 2 3 r l l = - (7) Where r 12 is the distance difference between first and second receiver; r 13 is the distance difference between first and third receiver and r 23 is the distance difference between second and third receiver and calculated in (5-7) The TDOAs are calculated as given below where c is the signal velocity in the medium and known exactly and Δ is the exact TDOA for the related receivers. Novel Composite Method for Determining the Location of the Transmitter using Particle Swarm Optimization Oguzhan Cakir, Ayhan Yazgan, Omer Cakir, Emin Tugcu, and Ismail Kaya 335 978-1-4673-1118-2/12/$31.00 ©2012 IEEE TSP 2012