The modeling of propagation path loss calculations by using artificial neural networks

The analysis of indoor radio propagation is very important for the maintenance of a sustainable and an effective communication link in Wireless systems. A sufficient radio propagation must be maintained to avoid the interruption of the data transmission. The positions of the transmitters and receivers can be determined with the various models predicting the radio signal power level versus the distance between the antennas. With the use of these models, an efficient wireless network can be designed. The maximum distance, where the communication is still maintained, is called the range of the system and this value is identified by the ?propagation path loss? term. The path loss calculations are modelled due to the characteristics of the communication system and the propagation environment and thus, a reliable coverage area prediciton can be done. In this assignment, the Wireless Local Area Network (WLAN) system operating at 2.4 GHz frequency band and locating in the entrance floor of the Yildiz Technical University Besiktas Campus T-Block have been researched and the related radio propagation path loss calculations have been modelled. Firstly, the general concepts and technical standards of the WLAN system have been introduced. Then, the mechanisms affecting the large-scale path loss have been identified and various emprical and determinisitical prediction models have been defined. Furthermore, an alternative Artificial Neural Network (ANN) approach have been developed and implemented. The Electric Field Strength measurements were performed in the propagation environment and these values were used to train the ANN model with Levenberg- Marquardt training algorithm.. By using MATLAB, the ANN prediction results have been compared with the emprical models and the measurement values. Moreover, the achieved results during this project have been compiled in two articles which are given in the acknowledgement section. Key words: Wireless Local Area Networks (WLAN), Indoor Propagation, Propagation Path Loss, Artificial Neural Networks,Levenberg-Marquardt.