Statistical tuning of Walfisch-Ikegami propagation model using Particle Swarm Optimization

Abstract

Propagation models play a vital role in the characterization and design of wireless and mobile communications networks. However, if they are utilized in a different environment than the one they were formulated for, propagation path-loss models may produce unacceptable deviation in predictions. This paper proposes a statistical tuning technique based on Particle Swarm Optimization (PSO) to calibrate the COST-231-Walfisch-Ikegami path loss propagation model using collected received signal power measurements from a deployed 3G network. The parameters of the modified model (partial coefficients of regression) are estimated considering the roof height, distance between buildings, and road orientation angle with respect to an incident radio signal as random variables. Appropriateness and validity of the tuned propagation model was demonstrated through comparisons and statistical indicators.

Cellular Network Planning Toward 2021 Case Study: Bukit Raya District Pekanbaru

“…Several propagation models can be found in the literature review such as Okumura-Hatta and Cost 231 Walfisch-Ikegami [3]- [5]. Cost-231-WI model is a combination of Walfisch and Ikegami-Bertoni model based on numerous site tests and analysis.…”

“…It calculates the multiple screen forward diffraction loss of base station antenna. This gives a better path loss prediction [5]. Therefore, this model is used in this paper.…”

“…The free-space path loss is an analytical model that predicts the strength of the signal received when a clear line of sight path exists between transmitter and receiver. The path loss depends on the distance between transmitter and receiver and frequency [5].…”

“…L p = 32.4478+ 20 log (f ) + 20 log(d ) + 60 (7) where the f is the frequency (MHz), d is the distance (m). The free space loss is calculated by following equation [5]:…”

“…The rooftop to street diffraction and scatter loss are determined by [4]: Simulation parameters used in this simulation are listed in Table 3. The multiscreen loss is calculated by using the following equation (10) [5]:…”