A Semi Empirical Path Loss Model by Using Artificial Neural Networks (Ann)
Sreevardhan Cheerla1, D Venkata Ratnam2, Chandra Vinay3, Narayanasetti Avinash4, Illa Anish5, Shaik Baji Imran6
1Sreevardhan Cheerla, Assistant Professor, Department of ECE, KLEF, Guntur (Andhra Pradesh), India.
2D. Venkata Ratnam, Professor, Department of ECE, KLEF, Guntur (Andhra Pradesh), India.
3Chandra Vinay, Pursuing Bachelor of Technology, Department of ECE, KLEF, Guntur (Andhra Pradesh), India.
4Narayanasetti Avinash, Pursuing Bachelor of Technology, Department of ECE, KLEF, Guntur (Andhra Pradesh), India.
5Illa Anish, Pursuing Bachelor of Technology, Department of ECE, KLEF, Guntur (Andhra Pradesh), India.
6Shaik Baji Imran, Pursuing Bachelor of Technology, Department of ECE, KLEF, Guntur (Andhra Pradesh), India.
Manuscript received on 05 May 2019 | Revised Manuscript received on 12 May 2019 | Manuscript published on 30 May 2019 | PP: 835-841 | Volume-8 Issue-7, May 2019 | Retrieval Number: G5786058719/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Path loss Models are essential for determining strength of received signals in hostile mobile propagation environment. In this paper, a method for propagation path prediction for urban, suburban and rural areas at 800 MHz, 1800 MHz is presented based on Artificial Neural Networks (ANN). The application of feed forward ANNs makes it likely to overcome drawbacks that we come across when we use prediction models, including both statistical and deterministic models. The Model uses the back propagation algorithm and considers the semi empirical model (COST-231 Walfisch Ikegami) as the reference standard, from which we consider the inputs. The obtained measurements are splitted into three sets, of which the first set is utilized for Model Training, second set for Model Testing and last set for Model Validation. The ANN Model’s performance for frequencies 800MHz and 1800 MHz demonstrates that the Mean Absolute Error (MAE) is 3.24 (Urban), 2.51 (Rural) and 1.91 (Sub urban) regions, corresponding MAE for 1800 MHz are is 2.52 (Urban), 2.18 (Rural) and 1.72 (Sub urban).
Keyword: Path loss, Artificial Neural Networks, Multilayer perceptron, Costs 231 WI Model, Levenberg Marquardt (LM) algorithm.
Scope of the Article: Artificial Neural Networks