ACTA IMEKO ISSN: 2221-870X March 2022, Volume 11, Number 1, 1 - 7 ACTA IMEKO | www.imeko.org March 2022 | Volume 11 | Number 1 | 1 Mitigation of spectrum sensing data falsification attack using multilayer perception in cognitive radio networks Mahesh Kumar Nanjundaswamy 1 , Ane Ashok Babu 2 , Sathish Shet 3 , Nithya Selvaraj 4 , Jamal Kovelakuntla 5 1 Department of Electronics and Communication Engineering, Dayananda Sagar College of Engineering, Bengaluru, Karnataka-560078, India 2 Department of Electronics and Communication Engineering, PVP SIDDHARTHA INSTITUTE OF TECHNOLOGY, Vijayawada, Andhra Pradesh- 520007, India 3 Department of Electronics and Communication Engineering, JSS Academy of Technical Education, Bengaluru, Karnataka- 560060, India 4 Department of Electronics and Communication Engineering, K. Ramakrishnan College of technology, Tiruchirappalli-621112, Tamilnadu, India 5 Department of Electronics and Communication Engineering, Gokaraju Rangaraju Institute of Engineering and Technology (GRIET), Hyderabad, Telangana-500090, India Section: RESEARCH PAPER Keywords: Cognitive radio network; cooperative spectrum sensing; energy statistic; machine learning model; spectrum sensing data falsification Citation: Mahesh Kumar Nanjundaswamy, Ane Ashok Babu, Sathish Shet, Nithya Selvaraj, Jamal Kovelakuntla, Mitigation of spectrum sensing data falsification attack using multilayer perception in cognitive radio networks, Acta IMEKO, vol. 11, no. 1, article 21, March 2022, identifier: IMEKO-ACTA- 11 (2022)-01-21 Section Editor: Md Zia Ur Rahman, Koneru Lakshmaiah Education Foundation, Guntur, India Received November 20, 2021; In final form March 1, 2022; Published March 2022 Copyright: This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Corresponding author: Mahesh Kumar Nanjundaswamy, e-mail: mkumar.n19@gmail.com 1. INTRODUCTION Over the past years, the world has witnessed a tremendous growth in the field of wireless communication technologies due to the popularity of telemedicine, smart home, smartphones, autonomous vehicles, mobile televisions and smart cities. The increasing demand for wireless communications has brought the problem of spectrum scarcity. Energy detection and measurement is a key task in spectrum sensing in cognitive radio networks. As a result, development of hybrid machine learning or signal processing algorithms becomes an intense research area for both measurement technology as well as in cognitive radio communications. The Federal Communications Commission ABSTRACT Cognitive radio network (CRN) is used to solve spectrum scarcity and low spectrum utilization problems in wireless communication systems. Spectrum sensing is a vital process in CRNs, which needs continuous measurement of energy. It enables the sensors to sense the primary signal. Cooperative Spectrum Sensing (CSS) has recommended to sense spectrum accurately and to enhance detection performance. However, Spectrum Sensing Data Falsification (SSDF) attack being launched by malicious users can lead to wrong global decision on the availability of spectrum. It is an extremely challenging task to alleviate impact of SSDF attack. Over the years, numerous strategies have been proposed to mitigate SSDF attack ranging from statistical to machine learning models. Energy measurement through statistical models is based on some predefined criteria. On the other hand, machine learning models have low sensing performance. Therefore, it is necessary to develop an efficient method to mitigate the negative impact of SSDF attack. This paper intends to propose a Multilayer Perceptron (MLP) classifier to identify falsified data in CSS to prevent SSDF attack. The statistical features of the received signals are measured and taken as feature vectors to be trained by MLP. In this manner, measurement of these statistical features using MLP becomes a key task in cognitive radio networks. Trained network is employed to differentiate malicious users signal from honest users’ signal. The network is trained with the Levenberg-Marquart algorithm and then employed for eliminating the effect of attacks due to the SSDF process. Once the simulated results are observed, it can be revealed that the proposed model could efficiently reduce the impact of malicious users in CRN.