Improving of the Interference Classification Techniques under the Smart Farming Environment using iSVM

Abstract

Smart farming is one of the recent concepts to increase the capability of the agriculture sector. This concept combines a set of algorithms, electronic sensors or devices, and technologies. The Internet of things (IoT), big data, and artificial intelligence (AI) play a significant role in providing and supporting the solution and optimization ways with the massive data inside the farm. Due to a large number of data inside the farm, smart farming needs to deploy the IoT tech-nology to communicate and transmit the data. However, the interference signals from the adjacent sensors or channels are a critical problem to reduce the reliability of the transmitted data. Therefore, we propose the i$S$VM experiment to observe and classify the interference signal from the received signal. The iSVM experiment compared the classical support vector machine (SVM), SVM with the radial basis function (RBF) kernel, and SVM with the different degrees of the polynomial kernel. Before implementing the i$S$VM experiment, this paper generated an IoT in smart farming with the effects of the actual environment, i.e., the path loss exponent, the additive white Gaussian noise (AWGN) noise, and the small scale fading. Next, this paper implemented the i$S$VM to classify and suppress the interference signal. Moreover, an i$S$VM was compared with the minimum mean square error (MMSE) filter and the received without the suppression technique. From our numerical results, SVM with the polynomial of degree 4 can perform with 80 percent (%) of the average accuracy.