Improved RSS-Based DoA Estimation Accuracy in Low-Profile ESPAR Antenna Using SVM Approach

In this paper, we have shown how the overall performance of direction-of-arrival (DoA) estimation using low-profile electronically steerable parasitic array radiator (ESPAR) antenna, which has been proposed for Internet of Things (IoT) applications, can significantly be improved when support vector machine (SVM) approach is applied. Because the SVM-based DoA estimation method used herein relies solely on received signal strength (RSS) values recorded at the antenna output port for different directional radiation patterns produced by the antenna steering circuit, the algorithm is well-suited for IoT nodes based on inexpensive radio transceivers. Measurement results indicate that, although the antenna can provide 8 unique main beam directions, SVM-based DoA of unknown incoming signals can successfully be estimated with good accuracy in a fast way using limited number of radiation patterns. Consequently, such an approach can be used in efficient location-based security methods in Industrial Internet of Things (IIoT) applications.