Research Article
SVM Multi-class Classification Method for Device Identification Using Eye Diagram Parameters
@INPROCEEDINGS{10.1007/978-3-031-70507-6_17, author={Jian Yuan and Aiqun Hu}, title={SVM Multi-class Classification Method for Device Identification Using Eye Diagram Parameters}, proceedings={IoT as a Service. 9th EAI International Conference, IoTaaS 2023, Nanjing, China, October 27-29, 2023, Proceedings}, proceedings_a={IOTAAS}, year={2024}, month={10}, keywords={device identification eye diagram supporting vector machine physical layer security}, doi={10.1007/978-3-031-70507-6_17} }- Jian Yuan
Aiqun Hu
Year: 2024
SVM Multi-class Classification Method for Device Identification Using Eye Diagram Parameters
IOTAAS
Springer
DOI: 10.1007/978-3-031-70507-6_17
Abstract
In this paper, we investigate the problem of identification of RS-485 devices. The proposed method utilizes their physical layer features in terms of time-domain overlapping traces. The first step is the acquisition of the physical layer waveform with an oscilloscope. The waveform is then preprocessed, a pulse shaping filter is applied, and the eye diagram is generated. These eye diagram parameters are estimated and stored as datasets. Then, the datasets are divided into two conventional categories. The first one is defined as the training set for deriving key parameters, the second one is used as the test set to verify the proposed algorithm. The eye diagram parameters of the test set are used for training, and the parameters of multi-class the supporting vector machine (SVM) are obtained. The data from the testing set is used to classify the devices. The results of the classification are used for device identification. Our experimental results show that classification accuracies of RS-485 devices can be higher than 88%, which indicate that our proposed method is practical and thus has a potential to be used in a number of applications including industrial Internet of Things (IoT) device identifications.
