2017/11/03 by Aidin Ferdowsi, Walid Saad, Ferdowsi, Aidin +1
Computer Science · #Advanced Malware Detection Techniques #Cryptography and Security (cs.CR) #FOS: Computer and information sciences #Information Theory (cs.IT) #Internet Traffic Analysis and Secure E-voting #Multimedia (cs.MM) #Network Security and Intrusion Detection
paper · pdf · doi:10.48550/arxiv.1711.01306
openalex publication_date 2017/11/03 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Securing the Internet of Things (IoT) is a necessary milestone toward\nexpediting the deployment of its applications and services. In particular, the\nfunctionality of the IoT devices is extremely dependent on the reliability of\ntheir message transmission. Cyber attacks such as data injection,\neavesdropping, and man-in-the-middle threats can lead to security challenges.\nSecuring IoT devices against such attacks requires accounting for their\nstringent computational power and need for low-latency operations. In this\npaper, a novel deep learning method is proposed for dynamic watermarking of IoT\nsignals to detect cyber attacks. The proposed learning framework, based on a\nlong short-term memory (LSTM) structure, enables the IoT devices to extract a\nset of stochastic features from their generated signal and dynamically\nwatermark these features into the signal. This method enables the IoT's cloud\ncenter, which collects signals from the IoT devices, to effectively\nauthenticate the reliability of the signals. Furthermore, the proposed method\nprevents complicated attack scenarios such as eavesdropping in which the cyber\nattacker collects the data from the IoT devices and aims to break the\nwatermarking algorithm. Simulation results show that, with an attack detection\ndelay of under 1 second the messages can be transmitted from IoT devices with\nan almost 100% reliability.\n