Prediction Method for Abnormal Behaviors of Embedded Systems

With the rise of Artificial Intelligence Internet of Things (AIoT), the demand for embedded systems has increased. In the past, different testing tools were required for the security testing of different embedded systems, resulting in increased manpower and time costs. This study proposes a method for detecting and predicting abnormal behaviors of embedded systems, which can be adapted to different embedded systems, avoiding the need for the establishment of specific abnormal detection systems on different embedded systems. This can reduce the development and time costs of some anomaly detection systems. In addition, the server can be used to monitor embedded systems and manage different embedded systems. This study uses Arduino Mega2560 Rev3 as the development and verification platform of the embedded system. Different malicious behaviors in four experiments are used to verify the feasibility of the proposed method. The experimental results show that using the instruction sequence of the proposed embedded system as behavioral feature data is an effective method of detecting abnormal behaviors within the system.