The main goal of the project is to enhance the cyber security of digital devices, such as Embedded and Mission‐Specific Devices (EMSD), by developing a monitor capable of detecting attackers by analyzing involuntary analog emissions across a variety of modalities, including power consumption, electromagnetic, and acoustic emissions. These involuntary analog emissions, also known as side channels, enable a new protection paradigm that separates security monitoring functionality from the protected system.
We attempt to integrate sensors with high spatial discrimination and smart algorithms to boost signals and reject interference, leverage sophisticated machine learning algorithms coupled with robust feature extraction techniques and synchronization to cope with complexity, boost fidelity using program instrumentation, and evaluate the performance using different deception techniques to stress the system and show effectiveness in a realistic environment. The main novelty of this approach, along with strong synchronization emphasis, is the development of a sensor-agnostic toolset to enable deployment flexibility and scalability.