Developing security measures against electromagnetic signal injection

点击翻译按钮获取中文摘要

Cyber security threats are increasing for businesses, critical infrastructure, and end users. The attacks come in two primary varieties, software and hardware. Software attacks are usually done remotely via the internet, but may be local, requiring physical access to connect a malicious pen drive. Hardware attacks are also performed locally (by tampering with equipment), however they may also be done remotely. The threats posed by all but the simplest remote hardware attacks are often overlooked. The primary attack vector for this is electromagnetic interference, the simplest and most well-known example of which is jamming. A significantly more sophisticated method is signal injection into sensors: many critical computer systems rely on their sensor data in order to make predictions or decisions (such as security systems, industrial control equipment, auto pilots, ...). At their core, most electronic sensors are converting real world effects into a small voltage, and then performing an accurate measurement of the voltage (e.g. microphones, cameras, antennas), which means they are inherently susceptible to electromagnetic interference inducing a small voltage. For similar reasons, many communication wires are susceptible to injection of data with electromagnetic interference. The research aims to investigate potential attacks by electromagnetic signal injection on a variety of systems used in the previously mentioned critical applications, studying their downstream effects, and determining the required countermeasures. Part of this is examining the operating principle of the device, to theoretically deduce what electromagnetic signal would be the most effective, and performing real world tests. On the downstream side, the research will examine the complex ways in which senor measurements are processed and used by software, and understanding how tampered values can lead to potentially disastrous outcomes. Finally, in security research it is very important to provide advice to manufacturers on how the hardware issues can be mitigated in the future, to consumers on how affected hardware can be identified and upgraded, and to software developers on how malicious values may be detected and safely handled. The research will focus on novel devices (such as electric and self-driving vehicles, IoT devices), on novel attack methods, and on novel countermeasures which can be implemented cheaply and easily, in order to ensure that they are swiftly adopted. Some of the work may be undertaken in collaboration with the research division of Armasuisse.This project falls within the EPSRC "Information and Communication Technologies" research area.