Near-ultrasonic covert channels using software-defined radio techniques

Traditional cybersecurity practices rely on computers only communicating through well-defined expected channels. If malware was developed to use covert channels, such as one created using ultrasonic sound, then this could bypass certain security measures found in computer networks. This paper aims to demonstrate the viability of acoustic covert channels by creating a low-bandwidth ultrasonic frequency channel utilising software-defined radio (SDR) techniques. Previous work was evaluated to identify the strengths and weaknesses of their implementations. Software-defined radio techniques were then applied to improve the performance and reliability of the acoustic covert channel. The proposed implementation was then evaluated over a range of hardware and compared to previous implantations based on the attributes of their throughput, range, and reliability. The outcome of this research was an ultrasonic covert channel implemented in GNU Radio. The proposed implementation was found to provide 47% higher throughput than previous work while using less signal bandwidth. Utilising software-defined radio techniques improves the performance of the acoustic covert channels over previous implementations. It is expected that this technique would be effective in an office environment, but less effective in high security or server environments due to the lack of audio equipment available in these spaces.