For a long time, it was assumed that underwater communications were secure from eavesdropping. That assumption has now been shattered.
Researchers from Princeton and MIT have developed a method to intercept underwater messages and locate an underwater transmitter from the air using radar. Their innovative device decodes tiny vibrations created by sonar signals on the water’s surface. The findings, presented at ACM MobiCom, detail the technology and offer strategies to guard against the attacks it enables. The demonstration was conducted on Lake Carnegie, a small artificial lake in Princeton, as the open ocean would present additional challenges. However, the team believes that with further engineering improvements, eavesdropping in open ocean environments could become feasible. The goal of their research was to expose the vulnerabilities in underwater transmissions and provide methods to prevent interception.
In 2018, MIT developed a system that required cooperation between the air and sea parties, with shared data rates, frequencies, and technical details provided in advance. The current research takes this a step further, enabling the interception of messages without prior knowledge of technical parameters. This advancement highlights significant security risks, as the technology could potentially intercept sensitive data transmitted by climate monitoring sensors, oil and gas equipment, and submarines.
The basis for this breakthrough stems from the discovery in 2018 that sound waves leave a sort of “fingerprint” on the water’s surface in the form of tiny vibrations corresponding to underwater signals. The team mounted a radar system on a drone to detect these surface vibrations and used algorithms to interpret the patterns, decode the signal, and extract the message.
Previously, the MIT system relied on advance knowledge of physical parameters, such as the transmission’s frequency and modulation type. In contrast, the latest research developed new algorithms that exploit the differences between radar and sonar to identify these parameters and decode the message without cooperation from the underwater transmitter.
To test their method, the researchers used an inexpensive commercial drone equipped with radar and conducted initial trials in a swimming pool. They placed a speaker underwater and flew the drone over the surface while swimmers provided interference. The drone sent brief radar chirps toward the water, and the reflected signals revealed the vibration patterns caused by the sound waves. These patterns were then analyzed and decoded by the system.
When the system was tested at Carnegie Lake in Princeton, it successfully identified unknown parameters and decoded messages from the underwater speaker, even with interference from wind and waves. One key achievement was determining the modulation type, a critical parameter, with 97.58% accuracy.
The research also found that the design parameters of underwater communication links greatly influence their susceptibility to such attacks. The paper includes recommendations for designing transmitters that are more resistant to eavesdropping, aiming to improve the security of underwater communications.
