1. Underwater Networking and MUAC
Masked Underwater Acoustic Communication
POCs:
Justin Rohrer [PI]
Carl Prince

2. Importance
The underwater domain is becoming an increasingly important area for the Navy. There are important systems at sea that can not give away their location:
Underwater Gliders
Underwater sensors
Underwater drones
Manned submarines
There needs to be a network to land in order to facilitate information this network may involve routers/hubs, etc.
Modeling this network is important

3. Underwater Networking
Create an underwater software defined network U-SDN, modeling underwater delays include detectors both fixed and movable
Map out how devices will connect and communicate to the land
Design of infrastructure

4. Use cases
Sensors transmitting OK -- location and good operation, nothing found – can be performed with very minimal data rates, is performed relatively often compared to sending detection data, it is important that this is performed so as not to be detected
Transmitting detection data – more important data, non-trivial data amounts may need to be sent relatively quickly

5. MUAC Masking Sounds
Since some underwater devices can not give away their location we must communicate in a way such that the communication is undetectable, this can be performed using masking sounds:
Biologics
dolphins
shrimp
whales
crabs
other
Naturally occurring sounds
bubbles
water
underwater background noise
earth quake noise

6. Acoustical Communication
Native acoustic communication can be performed by
Putting together sounds to send symbols
Using sound segments to send multiple symbols
Data rates can be more than ‘transmitting OK’
Native acoustics as a carrier wave with steganographic data sent on the carrier wave to achieve a low probability of detection [Passerieux]
Data rates are very slow, 1-4 symbols per sec, but very hard to detect
Can send on naturally occurring sounds
travel distance can be characterized as shorter
Can send on biologics
Travel distance can be longer
Current method based on Passerieux Patent US A1

7. Current Work
Focused on steganography, or the data hiding art of hiding messages in plain sight.
Use of spread spectrum using steganography to send messages
Steganographic algorithm is implemented by hiding bits/symbols in phase centered on a carrier frequency (signal processing).
The carrier transmission masks the underlying message.
Carrier transmissions are previously recorded or database stored sound clips in order to be representative of current environment and background noise conditions – can be ambient discrete, or continuous, and natural, or biologics
Exploring effects of modulation schemes on data rates and imperceptibility.
Goal is to send data and to recover messages, while at the same time being undetected by a trained adversary sonar operator with time-frequency displays.
Optimize which carrier sounds act as the best covers and trade offs of distance traveled vs. perceptibility.

8. MUAC Future Work
There is no one chosen model in published work, there are many models used to model communication, is there an optimal model – a model that is accurate and able to model transmission using time and frequency domains?
The current method the student has working is a bi-polar model, can we use a quadrature-polar model?
Currently we use 360 degrees as one cycle, can we create a timing algorithm to include non-360 degree cycles to make detection even harder?
Are there better steganographic or covert waveform approaches to Passerieux’s covert UAC?
Can we make an algorithm or modulation scheme improvement(s) for robustness, embedding efficiency, imperceptibility?
How good is our method of steganalyis – need a detailed statistical analysis of artifacts as a result of the embedding algorithm, and or how to attack this scheme?
Will our method actually work at sea – testing?
Can this scheme be implemented -- take the results of this study, once proven feasible, and find a way to bridge the gap with a program of record Navy system onboard ships, submarines, sonobouys, etc.