Feature ArticleUnderwater Optical Communication Using a Modulating Retroreflector
By William Cox
North Carolina State University
Department of Electrical and Computer Engineering
Raleigh, North Carolina
Communicating with underwater sensors and vehicles is an integral part of comprehensive ocean exploration and observation. Due to the short propagation length of radio frequency (RF) waves in underwater environments, current methods of communicating underwater are limited to tethered cables, acoustic pressure waves or optical communication at visible wavelengths. Substantial work has been invested in furthering acoustic communication technology, but acoustic systems have limited bandwidth. Tethered communications are viable for stationary objects or for point-to-point systems, such as a surface vessel to a remotely operated vehicle, but are impractical for many types of autonomous systems or sensor networks.
As a practical consideration, making physical connections underwater can be difficult, and wireless systems add a level of operational convenience. Thus, underwater optical communication can be a good method for underwater communication at short ranges when the application requires high data rates, low latency or covert operation. These systems are limited by optical scattering and absorption in the underwater environment. In clear waters, point-to-point links can have ranges approaching 200 meters while in littoral waters, shorter ranges are practical. For point-to-point applications, transmitters can be placed on each platform, but sensor nodes and small autonomous vehicles are often power- and size-limited. For short ranges, one way to reduce the burden of having a pointing and tracking system and a transmitter is to use a modulating retroreflector (MRR) on one platform.
Picture of modulators mounted to a printed circuit board. Each glass substrate has 102 modulators patterned. (Photo credit: Jim Simpson)
Use of a Modulating Retroflector
An MRR is an optical device that imposes data on an interrogating light and reflects the light back to its source. This allows the burden of the light source and pointing system to be placed on the interrogating platform, such as a submarine or unmanned underwater vehicle. The power-constrained platform, such as a sensor node or small unmanned vehicle, hosts the MMR. Such an arrangement allows a system to have duplex optical communication at low power.
MRRs have been used in many terrestrial applications for providing optical communication. Several methods have been created, including a quantum well that operates at infrared wavelengths or microelectromechanical systems (MEMS) corner cube design where the surface is moved to modulate the return light. Most of these devices are specifically designed to work at infrared wavelengths and are therefore unsuitable for underwater environments that require light sources in the 'blue-green' window for transmission. To overcome this limitation, researchers built an MRR that uses a MEMS Fabry-Perot optical cavity for modulating light in the visible wavelengths for communication underwater. By electrically changing the cavity spacing, the intensity of the reflected or transmitted light can be modulated, which allows the device to utilize many different types of RF communication schemes. A basic experiment with quadrature phase-shift keying (QPSK) modulation demonstrated data transfer at rates of 250 kilobits per second, 500 kilobits per second and one megabit per second.
Making the Modulator
The modulator is fabricated using bulk micromachining. Aluminum is patterned on a one-by-one-inch glass substrate and a silicon nitride membrane to form a Fabry-Perot cavity, which consists of two partial mirrors separated by a small distance. The two mirrors are separated by indium bump bonds. The indium is used to bond the samples to each other and to provide an electrical connection to the silicon nitride mirror layer. The lateral dimensions of the Fabry-Perot cavity are typically 250-by-250 micrometers but can be larger or smaller depending on the desired data rate or optical configuration. Multiple membranes are created on each sample for testing. The glass substrate is mounted to a printed circuit board for easy electrical signaling.
By applying a voltage across the two plates of the mirror, electrostatic forces draw the silicon nitride membrane toward the aluminum-on-glass plate. This change in mirror distance shifts the cavity spectrum and allows the total reflectivity or transmission of the device to be altered. By applying an AC voltage across the cavity, the retroreflector can modulate the amount of light that is reflected at a certain wavelength by moving from a peak to a valley in the Fabry-Perot transmission spectrum. Since there is no current flowing through the cavity, the MRR dissipates very little power.
A cross-sectional diagram of retroreflector along with illustration of retroreflection link. Not to scale. (Photo credit: William Cox)
In order to test the modulator, a basic communication experiment was carried out involving QPSK modulation through a water tank. This modulation type was chosen in order to maximize the amount of data transferred given the bandwidth of the device. For the tested devices, the bandwidth was almost two megahertz, allowing a data rate of up to one megabit per second.
Data bits were generated on a PC and streamed via USB to a custom modulator that then used an arbitrary waveform generator to create a phase-modulated signal with a carrier frequency of 750 kilohertz. Individual data bits were mapped to shifts in phase to produce the signal. This modulation signal was combined with a DC voltage offset to tune the bridge to the chosen wavelength. The DC voltage produces a static deflection of the bridge cavity, and the modulation signal causes small deflections about this set point. The output signal was then amplified by a high-speed, high-voltage amplifier to electrostatically move the membrane.
The MRR, in a retroreflecting configuration, was mounted at one end of a water tank filled with municipal water. Maalox (aluminum-magnesium hydroxide), a commercial antacid typically used for simulating the scattering effects of ocean water in laboratory experiments, was added to the tank. The attenuation coefficient (light loss per unit distance) was measured using a C-Star transmissometer from WET Labs (Philomath, Oregon). A 532-nanometer solid-state laser was used to interrogate the retroreflecting modulator. The reflected light, modulated with the data, is returned along the same path and is collected by a two-inch parabolic mirror that focuses the return light onto a photodetector. The signal is digitized, demodulated and compared with the original signal to obtain bit error rate.
Using laboratory test tanks, the MRR was able to be tested at two distances, 3.7 meters and 7.7 meters. A 20-milliwatt laser was used for the tests, and the chosen wavelength was 532 nanometers (green). For each packet of data that was transmitted through the water, the water quality, measured in terms of its attenuation coefficient per meter, was measured along with the number of errors in the packet. A simple Reed-Solomon forward error correction scheme was also used to encode the data and correct some errors at the receiver. Using this scheme, the system was able to achieve error-free data transmission of up to 500 kilobits per second at 6.5 attenuation lengths, which is the multiple of the transmission distance and the attenuation coefficient of the water. At one megabit per second, the system was able to transmit data error-free at up to 5.5 attenuation lengths at 3.7 meters and up to 3.8 attenuation lengths over the 7.7 meter path. Since each attenuation length represents a loss of a power of e (63 percent, 86 percent, 95 percent, etc.), the received optical signal power at 6.5 attenuation lengths is close to 30 microwatts. More sophisticated error-correcting techniques would extend the system's performance and the optical power of the interrogating system. More recently at North Carolina State University, compressed video data have been streamed from the retroreflector through the 3.7-meter tank.
This experiment demonstrated an MRR in the blue-green portion of the spectrum. It is envisioned such a device could easily be integrated into low-cost sensor nodes or autonomous vehicles to allow for high data-rate communication underwater. This would enable covert retrieval of much larger datasets than are feasible with existing wireless communication methods underwater. The low power requirements of these devices will also enable such sensors or vehicles to have a longer operational lifetime.
The authors would like to thank Jim Simpson, Brandon Cochenour and Dr. Brian Hughes for their input into the project and Dr. Linda Mullen at Pautuxent River Naval Air station for access to the large tank for testing. Discussion with Dr. William Rabinovich at Naval Research Laboratory was appreciated. This work was supported by the Office of Naval Research under grant Small Business Technology Transfer N00014-07-M-0308 and by the National Science Foundation under grants CCF-0515164 and ECCS-0636603.
William Cox is a Ph.D. candidate at North Carolina State University specializing in underwater optical communication. He has an interest in digital communication and robotics.
Kory Gray is a Ph.D. student at North Carolina State University specializing in nanoelectronics and photonics. He has an interest in microelectrical mechanical system devices and is president of the triathlon club at his university.
Dr. John Muth is an associate professor of electrical engineering at North Carolina State University. He has published more than 100 papers in nanotechnology and underwater communications.