Underwater Acoustic Communications

The US Navy is increasingly looking to underwater acoustic communications as the basis for wireless networks for various missions. However, the underwater channel is a difficult one for communications, plagued by multipath and fading. The relatively slower propagation speed (compared to RF) makes Doppler a significant factor for mobile platforms and when signals are scattered from an ocean surface in motion due to waves. The bandwidth available for communication is severely limited by volume attenuation (to perhaps 100 kHz at ranges of 1-3 km) and our ability to design wideband transmitters. Current experimental signaling schemes for point-to-point communications achieve no more than 20 kbps*km throughput under the best of conditions, and are generally processed off-line in the laboratory. Commercial modem systems achieve much less throughput, with at times lower reliability. Multiple input multiple output (MIMO) signaling is an exciting new technology, only recently being adopted in RF wireless applications (e.g. 802-11n), that significantly improves communications performance and reliability. This MIMO (multiple input / multiple output) technique exploits a "rich" multipath environment by simultaneously transmitting across many independent virtual channels using the entire band (i.e. without multiplexing in time or frequency). Spectral efficiencies of 20-40 bits per second per Hertz have been demonstrated in the wireless community - such rates are not possible using traditional techniques. How this technique can be adapted to the underwater channel is the subject of this proposal. We propose to develop a MIMO modem that provides either significantly increased bit rates or greatly reduced error rates, depending on the modulation scheme, in typical shallow water ocean environments. In our Phase I effort, we will: 1) continue collecting data at sea as we push the performance envelope of our current MIMO algorithms, space-time trellis coding (STTC) with joint equalization and decoding and layered space time coding (LSTC) with iterative decoding and equalization, 2) develop strategies for making the modem operate autonomously, using in-modem channel measurements as the basis for automating all modem parameter settings, 3) develop a preliminary hardware design, to establish the feasibility of implementing the modem on a single board with FPGAs and DSP chips, to be ultimately integrated with a next-generation commercial modem.