High fidelity simulation tools for performance assessment of underwater acoustic communications modems

New attractive techniques, such as multi-carrier modulations, efficient channel coding, iterative reception algorithms, and use of multiple antennas at reception or transmission, are now emerging in underwater acoustic communications (UAC). According to the published results, these techniques could provide, within the next years, a significant breakthrough with respects to the critical issue of robust and high bitrate (a few kbits/sec) tactical communications with submarines and/or AUV). Thales Underwater Systems (TUS), as a UAC equipments manufacturer and integrator, is therefore very interested in performance assessment of these techniques, either by implementing the corresponding signaling schemes and algorithms, or by testing them as black boxes from external suppliers. As long as possible this assessment has to be done without redoing expensive at sea experiments, but by using simulated data based upon previous experiments and/or high fidelity acoustic propagation modeling. A critical point is that UAC is likely one the most demanding applications for underwater acoustic propagation modeling. Indeed, apart from an accurate knowledge of the Signal-to-Noise ratio, the prediction and simulation of the UAC channel require a comprehensive modeling of the its time and spatial fluctuations (UAC channel modeled as a doubly dispersive Ricean channel). Constraints on the equipments (e.g. rms and peak transmitted power) have also to be taken into account. Specific tools which address the above aspects will be presented, from the computation of Shannon channel capacity under the above assumptions to high fidelity simulations tools relying on an explicit modeling of scattering of sound by the moving rough sea surface (e.g. NARCISSUS [1] developed by TUS). Emphasis will finally be laid upon a new innovative “stochastic replay” technique [2], which consists in simulating a random channel whose response has exactly the same statistical properties as a previously at sea recorded channel response, thus allowing perfect comparison of different modulation schemes or trimmings under the same environmental conditions.