ROBLINKS contibutes to the objectives indicated in the MAST III Work Programme, Research Area C, 1 ask 1.2 (Generic Technology in Underwater Communication). It aims at developing and testing robust coherent acoustic communications at long ranges (tens of km) in shallow water (20/30 m) at data rates in excess of 1 kbit/s. Here, robust means insensitive to the acoustic channel conditions. Present communication systems have good performances but either they have poor bandwidth eff1ciency (non coherent methods) or they are limited by time spreading which is less than the signal duration (differentially coherent) or they require operator assistance to adjust the receiver's parameters to the channel characteristics (coherent methods with standard equalization approach). A robust coherent communication system should fulfil a very large range of possible missions which involve remote control of underwater instrumentation and vehicles or collection of data from undersea equipment: surveying and monitoring of the marine environment in shallow water (coastal zone) but also in the deep sea, sea resources exploitation (off-shore inspection), etc. The method to reach this objective is:
1. To develop signal concepts and algorithhms based on continuous identification of the channel response The channel identification process provides automatic assistance to drive the parameters of the symbol restitution process. Two competing strategies will be investigated.
The first one uses a permanent monitoring signal, which avoids the traditional learning sequences periodically injected into the communication channel as used typically in the equalization approach. The idea is that this permanent signal should firstly allow to track very fast variations of the channel response and secondly exclude symbol errors from the channel estimation process. The loss of signalling space devoted to the communication is recovered by a better channel identification ability and a better reliability, which in turn allows to use efficient coding schemes.
The second one is blind identification. The a priori knowledge about the distribution properties is exploited. The Constant Modulus Algorithm (CMA) and fourth order cumulant based methods fit into this framework. The single input, multiple output (SIMO) setting will be considered. We expect to develop new methods performing better than the existing algorithms and having moderate computational cost for on-line implementation.
2. To evaluate experimentally the performance of these waveforms and processing algorithms under real shallow water conditions (North Sea) and compare their performance to a well established existing technique appropriate for this communication situation. Selected waveforms and processing algorithms will be implemented on a real-time system and the real-time performance will be evaluated during sea trials.
The partnership is composed of three members of three different countries complementary to a large extent: Partner 1 is a large company with experience both in acoustic signal processing and communication theory and system development.
Partner 2 is a large research institute with experience both in underwater acoustic communication and field experiments, especially in North Sea coastal area. Both Partner I and Partner 2 are directly interested in future product developments.
Partner 3 is an university, which will contributes its expertise in use of the underwater channel model and its knowledge of mathematic tools specific to ROBLINKS techniques (system identification, high order statistics, non-stationary signal analysis, etc.).
Funding SchemeCSC - Cost-sharing contracts
2509 JG Den Haag ('S-gravenhage)