Innovation

Home

Contacts

 

Introduction

Innovation

Objectives

Sea Trials

 

Underwater acoustic networks face different problems from hardwired or RF networks. These problems arise, mainly, from the very limited bandwidth, the large signal propagation times and the overload on a receiving element by the local transmit power levels (the near/far problem). The restricted bandwidth means that the use of multi-channel techniques is very limited and the near/far problem means that an acoustic unit may not transmit and receive at the same time. Consequently, a network protocol must carefully schedule transmissions to avoid packet collisions. Unfortunately, the large propagation delays involved in acoustic communications, which can be of the order of several seconds, make scheduling a far from simple problem. Several researchers have attempted to address some or all of these difficulties.

Some workers have tried to avoid these problems altogether by using hardwired, rather than acoustic, networks. This technique was applied by Kerfoot and Marra [1] who investigated the use of fibre-optic cables to implement underwater communication networks. However, this method was very expensive and required high maintenance, especially because sea bottom cables have high risk to be damaged. Webb [2] used abandoned undersea telephone cables to retrieve seabed seismology data from lower sea traffic areas. He states that the retrieval of data in the absence of such cables is an unsolved problem.

In terms of acoustic networks, Sayers et al [3] closely considered the transmission delay problem applied to the acoustic control of a remotely operated sub-sea vehicle. Their approach was a move and wait strategy, which results in an inefficient usage of the limited bandwidth available. In an attempt to overcome this, a simulated prediction of the effect of control inputs was implemented to give the operator a time-look-ahead view of the units estimated response. Such an approach would not be useful in a data transfer network.

A strategy of low intensity transmission may be used to limit the risk of packet collisions. Frye et al [4], working from Woods Hole in the United States, have implemented a coastal observatory consisting of underwater oceanographic measuring instruments. Several instruments send data by acoustic link to buoys, moored within a few km, the buoys relay this data to a land station by RF signal. The network uses a random access protocol in which the risk of packet collision is reduced by using a unidirectional communication links and limiting observations to a few per hour. In the absence of bi-directional transmission, necessary for re-transmission requests, marginal data or colliding data packets are lost, unless a bandwidth consuming, redundant transmission system is employed. As re-transmissions increase themselves the risk of packet collision, the optimal number of re-transmissions varies for different systems. This method is relatively cheap, easily maintainable and easy to deploy, but it is inefficient in use of the limited bandwidth and becomes high risk if significant data loss cannot be tolerated. A similar strategy has been employed by Spiess et al [5] who use three, seafloor transponders communicating with a surface link to measure the movement of oceanic tectonic plates.

Zvonar, Brady and Catipovic [6], again working in North America, investigated a method of reducing multi-user interference to allow packet collisions to be resolved, after finding that this is the dominant cause of network interference. They implemented an acoustic local area network, in which several underwater nodes transmit to one surface base station. Interference cancellation is carried out using an adaptive fractional transverse filter which was able resolve multi-user interference for two colliding packets. The filter also helped in resolving inter-symbol interference in the signal but did require a known training sequence for each underwater node. In [7] the same authors used a multi-dimensional decision feedback equaliser, in which the output from all active users is fed back to resolve multi-user interference.

Raysin et al [8] have developed an acoustic underwater network protocol to avoid collisions and ensure receipt, or to organise retransmit requests. A request-to-send clear-to-send protocol is used. In some cases a confirm-receipt signal is sent on successful data exchange and in others a retransmit-request is sent if data is not received. The system uses dead time, based on expected propagation delay, to indicate lost transmissions. Power control is achieved by increasing output power for retransmits. With this from of acoustic network protocol, in which data packets lengths are often less than propagation delays, control messages occupy more available transmission slots than data packets. This work provides no indication of how scheduling is achieved.

Hou, Hinton, Adams and Sharif [9] propose the use of a master node free, TDMA type protocol, in which packets are interleaved using the underwater channel as a FIFO buffer. The scheduling algorithm ensures that no node transmits when it is intended to receive. This system requires that propagation delays between nodes are known to the system at design time. It is also noted that the protocol is best suited to small networks and that some form of multi-user interference cancellation may be required but it does reduce the need for control signals.

Most of the current efforts in Europe, in order to further develop and improve shallow water acoustic communication (multi)-node links, are carried out within the framework of MAST III projects. These projects (named ROBLINKS, SWAN and LOTUS) will continue until the end of 2000. The ROBLINKS project [10][11][12][13][14][15] aims to develop robust (insensitive or self adaptive to changing conditions) communication algorithms to enable long range high data-rate acoustic communication in shallow water. The objective of the SWAN project [16][17][18]is to develop communication/protocol algorithms for shallow water acoustic networks. The LOTUS project [9][19][20] investigates long range acoustic networks. These three projects have advanced the state of the art in underwater acoustic communication with algorithmic developments tested during sea experiments.

It is however clear that the understanding of the design of efficient sub-sea acoustic networks is in its infancy and that much further work needs to be done. There is a particular need to develop efficient protocols that can reserve the maximum amount of bandwidth for data transfer rather than network control and cope with the volatility that is experienced in real sub-sea channels. The ACME project will further advance the state of the art in sub-sea networks by addressing, for the first time, the needs of a real data gathering problem in a very harsh environment. The few previous investigations into acoustic networks have been at a very primitive level. They have taken place in very benign circumstances and have been concerned mainly with the problem of recovering weak signals in the presence of stronger ones. Indeed very little consideration has yet been given to the particular problems of managing a real acoustic network, where some communications routes are likely to disappear for significant periods and where data must be sent via alternative routes in order to maintain the integrity of the system. The ACME project will address just these aspects, by developing techniques around a real scenario, where signal paths will be subject, not only to the normal changes in channel conditions such as are found in any sub-sea environment, but also to interference from vessels passing close to the network nodes, from the acoustic noise and turbulence such vessels generate and from particles or gas bubbles forced into the water column by their passage. This will require new network protocols to be developed if acceptable data transfer rates are to be maintained under such difficult conditions.

Currently, apart from the present, MAST funded projects due for completion at the end of 2000, all work in this area is being carried out in North America. It is thus essential that Europe continues with research work into sub-sea acoustic networks, in order to provide European industry with a foundation upon which to develop products for the future exploitation of a world-wide market.


From an end users point of view, it is known that today's instruments and techniques are not capable of delivering a solution for monitoring the environment in many European coastal areas of interest. For example measuring the current in the middle of shipping lanes is now almost impossible. Options with cables are expensive and not long lasting because the cables are torn apart within days, buoys cant be placed in the middle of a shipping lane and a sensor pole near the shore is expensive and doesn't give accurate data. Even when real time collection of data is not necessary, it is most often crucial, for economic and operational reasons, to monitor the state of the sensors which compose the measurement network. For example, projects for modeling water quality partly failed in the past, due to the lack of available data, because the failure of half of the sensors could not be detected [21][22].

It is clear that the success of an acoustic network, like the network prototype that ACME aims at designing, building and testing, would open the door for a series of innovative applications in the fields of safety, environment or inspection of underwater structures, that were not even conceivable in the past.

References

  1. F.W. Kerfoot and W.C. Marra. "Undersea fiber optic networks: Past, present, and future", IEEE Journal on Selected Areas in Communications, 1998, Vol. 16, No. 7, pp. 1220-1225
  2. S.C. Webb. "Broadband seismology and noise under the ocean", Reviews of Geophysics, 1998, Vol. 36, No. 1, pp. 105-142
  3. C.P. Sayers, R.P. Paul, L.L. Whitcomb and D.R. Yoerger. "Teleprograming for subsea teleoperation using acoustic communication", IEEE Journal of Oceanic Engineering, 1998, Vol. 23, No. 1, pp. 60- 71
  4. D Frye, K. von der Heydt, M. Johnson, A. Maffei, S. Lerner and B. Butman. "New technologies for coastal observatories", Sea Technology, 1999, Vol. 40, No. 10, pp. 29-35
  5. F.N. Spiess, C.D. Chadwell, J.A. Hilderbrand, L.E. Young, G.H. Purcell and H. Dragert. "Precise GPS/Acoustic positioning of subsea reference points for tectonic studies", Physics of the Earth and Planetary Interiors, 1998, Vol. 108, No. 2, pp. 101-112
  6. Z. Zvonar, D. Brady and J. Catipovic. "An adaptive decentralized multiuser receiver for deep-water acoustic telemetry", The Journal of the Acoustical Society of America, 1997, Vol. 101, No. 4, pp. 2384-2387
  7. Z. Zvonar, D. Brady and J. Catipovic. "Adaptive detection for shallow-water acoustic telemetry with cochannel interference", IEEE Journal of Oceanic Engineering, 1996, Vol.21, No.4, pp.528-536
  8. K. Raysin, J.R. Rice, E. Dorman and S. Matheny. "Telesonar network modeling and simulation", Proc. Oceans'99 Seattle, 1999
  9. B. Hou, O.R. Hinton, A.E. Adams and B.S. Sharif, "A time-domain-orientated multiple access protocol for underwater acoustic network communications", Proc. Oceans'99 Seattle, 1999
  10. Daniel Cano, Martin van Gijzen, Andreas Waldhorst, "Long Range Shallow Water Robust Acoustic Communication Links ROBLINKS", In Third European Marine Science and Technology Conference, Lisbon, Volume III, pp. 1133-1136, 1998
  11. M.B. van Gijzen, P.A. van Walree, D. Cano, J.M. Passerieux, A. Waldhorst, R.B. Weber, C. Maillard, "The ROBLINKS underwater acoustic communication experiment", To appear in The proceedings of the fifth European Conference on Underwater Acoustics, 2000, Lyon, 2000
  12. P.A. van Walree, M.B. van Gijzen, D.G. Simons, "Analysis of a shallow-water acoustic communication channel", To appear in The proceedings of the fifth European Conference on Underwater Acoustics, 2000, Lyon, 2000
  13. J.M. Passerieux and D. Cano, "Robust shallow water acoustic communications based upon orthogonal sequences and real-time channel identification", To appear in The proceedings of the UDT Europe Conference, 27-29 June 2000, London (UK), 2000
  14. M.B. van Gijzen and P.A. van Walree, "Shallow-water acoustic communication with high bit rate BPSK signals", To appear in The proceedings of the Oceans 2000 Conference, Providence (USA), 2000
  15. A. Waldhorst, R. Weber and J.F. Böhme, "A Blind Receiver For Digital Communications in Shallow Water" To appear in The proceedings of the Oceans 2000 Conference, Providence (USA), 2000
  16. H.K. Yeo, B.S. Sharif, O.R. Hinton, A.E. Adams, "Analysis of a Multi-element Multi-user receiver for Shallow water Acoustic Network (SWAN) based on Recursive Successive Inteference Cancellation (RSIC)", Proc Oceanic'99 Conf, Seattle, Sept 1999
  17. H.K. Yeo, B.S. Sharif, O.R. Hinton, A.E. Adams, "Improved RLS Algorithm for Time Variant Underwater Acoustic Communications", Electronics Letters, Vol 36, No 2, 20th Jan 2000, pp191-2
  18. H.K. Yeo, B.S. Sharif, O.R. Hinton, A.E. Adams, "Multi-User Detection for Time Variant Multipath Environment", Proc. 2000 |Int Symp. on Underwater Tech, Tokyo, May 2000, pp399-404
  19. Tsimenidis C.C., Hinton O.R., Sharif B.S., Adams A.E.(25%), "An Adaptive Array Direct Sequence Code Division Multiple Access (DS-CDMA) Receiver for Shallow Water Asynchronous Multiuser Networks", Proc IEEE Oceans '99, Session 3D, Seattle, Sept. 1999
  20. A.E. Adams, O.R. Hinton, B.S. Sharif, G. Salles, N. Orr and C. C. Tsimenidis " An Experiment in Sub Sea Networks - The LOTUS Sea Trials", accepted for 5th. European Conference on Underwater Acoustics, 2000, Lyon, 2000
  21. "Sea grass Project", North Sea Experiments to evaluate effect of pollution on sea grass , 1995-1998
  22. "Sand transport Project", North Sea Experiments to evaluate effect of currents and storms on sand transport near the coast, 1992-199