Underwater Acoustic Sensor Network (UASN)
Disclaimer: This dissertation has been submitted by a student. This is not an example of the work written by our professional dissertation writers. You can view samples of our professional work here.
Any opinions, findings, conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of UK Essays.
Most of the earth surface is composed of water including fresh water from river, lakes etc and salt water from the sea. There are still many un-explored areas for such places. This needs significant research efforts and good communication systems. Wireless sensor network in aqueous medium has the ability to explore the underwater environment in details. For all applications of underwater, a good communication system as well as an effective routing protocol is needed. This will enable the underwater devices to communicate precisely. Underwater propagation speed varies with temperature, salinity and depth. By varying the underwater propagation speed at different depth, two scenarios can be achieved accurately namely: shallow and deep water. Shallow water consists of depth less than 200m and cylinder spreading. Deep water consists of depth greater or equal to 200 m and spherical spreading. In both shallow and deep water, different ambient noise and different spreading factor is applied.
CHAPTER 2: Study of Underwater Acoustic Sensor Network (UASN)
Application of UASN
Wireless sensor network in aqueous medium also known as underwater sensor network has enabled a broad range of applications including:
- Environmental Monitoring
Underwater sensor network can be used to monitor pollution like chemical, biological such as tracking of fish or micro-organisms, nuclear and oil leakage pollutions in bays, lakes or rivers . Underwater sensor network can also be used to improve weather forecast, detect climate change, predict the effect of human activities on marine ecosystems, ocean currents and temperature change e.g. the global warming effect to ocean.
- Under Ocean Exploration
Exploring minerals, oilfields or reservoir, determine routes for laying undersea cables and exploration valuable minerals can be done with such underwater sensor network.
- Disaster Prevention
Sensor network that measure seismic activity from remote locations can provide tsunami warning to coastal areas, or study the effects of submarine earthquakes (seaquakes) 
- Equipment Monitoring
Long-term equipment monitoring may be done with pre-installed infrastructure. Short-term equipment monitoring shares many requirements of long-term seismic monitoring, including the need for wireless (acoustic) communication, automatic configuration into a multihop network, localization (and hence time synchronization), and energy efficient operation
- Mine Reconnaissance
By using acoustic sensors and optical sensors together, mine detection can be accomplished quickly and effectively.
- Assisted Monitoring
Sensor can be used to discover danger on the seabed, locate dangerous rocks or shoals in shallow waters, mooring position, submerged wrecks and to perform bathymetry profiling.
- Information collection
The main goal of communication network is the exchange of information inside the network and outside the network via a gateway or switch center. This application is used to share information among nodes and autonomous underwater vehicles.
Characteristic of UASN
Underwater Acoustic Networks (UANs), including but not limited to, Underwater Acoustic Sensor Networks (UASNs) and Autonomous Underwater Vehicle Networks (AUVNs) , are defined as networks composed of more than two nodes, using acoustic signals to communicate, for the purpose of underwater applications. UASNs and AUVNs are two important kinds of UANs. The former is composed of many sensor nodes, mostly for a monitoring purpose. The nodes are usually without or with limited capacity to move. The latter is composed of autonomous or unmanned vehicles with high mobility, deployed for applications that need mobility, e.g., exploration. An UAN can be an UASN, or an AUVN, or a combination of both.
Acoustic communications, on the other hands, is defined as communication methods from one point to another by using acoustic signals. Network structure is not formed in acoustic point-to-point communications.
Sound travels best through the water in comparison with electromagnetic waves and optical signals. Acoustic signal is sound signal waveform, usually produced by sonar for underwater applications. Acoustic signal processing extracts information from acoustic signals in the presence of noise and uncertainty.
Underwater acoustic communications are mainly influenced by path loss, noise, multi-path, Doppler spread, and high and variable propagation delay. All these factors determine the temporal and spatial variability of the acoustic channel, and make the available bandwidth of the Underwater Acoustic channel (UW-A) limited and dramatically dependent on both range and frequency. Long-range systems that operate over several tens of kilometers may have a bandwidth of only a few kHz, while a short-range system operating over several tens of meters may have more than a hundred kHz bandwidth. These factors lead to low bit rate.
Underwater acoustic communication links can be classified according to their range as very long, long, medium, short, and very short links. Acoustic links are also roughly classified as vertical and horizontal, according to the direction of the sound ray. Their propagation characteristics differ consistently, especially with respect to time dispersion, multi-path spreads, and delay variance.
Acoustic signal is the only physical feasible tool that works in underwater environment. Compared with it, electromagnetic wave can only travel in water with short distance due to the high attenuation and absorption effect in underwater environment. It is found that the absorption of electromagnetic energy in sea water is about 45× ?f dB per kilometer, where f is frequency in Hertz; In contrast, the absorption of acoustic signal over most frequencies of interest is about three orders of magnitude lower .
Hereafter the factors that influence acoustic communications is analyzed in order to state the challenges posed by the underwater channels for underwater sensor networking. These include:
- Path loss
- Attenuation is mainly provoked by absorption due to conversion of acoustic energy into heat, which increases with distance and frequency. It is also caused by scattering a reverberation (on rough ocean surface and bottom), refraction, and dispersion (due to the displacement of the reflection point caused by wind on the surface). Water depth plays a key role in determining the attenuation.
- Geometric Spreading is the spreading of sound energy as a result of the expansion of the wavefronts. It increases with the propagation distance and is independent of frequency. There are two common kinds of geometric spreading: spherical (omni-directional point source), and cylindrical (horizontal radiation only).
- Man made noise is mainly caused by machinery noise (pumps, reduction gears, power plants, etc.), and shipping activity (hull fouling, animal life on hull, cavitations), especially in areas encumbered with heavy vessel traffic.
- Ambient Noise is related to hydrodynamics (movement of water including tides, current, storms, wind, rain, etc.), seismic and biological phenomena.
- Multi-path propagation may be responsible for severe degradation of the acoustic communication signal, since it generates Inter-Symbol Interference (ISI).
The multi-path geometry depends on the link configuration. Vertical channels are characterized by little time dispersion, whereas horizontal channels may have extremely long multi-path spreads.
The extent of the spreading is a strong function of depth and the distance between transmitter and receiver.
High delay and delay variance
- The propagation speed in the UW-A channel is five orders of magnitude lower than in the radio channel. This large propagation delay (0.67 s/km) can reduce the throughput of the system considerably.
- The very high delay variance is even more harmful for efficient protocol design, as it prevents from accurately estimating the round trip time (RTT), which is the key parameter for many common communication protocols.
- Doppler spread
- The Doppler frequency spread can be significant in UW-A channels, causing degradation in the performance of digital communications: transmissions at a high data rate because many adjacent symbols to interfere at the receiver, requiring sophisticated signal processing to deal with the generated ISI.
- The Doppler spreading generates:
- a simple frequency translation, which is relatively easy for a receiver to compensate for
- a continuous spreading of frequencies, which constitutes a non-shifted signal, which is more difficult for a receiver to compensate for.
If a channel has a Doppler spread with bandwidth B and a signal has symbol duration T, then there are approximately BT uncorrelated samples of its complex envelope. When BT is much less than unity, the channel is said to be under spread and the effects of the Doppler fading can be ignored, while, if greater than unity, it is overspread.
Most of the described factors are caused by the chemical-physical properties of the water medium such as temperature, salinity and density, and by their spatio-temporal variations. These variations, together with the wave guide nature of the channel, because the acoustic channel to be temporally and spatially variable. In particular, the horizontal channel is by far more rapidly varying than the vertical channel, in both deep and shallow water.
CHAPTER 3: Network Architecture
Underwater sensor nodes: The underwater sensor nodes are deployed on the sea floor anchored to the ocean bottom . The sensors are equipped with floating buoys to push the nodes upwards, thus they are relatively stationary nodes . Using acoustic links, they relay data to underwater sink directly or via multi-hop path.
Underwater sink nodes: Underwater sink nodes take charge of collecting data of underwater sensors deployed on the ocean bottom and then send to the surface sink node. They may be equipped with vertical and horizontal acoustic transducers. The horizontal transceiver is used to collect the sensors' data and the vertical transceiver provides transmitting link between underwater sink and the surface sink node.
Surface sink node: Surface sink node is attached on a floating buoy with satellite, radio frequency (RF) or cell phone technology to transmit data to shore in real time.
A reference architecture for two-dimensional underwater networks is shown in Figure. 1. A group of sensor nodes are anchored to the deep of the ocean. Underwater sensor nodes are interconnected to one or more underwater gateways by means of wireless acoustic links. Underwater-gateways are network devices in charge of relaying data from the ocean bottom network to a surface station. To achieve this objective, they are equipped with two acoustic transceivers, namely a vertical and a horizontal transceiver. The horizontal transceiver is used by the underwater-gateway to communicate with the sensor nodes in order to:
- send commands and configuration data to the sensors (underwater -gateway to sensors);
- collect monitored data (sensors to underwater -gateway). The vertical link is used by the underwater -gateways to relay data to a surface station.
In deep water applications, vertical transceivers must be long range transceivers. The surface station is equipped with an acoustic transceiver that is able to handle multiple parallel communications with the deployed underwater -gateways. It is also endowed with a long range RF and/or satellite transmitter to communicate with the onshore sink (os-sink) and/or to a surface sink (s-sink). In shallow water, bottom-deployed sensors/modems may directly communicate with the surface buoy, with no specialized bottom node (underwater -gateway).
Three-dimensional underwater networks are used to detect and observe phenomena that cannot be adequately observed by means of ocean bottom sensor nodes, i.e., to perform cooperative sampling of the 3D ocean environment. In three-dimensional underwater networks, sensor nodes float at different depths to observe a phenomenon. In this architecture, given in Figure 2, each sensor is anchored to the ocean bottom and equipped with a floating buoy that can be inflated by a pump. The buoy pushes the sensor towards the ocean surface. The depth of the sensor can then be regulated by adjusting the length of the wire that connects the sensor to the anchor, by means of an electronically controlled engine that resides on the sensor. Sensing and communication coverage in a 3D environment are rigorously investigated in . The diameter, minimum and maximum degree of the reachability graph that describes the network are derived as a function of the communication range, while different degrees of coverage for the 3D environment are characterized as a function of the sensing range.
3D Model with AUV
The above figure represents the third type of network architecture which consist of sensor nodes and Autonomous Underwater Vehicles (AUV) which act as mobile sensor nodes for ocean monitoring, underwater resource study, etc.
CHAPTER 4: Differences between underwater and terrestrial Sensor Network
An underwater acoustic channel is different from a ground-based radio channel from many aspects, including:
- Bandwidth is extremely limited. The attenuation of acoustic signal increases with frequency and range  . Consequently, the feasible band is extremely small. For example, a short range system operating over several tens of meters may have available bandwidth of a hundred kHz; a medium-range system operating over several kilometers has a bandwidth on the order of ten kHz; and a long-range system operating over several tens of kilometers is limited to only a few kHz of bandwidth .
- Propagation delay is long. The transmission speed of acoustic signals in salty water is around 1500 meter/s , which is a difference of five orders of magnitude lower than the speed of electromagnetic wave in free space. Correspondently, propagation delay in an underwater channel becomes significant. This is one of the essential characteristics of underwater channels and has profound implications on localization and time synchronization.
- The channel impulse response is not only spatially varied but also temporarily varied. The channel characteristics vary with time and highly depend on the location of the transmitter and receiver. The fluctuation nature of the channel causes the received signals easily distorted. There are two types of propagation paths: macro-multipaths, which are the deterministic propagation paths; and micro-multipath, which is a random signal fluctuation. The macro-multipaths are caused by both reflection at the boundaries (bottom, surface and any object in the water) and bending. Inter- Symbol Interference (ISI) thus occurs. Compared with the spread of its ground-based counterpart, which is on the order of several symbol intervals, ISI spreading in an underwater acoustic channel is several tens or hundred of symbol intervals for moderate to high data rate in the horizontal channel. Micro-multipath fluctuations are mainly caused by surface wave, which contributes the most to the time variability of shallow water channel.
- In deep water, internal waves impact the single-path random fluctuations .
- Probability of bit error is much higher and temporary loss of connectivity (shadow zone) sometimes occurs, due to the extreme characteristics of the channel.
- Cost. While terrestrial sensor nodes are expected to become increasingly inexpensive, underwater sensors are expensive devices. This is especially due to the more complex underwater transceivers and to the hardware protection needed in the extreme underwater environment. Also, because of the low economy of scale caused by a small relative number of suppliers, underwater sensors are characterized by high cost.
- Deployment. While terrestrial sensor networks are densely deployed, in underwater, the deployment is generally more sparse.
- Power. The power needed for acoustic underwater communications is higher than in terrestrial radio communications because of the different physical layer technology (acoustic vs. RF waves), the higher distances, and more complex signal processing techniques implemented at the receivers to compensate for the impairments of the channel.
- Memory. While terrestrial sensor nodes have very limited storage capacity, underwater-sensors may need to be able to do some data caching as the underwater channel may be intermittent.
- Spatial Correlation. While the readings from terrestrial sensors are often correlated, this is more unlikely to happen in underwater networks due to the higher distance among sensors.
CHAPTER 5: Layered of UASN
The underwater architecture network consists of five layers, application, transport, network, data link and physical layer as shown in the figure below. As typical underwater systems have limited processing capability, the protocol has been kept as simple as possible without significantly compromising performance.
The underwater sensor network specifications currently do not include any recommendations for authentication and encryption. These may be easily implemented at the application layer or via a spreading scheme at the physical layer.
Each layer is described by a SAPI. The SAPI is defined in terms of messages being passed to and from the layer. The clients (usually higher layers) of a layer invoke the layer via a request (REQ). The layer responds to each REQ by a response (RSP). Errors are reported via an ERR RSP with error codes. If the layer needs to send unsolicited messages to the client, it does so via a notification (NTF). A layer communicates logically with its peer layer via protocol data units (PDU). As the peer-to-peer communication is symmetric, a layer may send a REQ PDU to its peer layer at any time. It would optionally respond to such a PDU with a RSP PDU. This is logically depicted in Figure below
It may be desirable in some cases, that non-neighboring layers communicate with each other to achieve cross-layer optimization. This may be implemented by allowing REQ and RSP PDUs between any two layers in the protocol stack.
The underwater sensor network specifications define detailed message structures for all SAPI messages. These message structures include message identifiers, data formats to be used, parameters and their possible values
The physical layer provides framing, modulation and error correction capability (via FEC). It provides primitives for sending and receiving packets. It may also provide additional functionality such as parameter settings, parameter recommendation, carrier sensing, etc.
At first underwater channel development was based on non-coherent frequency shift keying (FSK) modulation, since it relies on energy detection. Thus, it does not require phase tracking, which is a very difficult task mainly because of the Doppler-spread in the underwater acoustic channel. Although non-coherent modulation schemes are characterized by high power efficiency, their low bandwidth efficiency makes them unsuitable for high data rate multiuser networks.
Hence, coherent modulation techniques have been developed for long-range, high-throughput systems. In the last years, fully coherent modulation techniques, such as phase shift keying (PSK) and quadrature amplitude modulation (QAM), have become practical due to the availability of powerful digital processing. Channel equalization techniques are exploited to leverage the effect of the inter-symbol interference (ISI), instead of trying to avoid or suppress it. Decision-feedback equalizers (DFEs) track the complex, relatively slowly varying channel response and thus provide high throughput when the channel is slowly varying. Conversely, when the channel varies faster, it is necessary to combine the DFE with a Phase Locked Loop (PLL) , which estimates and compensates for the phase offset in a rapid, stable manner. The use of decision feedback equalization and phase-locked loops is driven by the complexity and time variability of ocean channel impulse responses.
Differential phase shift keying (DPSK) serves as an intermediate solution between incoherent and fully coherent systems in terms of bandwidth efficiency. DPSK encodes information relative to the previous symbol rather than to an arbitrary fixed reference in the signal phase and may be referred to as a partially coherent modulation. While this strategy substantially alleviates carrier phase-tracking requirements, the penalty is an increased error probability over PSK at an equivalent data rate.
Another promising solution for underwater communications is the orthogonal frequency division multiplexing (OFDM) spread spectrum technique, which is particularly efficient when noise is spread over a large portion of the available bandwidth. OFDM is frequently referred to as multicarrier modulation because it transmits signals over multiple sub-carriers simultaneously. In particular, sub-carriers that experience higher SNR, are allotted with a higher number of bits, whereas less bits are allotted to sub-carriers experiencing attenuation, according to the concept of bit loading, which requires channel estimation. Since the symbol duration for each individual carrier increases, OFDM systems perform robustly in severe multi-path environments, and achieve a high spectral efficiency.
Many of the techniques discussed above require underwater channel estimation, which can be achieved by means of probe packets . An accurate estimate of the channel can be obtained with a high probing rate and/or with a large probe packet size, which however result in high overhead, and in the consequent drain of channel capacity and energy.
Data link layer (MAC layer)
The data link layer provides single hop data transmission capability; it will not be able to transmit a packet successfully if the destination node is not directly accessible from the source node. It may include some degree of reliability. It may also provide error detection capability (e.g. CRC check). In case of a shared medium, the data link layer must include the medium access control (MAC) sub-layer.
Frequency division multiple access (FDMA) is not suitable for underwater sensor network due to the narrow bandwidth in underwater acoustic channels and the vulnerability of limited band systems to fading and multipath.
Time division multiple access (TDMA) shows limited bandwidth efficiency because of the long time guards required in the underwater acoustic channel. In fact, long time guards must be designed to account for the large propagation delay and delay variance of the underwater channel in order to minimize packet collisions from adjacent time slots. Moreover, the variable delay makes it very challenging to realize a precise synchronization, with a common timing reference, which is required for TDMA.
Carrier sense multiple access (CSMA) prevents collisions with the ongoing transmission at the transmitter side. To prevent collisions at the receiver side, however, it is necessary to add a guard time between transmissions dimensioned according to the maximum propagation delay in the network. This makes the protocol dramatically inefficient for underwater acoustic sensor network.
The use of contention-based techniques that rely on handshaking mechanisms such as RTS/ CTS in shared medium access is impractical in underwater, for the following reasons:
- large delays in the propagation of RTS/CTS control packets lead to low throughput;
- due to the high propagation delay of underwater acoustic channels, when carrier sense is used, as in 802.11, it is more likely that the channel be sensed idle while a transmission is ongoing, since the signal may not have reached the receiver yet;
- the high variability of delay in handshaking packets makes it impractical to predict the start and finish time of the transmissions of other stations. Thus, collisions are highly likely to occur.
Code division multiple access (CDMA) is quite robust to frequency selective fading caused by underwater multi-paths, since it distinguishes simultaneous signals transmitted by multiple devices by means of pseudo-noise codes that are used for spreading the user signal over the entire available band. CDMA allows reducing the number of packet retransmissions, which results in decreased battery consumption and increased network throughput.
In conclusion, although the high delay spread which characterizes the horizontal link in underwater channels makes it difficult to maintain synchronization among the stations, especially when orthogonal code techniques are used , CDMA is a promising multiple access technique for underwater acoustic networks. This is particularly true in shallow water, where multi-paths and Doppler- spreading plays a key role in the communication performance.
Network layer (Routing)
The network layer is in charge of determining the path between a source (the sensor that samples a physical phenomenon) and a destination node (usually the surface station). In general, while many impairments of the underwater acoustic channel are adequately addressed at the physical and data link layers, some other characteristics, such as the extremely long propagation delays, are better addressed at the network layer.
Basically, there are two methods of routing. The first one is virtual circuit routing and the second one is packet-switch routing.
In virtual circuit routing, the networks use virtual circuits to decide on the path at the beginning of the network operation. Virtual-circuit-switch routing protocols can be a better choice for underwater acoustic networks. The reasons are:
- Underwater acoustic networks are typical asymmetric instead of symmetric. However, packet switched routing protocols are proposed for symmetric network architecture;
- Virtual-circuit-switch routing protocols work robust against link failure, which is critical in underwater environment; and
- Virtual-circuit-switch routing protocols have less signal overhead and low latency, which are needed for underwater acoustic channel environment.
However, virtual-circuit-switch routing protocols usually lack of flexibility.
In packet-switch routing, every node that is part of the transmission makes its own routing decision, i.e., decides its next hop to relay the packet. Packet-switch routing can be further classified into Proactive routing, Reactive and geographical routing protocols. Most routing protocols for ground-based wireless networks are packet-switch based.
Proactive routing protocols attempt to minimize the message latency by maintaining up-to-date routing information at all times from each node to any other node. It broadcasts control packets that contain routing table information. Typical protocols include Destination Sequence Distance Vector (DSDV)  and Temporally Ordered Routing Algorithm (TORA).
However, proactive routing protocols provoke a large signaling overhead to establish routes for the first time and each time the network topology changes. It may not be a good fit in underwater environment due to the high probability of link failure and extremely limited bandwidth there.
Reactive routing protocols only initiate a route discovery process upon request. Correspondently, each node does not need to maintain a sizable "look-up" table for routing. This kind of routing protocols is more suitable for dynamic environment like ad hoc wireless networks. Typical protocol examples are Ad hoc On-demand Distance Vector (AODV) , and Dynamic Source Routing (DSR) .
The shortage of reactive routing protocols is its high latency to establish routing. Similar to its proactive counterpart, flooding of control packets to establish paths is needed, which brings significant signal overhead. The high latency could become much deteriorated in underwater environment because of the much slower propagation speed of acoustic signal compared with the radio wave in the air.
Geographic routing (also called georouting or position-based routing) is a routing principle that relies on geographic position information. It is mainly proposed for wireless networks and based on the idea that the source sends a message to the geographic location of the destination instead of using the network address.
Geographic routing requires that each node can determine its own location and that the source is aware of the location of the destination. With this information a message can be routed to the destination without knowledge of the network topology or a prior route discovery.
A transport layer protocol is needed in underwater sensor network not only to achieve reliable collective transport of event features, but also to perform flow control and congestion control. The primary objective is to save scarce sensor resources and increase the network efficiency. A reliable transport protocol should guarantee that the applications be able to correctly identify event features estimated by the sensor network. Congestion control is needed to prevent the network from being congested by excessive data with respect to the network capacity, while flow control is needed to avoid that network devices with limited memory are overwhelmed by data transmissions.
Most existing TCP implementations are unsuited for the underwater environment, since the flow control functionality is based on a window- based mechanism that relies on an accurate esteem of the round trip time (RTT), which is twice the end-to-end delay from source to destination.
Rate-based transport protocols seem also unsuited for this challenging environment. They still rely on feedback control messages sent back by the destination to dynamically adapt the transmission rate, i.e., to decrease the transmission rate when packet loss is experienced or to increase it otherwise. The high delay and delay variance can thus cause instability in the feedback control.
Furthermore, due to the unreliability of the acoustic channel, it is necessary to distinguish between packet losses due to the high bit error rate of the acoustic channel, from those caused by packets being dropped from the queues of sensor nodes due to network congestion. In terrestrial, assume that congestion is the only cause for packet loss and the solution lies on decreasing the transmission rate, but in underwater sensor network if the packet loss is due to bad channel then the transmission rate should not be decreased to preserve throughput efficiency.
Transport layer functionalities can be tightly integrated with data link layer functionalities in a cross-layer module. The purpose of such an integrated module is to make the information about the condition of the variable underwater channel available also at the transport layer. In fact, usually the state of the channel is known only at the physical and channel access sub-layers, while the design principle of layer separation makes this information transparent to the higher layers. This integration allows maximizing the efficiency of the transport functionalities and the behavior of data link and transport layer protocols can be dynamically adapted to the variability of the underwater environment.
Although many application areas for underwater sensor networks can be outlined, to the best of our knowledge the definition of an application layer protocol for underwater sensor network remains largely unexplored.
The purpose of an application layer is multifold:
- to provide a network management protocol that makes hardware and software details of the lower layers transparent to management applications;
- to provide a language for querying the sensor network as a whole;
- to assign tasks and to advertise events and data.
No efforts in these areas have been made to date that address the specific needs of the underwater acoustic environment. A deeper understanding of the application areas and of the communication problems in underwater sensor networks is crucial to outline some design principles on how to extend or reshape existing application layer protocols for terrestrial sensor networks.
Some of the latest developments in middleware may be studied and adapted to realize a versatile application layer for underwater sensor networks. For example, the San Diego Supercomputing Center
Storage Resource Broker (SRB) is a client- server middleware that provides a uniform interface for connecting to heterogeneous data resources over a network, and accessing replicated data sets. SRB provides a way to access data sets and resources based on their attributes and/or logical names rather than their names or physical locations.
CHAPTER 6: Routing Protocol
Routing Protocol for wireless sensor nodes
Fully connected peer-to-peer topologies without the need for routing were commonly used earlier. But such networks could suffer from near-far power problems and multi-hop routing is now preferred for large networks. In clustered network routing topology, only some nodes have routing (gateways) functionality and ordinary nodes within one-hop distance to gateways send data to them for routing. In fully distributed routing topology, all nodes are equal and perform routing to neighbors as required.
In last few years, many energy efficient routing protocols have been proposed for terrestrial sensor networks, such as Directed Diffusion , Two-Tier Data Dissemination, GRAdient, Rumor routing , and SPIN . In the following, we brief these protocols and discuss why they are not suitable for the underwater sensor network environments.
- Directed Diffusion is proposed in . In the target application scenario, the sink floods its interest into the network, and the source node responds with data. The data are first forwarded to the sink along all possible paths.
Then, an optimal path is enforced from the sink to the source recursively based on the quality of the received data. Directed Diffusion works well in low dynamic networks, where most nodes are stationary and forwarding paths are relatively stable. However, if we apply Directed Diffusion in UWSNs, it will consume a large amount of resource (including energy) to maintain the forwarding paths due to node mobility.
In the Rumor routing algorithm , both event notifications and data queries are forwarded randomly. The successful data delivery depends on the chance that these two types of forwarding paths interleave. In stationary networks, it is most likely that these paths will meet since they are relatively stable. However, in underwater sensor networks, this is unlikely to happen. Even in networks with low mobility, the instability of forwarding paths is worsen by the low propagation speed of acoustic signals.
- SPIN (Sensor Protocol Information via Negotiation) is proposed for low data rate networks . When a node wants to send data, it broadcasts a description message of the data, and each neighbor decides whether to accept the data based on its local resource condition. Once again, the high propagation delay in UWSNs makes this protocol's throughput low. Moreover, flooding in SPIN depletes the energy of the network, especially for medium or high data rate networks.
- TTDD (Two-Tier Data Dissemination) addresses the problem caused by mobile sinks . In this protocol, the source sensor initiates the process to construct a grid covering the whole field. Data and queries are forwarded along the cross points in the grid. The impact of the sink mobility is confined within each cell. When most nodes in the network are fixed, it costs less energy to maintain the grid. However, the overhead to maintain the grid will increase significantly when most nodes are mobile (for example, in UWSNs).
- GRAdient broadcast is a robust data delivery protocol proposed in . In GRAdient, a cost field is built in the whole network by the sink, which has the lowest cost. Data packets are forwarded along the direction from higher cost nodes to lower cost nodes. The width of the path is controlled by the credits of each packet. In this way,
GRAdient improves robustness. However, in UWSNs where sensor nodes are mobile, the protocol will consume a large amount of scarce energy to update the cost field in order to keep relatively accurate paths from the source to the sink.
In short, the routing protocols for UWSNs have to address the node mobility issue at minimum energy expenditure. However, existing routing protocols designed for terrestrial sensor networks can not satisfy this requirement. When applied directly in the underwater sensor network environments, these proposals become very expensive in terms of energy due to node mobility.
While going through the literature review in section 5.3 network layer, among the three types of routing, geographical routing protocol is more suitable for underwater routing protocol. Thus in this paper, a Vector Based Forwarding underwater routing protocol has been proposed to be efficient and robust for routing underwater wireless sensor nodes.
Overview of Vector Based Forwarding (VBF)
In some recent papers , a new underwater routing protocol has been proposed by Peng Xie, Jun-Hong Cui, Li Lao, called Vector Based Forwarding VBF. The VBF routing protocol is essentially a location based routing approach. VBF main aim is to provide a robust, scalable and energy efficient routing. In VBF, each packet carries the positions of the sender, the target and the forwarder (i.e., the node which forwards this packet). The forwarding path is specified by the routing vector from the sender to the target. Upon receiving a packet, a node computes its relative position to the forwarder by measuring its distance to the forwarder and the angle of arrival (AOA) of the signal. Recursively, all the nodes receiving the packet compute their positions. If a node determines that it is close to the routing vector enough (e.g., less than a predefined distance threshold), it puts its own computed position in the packet and continues forwarding the packet; otherwise, it simply discards the packet. Therefore, the forwarding path is virtually a routing "pipe" from the source to the target: the sensor nodes inside this pipe are eligible for packet forwarding, and those outside the pipe do not forward.
OP - Original Point
TP - Target Point
FP - Forwarder Point
RANGE - It is used to broadcast to a certain range specify by the TP
RADIUS - It is the pre-defined threshold distance from the routing vector.
How VBF works
The above figure shows an underwater wireless sensor network which consist of multiple sensor nodes in 3-diamension. All sensor nodes are equipped with some devices which enable them to measure the distance and the angle of arrival (AOA) of the signal. Node S0 is the source and node S1 is the sink. When S0 want to sent data packets to sink S1, it first establishes a routing vector (S0S1) as shown in the above figure. W is the threshold distance from the routing vector, which makes a cylinder pipe with central axis S1S0 and radius W. S0 broadcast the packet with target S1. Upon receiving the packet the node calculate its distance from the routing vector. If it found that it is in the range then it forward the packet to the next node. But if it is not in the range on the cylinder pipe, it simply discards the packet.
Source routing allows a sender of a packet to partially or completely specify the route the packet takes through the network. In contrast, in non-source routing protocols, routers in the network determine the path based on the packet's destination. VBF is a source routing protocol but like all other source routing protocol, VBF required no state information is required at each node. Hence it is scalable to the size of the network. Moreover only the nodes found in the cylinder pipe participate in packet forwarding, thus the network energy is saved.
Self adaptation algorithms
In VBF protocol, all nodes which are close to the routing vector (within the routing pipe) are participated in the routing path of data packet. However when sensor nodes are densely deployed VBF may involve too many nodes in data forwarding, therefore the energy consumption is increased. A forwarding policy cannot be applied based on the node density as the network is dynamically changing due to mobility. The mobility is causes by the underwater vehicles, thus a Self-Adaptation Algorithm  for VBF proposed by Peng Xie has been used to allow each node to estimate in its neighborhood (based on local information) and forward the packets adaptively. The self - adaptation algorithms make use of desirableness factor.
The desirableness factor ?, is the measurement if a node is suitable or not to forward a packet and is given by:
S0 - source node
S1 - sink node
F - Forwarder node
p - Projection of node A to the routing vector S0S1
d - Distance between node A and node F
? - Angle between vector FS1 and vector FA
R - Transmission range
W - Radius of the routing pipe (the threshold distance)
From the above figure node F is the forwarder and it must choose between node A, B or C to be the next forwarder. Once node A, B and C receive the packet they compute the desirableness factor. If the node has a large desirableness factor then either its projection to the routing vector is large or it is not far away from the forwarder. Hence this node will not be part of the routing path. Otherwise if the node (A) has a zero desirableness factor, then this node is on both the routing vector and the edge of the transmission range of the forwarder F. This node is called optimal node and its position is known as best position.
By using the self adaptation algorithm, the most desirable nodes are selected as forwarders. In this algorithm when a node receive a packet and if the node is close enough to the routing vector, it holds the packet for a time period based on the desirableness factor. The time period is given by:
Tadaptation - pre-defined maximum delay
?o - propagation speed of acoustic signals in water
d - distance between this node and the forwarder.
From the above equation if follows, the smaller the desirableness factor, the less time to wait. During the delayed time period Tadaptation, if a node receives duplicate packets from n other nodes, then this node has to compute its desirableness factors relative to these nodes, ?1,..., ?n, and the original forwarder, ?0. If min (?0, ?1,..., ?n) < ?c/2^(n), where ?c is a pre-defined initial value of the desirableness factor (0 ? ?c ? 3), then this node forwards the packet; otherwise it discards the packet.
The self-adaptation algorithm gives higher priority to the desirable node to continue broadcasting the packet, and it also allows a less desirable node to have chances to re-evaluate its priority in the neighborhood. After receiving the same packets from its neighbors, the less desirable node can measure its importance by computing its desirableness factor relative to its neighbors. If there are many more desirable nodes in the neighborhood, the probability of this node to forward the packet is reduced. That is, it is useless for this node to forward the packet anymore since many other more desirable nodes have forwarded the packet. In fact, if a node receives more than two duplicate packets during its waiting time, it is most likely that this node will not forward the packet no matter what initial value ?c takes. Hence the computation overhead is reduced by skipping the re-evaluation of the desirableness factor.
The lower Bound for Tdelay
The purpose of the delay is to distinguish the importance of the nodes in the transmission range of a forwarder. When maximum delay, Tdelay is set small, the end- to-end delay is reduced. In VBF Tdelay is set large enough due to the purpose of the delay time.
If at any given time, two adjacent nodes are approaching zero desirableness factor, then these nodes are close to the best position of the forwarder. The self-adaptation algorithm attempts to differentiate the delay time of these nodes to the extent such that the difference between their delay time period is large enough to allow the optimal node to suppress the other node.
Types of queries in VBF
Routing in VBF is initiated by query packets. There are two types of queries which can be routed effectively by VBF. They are:
- Location dependent queries
In this case, the sink is interested in some specific area and knows the location of the area. So it issues an INTEREST query packet. It broadcast the query packet to the specify source to get the required information. Only those nodes which are from a distance less than RADIUS (i.e. the threshold distance) from the routing vector, participate in the routing path.
- Location Independent queries
In this case, the sink is interested in only some data information regardless to the location. It broadcast a query packet. Then the sink receives back an INTEREST packet which contains the path and coordinate of the source node. Therefore data are being sent from the source to the sink. An example of this can be when a sink want to know where an abnormal event has happen in the network.
Handling Source Mobility in VBF
Mobility plays an important role in underwater sensor network. The sink sent an INTEREST packet to a source node for having data in that particular area. The source node sent back data packets through routing path. If at any instant the source node move out from that interested area, the data received by the sink could be inaccurate. Therefore a sink- assisted approach is used to solve this problem.
The source keeps sending packets to the sink and the sink can make use of the source location information carried in the packets to find out if the source is moving out from the interest area. If so the sink sends the SOURCE_DENY packet to stop that source from sending data. At the same time the sink initiates another interest query to find a new source.
CHAPTER 7: Ocean Model Characteristic
Underwater speed propagation
The ocean acoustic environment
The ocean is an acoustic waveguide limited above by the sea surface and below by the seafloor. The speed of sound in the waveguide plays the same role as the index of refraction does in optics. Sound speed is normally related to density and compressibility. In the ocean, density is related to static pressure, salinity, and temperature. The sound speed in the ocean is an increasing function of temperature, salinity, and pressure, the latter being a function of depth. It is customary to express sound speed (c) as an empirical function of three independent variables: temperature (T) in degrees centigrade, salinity (S) in parts per thousand, and depth (z) in meters. A simplified expression for this dependence is
c = 1449.05 + 45.7t - 5.21t2 + 0.23t3 + (1.333 - 0.126t + 0.009t2) (S - 35) + (16.23 + 0.253t) z + (0.213-0.1t) z2 + [0.016 + 0.0002(S-35)] (S - 35)tz .......equ(6)
The above equation is valid for 0ºC ? T ? 35ºC, 0 ? S ? 45%, and 0 ?z ? 4000m and t = T/10.
The above equation for the speed of sound in sea-water as a function of temperature, salinity and depth is given by Coppens equation (1981).
Seasonal and diurnal changes affect the oceanographic parameters in the upper ocean. In addition, all of these parameters are a function of geography. The above figure shows a typical set of sound-speed profiles indicating greatest variability near the surface as function of season and time of day. In a warmer season (or warmer part of the day) the temperature increases near the surface and hence the speed of sound increases toward the sea surface.
In non polar regions, the oceanographic properties of the water near the surface result from mixing due to wind and wave activity at air-sea interface. This near-surface mixed layer has a constant temperature (except in calm, warm surface conditions). Hence, in this isothermal mixed layer we have a sound-speed profile which increases with depth because of the pressure gradient effect, the last term in the above equation of sound speed in water. This is the surface duct region, and its existence depends on the near-surface oceanographic conditions. Note that the more agitated the upper layer is, the deeper the mixed layer and the less likely will there be any departure from the mixed layer part of the profile depicted in the above figure. Hence, an atmospheric storm passing over a region mixes the near-surface waters so that a surface duct is created or an existing one deepened or enhanced.
Below the mixed layer is the thermocline where the temperature decreases with depth and therefore the sound speed also decreases with depth. Below the thermocline, the temperature is constant (about 2ºC - a thermodynamic property of salt water at high pressure. Therefore, between the deep isothermal region and the mixed layer, a minimum sound speed is required which is often referred to as the axis of the deep sound channel. However in polar regions, the water is coldest near the surface and hence the minimum sound speed is at the ocean-air (or ice) interface as in the above figure. In continental shelf regions (shallow water) with water depth of the order of a few hundred meters, only the upper part of the sound-speed profile in the above figure is relevant. This upper region is dependent on season and time of day, which, in turn, affects sound propagation in water column.
Underwater acoustic propagation depends on many factors. The direction of sound propagation is determined by the sound speed gradients in the water. In the sea the vertical gradients are generally much larger than the horizontal ones. These facts, combined with a tendency for increasing sound speed with increasing depth due to the increasing pressure in the deep sea reverses the sound speed gradient in the thermocline creating an efficient waveguide at the depth corresponding to the minimum sound speed. The sound speed profile may cause regions of low sound intensity called "Shadow Zones" and regions of high intensity called "Caustics". 
Top and bottom boundaries
Along with the sound speed profile, information about the top and bottom boundary conditions are important descriptions of an ocean region. The top boundary may be described as a vacuum, in which no sound travels, or as an acoustic half-space that is described by its own sound speed profile [6, 7]. This description will depend on the deployment location and available knowledge about propagation past the ocean boundary.
Depending on the material consistency of the seabed, the bottom boundary can be treated as perfectly rigid or again as having its own sound propagation profile. The roughness and bathymetry profile of the bottom may also be considered. This feature of the model is especially useful for bottom dwelling networks in which bottom reflections will be prevalent [14, 15].
In general, as sound propagates underwater there is a reduction in the sound intensity over increasing ranges, though in some circumstances a gain can be obtained due to focusing. Propagation loss (sometimes referred to as transmission loss) is a quantitative measure of the reduction in sound intensity between two points, normally the sound source and a distant receiver. If Is is the far field intensity of the source referred to a point 1 m from its acoustic centre and Ir is the intensity at the receiver, then the propagation loss is given by PL = 10log(Is / Ir). At short range the propagation loss is dominated by spreading while at long range it is dominated by absorption and/or scattering losses. 
Signal to noise ratio
SNR is usually expressed as Eb / No or energy per bit over noise spectral density , .
The received SNR depends on a few basic factors: the transmitter power, the data rate being sent, the noise level at the receiver, and the signal attenuation between the transmitter and receiver.
Transmit Power: There is no fundamental limit to transmitter power, but it can have a major effect on the energy budget for the system. For energy efficiency and to minimize interference with neighboring transmitters we wish to use the smallest possible transmitter power.
Data Rate: This is a tradeoff between available power and channel bandwidth. Because acoustic communications are possible only over fairly limited bandwidths, a fairly low data rate is expected by comparison to most radios. A range of data rate of 5kb/s to 20 kb/s is used. In application such as robotic control, the ability to communicate at all (even at a low rate) is much more important than the ability to send large amounts of data quickly.
Noise Level: Noise levels in the ocean have a critical effect on sonar performance, and have been studied extensively. Burdic  and Urick  are two standard references. The frequency range interested is between 200 Hz and 50 kHz (the midfrequency band). In this frequency range the dominant noise source is wind acting on the sea surface. Knudsen  has shown a correlation between ambient noise and wind force or sea state. Ambient noise increases about 5dB as the wind strength doubles. Peak wind noise occurs around 500 Hz, and then decreases about -6dB per octave. At a frequency of 10,000 Hz the ambient noise spectral density is expected to range between 28 dB/Hz and 50 dB/Hz relative to 1 micro Pascal. This suggests the need for wide range control of transmitter power.
Noise in the underwater channel is a combination of various sources. These sources include man-made sources, such as ship traffic, and environmental factors, such as surface action from wind, thermal noise, and turbulent activity in the water [5-7].
For a target environment, each of the noise sources will need to be described so as to capture the specific noise characteristics. In this paper, noise calculations follow the work of Harris and Zorzi , in which the frequency dependant noise calculation includes factors forwater turbulence (Nt ), wind (Nw), ship activity (Ns), and temperature (Nth).
Noise calculation is frequency dependant and allows for inclusion of a shipping factor s and wind speed w. The shipping factor is a value between 0 and 1, representing a scale of minimal activity (0) to high activity (1).Wind speed is simply given as m/s and may represent an average value, or may vary throughout testing to match actual wind pattern data that were collected.
The following relationships show the calculation of noise factors 
10 log Nt (f ) = 17 - 30 log f
10 log Ns (f ) = 40 + 20(s - 0.5) + 26 log f -60 log(f + 0.03)
10 log Nw (f ) = 50 + 7.5w½ + 20 log f -40 log(f + 0.4)
10 log Nth (f ) = -15 + 20 log f
Where Nt (f ) is the noise due to turbulence, Ns(f) is shipping noise, s is the shipping factor, Nw (f ) is noise due to wind, w is the wind speed in m/s, and Nth(f) is noise due to thermal effects.
The total noise is then calculated as:
Noise (f ) = Nt (f ) + Nw (f ) + Ns (f ) + Nth (f ) .......equ(7)
Where Noise (f) is the total noise.
Signal Attenuation: Attenuation is due to a variety of factors. Both radio waves and acoustic waves experience 1=R2 attenuation due to spherical spreading. There are also absorptive losses caused by the transmission media. Unlike in-the-air RF, absorptive losses in underwater acoustics are significant, and very dependent on frequency. At 12.5kHz absorption it is 1dB/km or less. At 70kHz it can exceed 20dB/km. This places a practical upper limit on our carrier frequency at about 100kHz. There are additional loss effects, mostly associated with scattering, refraction and reflections .
An underwater acoustic channel is characterized by a path loss that depends on both distance d and signal frequency f as
A (d,f) = dk × [10(?(f)/10) ] d ....equ(8)
Where k is the spreading factor and ?(f) is the absorption coefficient.
The spreading factor describes the geometry of propagation, k = 2 correspond to spherical spreading, k = 1 to cylindrical spreading, and k = 1.5 to practical spreading. 
The frequency dependent absorption coefficient ? is calculated by Thorp's expression in  for frequencies above a few hundred Hertz as:
Where ? is in dB/km and f is the frequency in kHz.
For lower frequencies, ? is expressed as:
Where ? is in dB/km and f is the frequency in kHz.
The transmission loss is mainly caused by distance dependent attenuation and frequency dependent absorption both in shallow water and deep sea.
Characteristic of Shallow water
Two definitions of shallow water may be made: hypsometric and acoustic. The hypsometric definition is based on the fact that most continents have continental shelves bordered by the 200m contour line, beyond which the bottom generally falls off rapidly into deep water. Therefore, shallow water is taken to mean continental shelf waters shallower than 200m; shallow water represents about 7.5% of the total ocean area.
Acoustically, shallow water conditions exist whenever the propagation is characterized by numerous encounters with both the sea surface and the sea floor. By this definition, some hypsometrically shallow water areas are acoustically deep; alternatively, the deep ocean may be considered shallow when low-frequency, long-range propagation conditions are achieved through repeated interactions with both the surface and the bottom. Shallow water regions are distinguished from deep-water regions by the relatively greater role played in shallow water by the reflecting and scattering boundaries. Also, differences from one shallow-water region to another are primarily driven by differences in the structure and composition of the sea floor. The sea floor is perhaps the most important part of the marine environment that distinguishes shallow water propagation from deep water propagation.
Ambient noise in shallow water tends to be highly variable, both in time and location. As in deep water, wind speed seems to be the major factor in determining the levels of ambient noise above a frequency of about 100 Hz. The noise spectral levels from several studies of ambient noise in shallow water are shown in figure 12. At very low frequency below 500 Hz, the noise is due to shipping activities. Turbulence noise occurs at frequency below 1 Hz. Above 500 Hz, levels are often 5-10 dB higher in shallow water than in deep water. [44,52]
In shallow water, acoustic signals propagate within a cylinder bounded by the sea surface and sea floor .
Where I is the intensity in W/m2, R is the range in meter and D is the distance in meter from the surface to the bottom of the sea.
Characteristic of Deep water
The principal characteristic of deep-water propagation is the existence of an upward-refracting sound-speed profile which permits long-range propagation without significant bottom interaction. Hence using reciprocates of shallow water, the deep water is considered at the depth of 200 m and above.
In deep water, frequencies from 1 to 20 Hz are affected by noise of ship, wind speed and oceanic turbulence is also an important source of noise at these frequencies. This turbulence takes the form of "irregular random water currents of large or small scale". Below 1 Hz ambient noise arises from seismic source, turbulence, tides and waves. From 20-300Hz, noise levels from distant shipping usually exceed wind-related noise. Above 300Hz shipping sounds may or may not be significant depending on the level of wind-dependent ambient noise. From 500Hz to 50 kHz, wind wave and (intermittently) precipitation noise dominate. Above 30-50 kHz, thermal agitation may dominate, especially when wind speed is low [44, 52]. These are shown in figure 14 below.
In deep water the propagation range is not bounded by the sea floor and the surface, so that spherical spreading applies. In deep sea, transmission loss is caused by spherical spreading and absorption.
The transmission power is directly proportional to the square of the distance between sensor nodes in deep water scenarios.
Where I is the intensity in W/m2 and R is the range in meter
CHAPTER 8: Implementation of VBF Routing Protocol using ns-2
The simulations were built using the underwater package Aqua-sim in ns-2.30 .Developed on the basis of NS-2, Aqua-Sim can effectively simulate acoustic signal attenuation and packet collisions in underwater sensor networks. Moreover, Aqua-Sim supports three-dimensional deployment. Further, Aqua-Sim can easily be integrated with the existing codes in NS-2. Aqua-Sim is in parallel with the CMU wireless simulation package (describes the wireless model that was originally ported as CMU's Monarch group's mobility extension to ns). As shown in the figure below, Aqua-Sim is independent of the wireless simulation package and is not affected by any change in the wireless package. On the other hand, any change to Aqua-Sim is also confined to itself and does not have any impact on other packages in NS-2. In this way, Aqua-Sim can evolve independently. 
Aqua-Sim follows the object-oriented design style of NS- 2, and all network entities are implemented as classes in C++. Currently, Aqua-Sim is organized into four folders, uw-common, uw-mac, uw-routing and uw-tcl. The codes simulating underwater sensor nodes and traffic are grouped in folder uw-common; the codes simulating acoustic channels and MAC protocols are organized in the folder of uw-mac. The folder uw-routing contains all routing protocols. The folder uw-tcl includes all Otcl script examples to validate Aqua-Sim.
An Otcl script is being written which include all basic information about the ocean environment and the underwater channel. It makes use of the above modules found in uw-common, uw-mac and uw-routing. Some modification has been made in these underwater modules in order to vary the underwater propagation speed and some formulas have been added to calculate power received.
The performance metrics which have been proposed are:
- Propagation delay
Propagation delay is the total time delay to send a number of packets from a source to a destination through Vector Based Forwarding routing protocol. Unit is second.
Where distance travel is in meter and the propagation speed is in m/s.
- Signal to Noise Ratio (SNR)
SNR is the ratio of the total power transmitted and the total noise in the network to send a number of packets from a source to a destination through Vector Based Forwarding.
Where the signal strength is in dB re µ Pa.
In all the simulations describe in this section, sensor nodes are randomly deployed 3D network architecture. There are one
Cite This Dissertation
To export a reference to this article please select a referencing stye below: