sensorsArticleA Novel Cooperative Opportunistic Routing Schemefor Underwater Sensor NetworksSeyed Mohammad Ghoreyshi *, Alireza Shahrabi and Tuleen BoutalebSchool of Engineering and Built Environment, Glasgow Caledonian University, Lanarkshire, Glasgow G4 0BA,UK; A.Shahrabigcu.ac.uk (A.S.); T.Boutalebgcu.ac.uk (T.B.)* Correspondence: Seyed.MohammadGhoreyshigcu.ac.uk; Tel.: +44-141-331-8547Academic Editor: Jaime Lloret MauriReceived: 15 December 2015; Accepted: 22 February 2016; Published: 26 February 2016Abstract:Increasing attention has recently been devoted to underwater sensor networks (UWSNs)because of their capabilities in the ocean monitoring and resource discovery. UWSNs are faced withdifferent challenges, the most notable of which is perhaps how to efficiently deliver packets takinginto account all of the constraints of the available acoustic communication channel. The opportunisticrouting provides a reliable solution with the aid of intermediate nodes collaboration to relay a packettoward the destination. In this paper, we propose a new routing protocol, called opportunistic voidavoidance routing (OVAR), to address the void problem and also the energy-reliability trade-off in theforwarding set selection. OVAR takes advantage of distributed beaconing, constructs the adjacencygraph at each hop and selects a forwarding set that holds the best trade-off between reliability andenergy efficiency. The unique features of OVAR in selecting the candidate nodes in the vicinityof each other leads to the resolution of the hidden node problem. OVAR is also able to select theforwarding set in any direction from the sender, which increases its flexibility to bypass any kind ofvoid area with the minimum deviation from the optimal path. The results of our extensive simulationstudy show that OVAR outperforms other protocols in terms of the packet delivery ratio, energyconsumption, end-to-end delay, hop count and traversed distance.Keywords: routing protocol; void area; underwater sensor network1. IntroductionNowadays, resource discovery in the underwater environment has become one of the importantgoals to reduce dependency on land resources. However, it is a difficult and costly task to monitor anddiscover the underwater environment. Underwater sensor networks (UWSNs) have recently attractedmuch attention due to their significant ability in ocean monitoring and resource discovery. Due torestrictions on the use of radio waves, acoustic transmission is most commonly used in the underwaterenvironment. Required data are collected by the underwater sensors and directed towards the sinkon the surface. Afterwards, the sink can transmit collected information to the monitoring centre viasatellite for further analysis 14, as shown in Figure 1.Some unique features of UWSNs make data forwarding in this environment a challenging task.This includes node movement, low available bandwidth, slow propagation speed, high deploymentcost and a lossy environment 57. It also should be mentioned that the Global Positioning System(GPS) cannot be used in an underwater environment as a localisation system because of the quickattenuation of its waves in water 5. Furthermore, nodes cannot be aware of their positions bypre-configuration, because they are not stationary due to the water current. Nevertheless, the depthof each node in the water can be estimated through an embedded pressure gauge 8. Then, depthinformation can be used during the data forwarding procedure.Sensors 2016, 16, 297; doi:10.3390/ 2016, 16, 2972 of 29The presence of void areas, a high bit error rate and energy conservation are perhaps the mostchallenging issues from the perspective of routing protocols in UWSNs. A void communication area isa three-dimensional region between underwater nodes that lacks any nodes inside (similar to holes).The void area can prevent communication between some of the nodes in thenetwork 9.There arevarious reasons for the presence of void areas, such as sparse topology, temporary obstacles, unreliablenodes or links, etc. 10. In most cases, the lack of employing enough sensor nodes, due to their highcost, while covering a large monitoring area might lead to sparse deployment of the sensors and,consequently, the creation of some void area. Moreover, the relocation of underwater sensor nodes bythe water current can potentially create a void area 11,12.On the other hand, the adverse characteristics of the underwater channel can cause a high biterror, resulting from high attenuation, channel fading, noise, Doppler spread, etc. The communicationchannel quality varies at different ocean depths under varying pressure, temperature and salinity. Thelimited bandwidth of acoustic transmission also reduces the efficiency of communication betweenunderwater nodes 13,14. Generally, nodes are considered connected to each other if the transferredsignal between them can be decoded without any error.In terms of energy consumption, there are also some restrictions due