Data Collection, Storage, and Retrievalwith an Underwater Sensor NetworkI. Vasilescu, K. Kotay, and D. RusMIT CSAILThe Stata Center, 32 Vassar StreetCambridge, MA 02139fiuliuv,kotay,[email protected]. Dunbabin and P. CorkeCSIRO ICT Centre1 Technology CtBrisbane, Australiafmatthew.dunbabin,[email protected] this paper we present a novel platform for underwater sen-sor networks to be used for long-term monitoring of coralreefs and fisheries. The sensor network consists of staticand mobile underwater sensor nodes. The nodes commu-nicate point-to-point using a novel high-speed optical com-munication system integrated into the TinyOS stack, andthey broadcast using an acoustic protocol integrated in theTinyOS stack. The nodes have a variety of sensing capa-bilities, including cameras, water temperature, and pres-sure. The mobile nodes can locate and hover above thestatic nodes for data muling, and they can perform networkmaintenance functions such as deployment, relocation, andrecovery. In this paper we describe the hardware and soft-ware architecture of this underwater sensor network. Wethen describe the optical and acoustic networking protocolsand present experimental networking and data collected ina pool, in rivers, and in the ocean. Finally, we describe ourexperiments with mobility for data muling in this network.Categories and Subject DescriptorsC.2.4 [Computer-Communications Networks]: Distrib-uted Systems; C.3 [Special-Purpose and Application-Based Systems]: Real-time and embedded systemsGeneral TermsAlgorithms, Design, ExperimentationKeywordsmobile sensor networks, underwater networks, data muling1. INTRODUCTIONThe application of wireless sensor networks to the under-water domain has huge potential for monitoring the health ofriver and marine environments. The oceans alone cover 70%Permission to make digital or hard copies of all or part of this work forpersonal or classroom use is granted without fee provided that copies arenot made or distributed for profit or commercial advantage and that copiesbear this notice and the full citation on the first page. To copy otherwise, torepublish, to post on servers or to redistribute to lists, requires prior specificpermission and/or a fee.SenSys’05, November 2–4, 2005, San Diego, California, USA.Copyright 2005 ACM 1-59593-054-X/05/0011 ...$5.00.of our planet and along with rivers and lakes are critical toour well-being. Monitoring these environments is difficultand costly for humans: divers are regulated in the hoursand depths at which they can work, and require a boat onthe surface that is costly to operate and subject to weatherconditions. A sensor network deployed underwater couldmonitor physical variables such as water temperature andpressure as well as variables such as conductivity, turbid-ity and certain pollutants. The network could track plumesof silt due to dredging operations or pollutants flowing infrom land, and it could monitor and model the behavior ofunderwater ecosystems. Imaging sensors could be used tomeasure visible change in the environment or count, andperhaps even classifyspecies.However a number of problems confront us in achievingthis goal. Some such as power efficiency, deployment andrepair are common to wireless sensor network deploymentson land, though more difficult in the underwater environ-ment. Other issues render the problem radically different.A key issue is communications — current terrestrial wire-less sensor network applications to date have used radio. Atfrequencies that are practical with low-cost radio chips andcompact antennas, radio waves are attenuated so stronglyin salt water that radio communications is impractical.In this paper we describe an underwater sensor networksystem that consists of static and mobile sensor nodes (seeFigure 1). The system is networked with two communicationmodalities: ultrasonic and optical. Ultrasonic communica-tions has a long history for underwater applications and iswidely used with autonomous underwater vehicles. It hasmany similarities to radio: it is a shared medium that sup-ports broadcasting, but the low propagation speed1posesa challenge for carrier-sense transmission strategies. Opti-cal communication is capable of much higher data trans-mission rates and the propagation speed is much closer tothe speed of light2. Unlike ultrasonic and radio commu-nications, optical communication is essentially directional.This dual networking scheme enables many underwater sen-sor network applications, as it supports low-speed broadcast(e.g. necessary for localization) and high-speed directionaldata transfer (e.g. for monitoring).An important characteristic of the underwater sensor net-work we built is mobility, whose benefits have been exploredpreviously in [6, 15–17]. The sensor network includes both1Due to relatively low speed of sound in water of approxi-mately 1500m/s.2v = c/n where n is the refractive index of the medium.154static and mobile nodes. Mobility enhances the performanceof this sensor network in several ways. First, it provides ameans for deploying, reconfiguring, and retrieving the nodesin the network. Second, it permits large area coverage withsparse networks which is especially important in an under-water environment — a much harder space to access thanterrestrial space. The mobile nodes can move across the fieldto ensure the necessary connectivity. Third, mobile nodescan act as data mules and travel from node to node across asparsely deployed sensor network to collect data. Commu-nications is enabled only when the sensors and the mobilemules are in close proximity. Transmitting data over theseshorter distances reduces the power consumption on eachsensor and alleviates the hot spot problem on the sensorsnear the destination. Moreover, since underwater acousticcommunication is characterized by low data rates, and op-tical underwater communication is subject to short ranges,mobility enables a time-efficient and more energy-efficientmeans to collect and transmit the data.Specifically, we use the mobile nodes as data mules. Ourdata muling solution relies on a mobile node which is anautonomous underwater vehicle (AUV) with a matching op-tical communications link. This node can locate the staticnodes using an optical location system. The AUV visitsthe nodes periodically to upload the data. The nodes aremostly in a deep sleep mode and wake every few seconds todetermine if they are being optically signalled. This createsa desirable