07-05-2011, 03:47 PM
Abstract
A SEA Swarm (Sensor Equipped Aquatic Swarm) isa collection of mobile underwater sensors that moves as a groupwith water current and enables 4D (space and time) monitoringof local underwater events such as contaminants and intruders.For prompt alert reporting, mobile sensors routes events to mobilesinks (i.e., autonomous underwater vehicles) via geographicrouting that is known to be most efficient under mobility andscarce acoustic bandwidth. In order for a packet to be routed tothe destination using geographical routing, it requires to knowthe location of the destination. This is accomplished by havinga location service that returns the location of a requested node.Our goal is to design such location service for SEA Swarm.In this paper, we analyze various design choices to realize anefficient location service in SEA Swarm scenarios. We find thatconventional ad hoc network location service protocols cannotbe directly used, because the entire swarm moves along watercurrent. We prove that maintaining location information in a2D plane is a better design choice. Given this, we propose abio-inspired location service called a Phero-Trail location serviceprotocol. In Phero-Trail, location information is stored in a 2Dupper hull of a SEA Swarm, and a mobile sink uses its trajectory(`a la a pheromone trail of ants) projected to the 2D hull tomaintain location information. This enables mobile sensors toefficiently locate a mobile sink. Our results show that Phero-Trail performs better than existing approaches.Index Terms—Underwater sensor networks, geographic locationservices, mobile networks, pheromone trail.
I. INTRODUCTION
ALARGE-SCALE Underwater Sensor Network (UWSN)architecture has recently been proposed to explore theocean and in particular, to support solutions for time-criticalaquatic applications such as submarine tracking and harbormonitoring [31]. Unlike traditional aquatic monitoring orsurveillance applications where sensors are usually tetheredto the sea floor or attached to pillars or surface buoys,we assume that a large number of underwater sensor nodesare air-dropped to the venue of interest to create a SEASwarm (Sensor Equipped Aquatic Swarm) [32]. Each node is equipped with a low bandwidth acoustic modem and withvarious sensors. It can dynamically control its depth through afish-like bladder apparatus and a pressure gauge, e.g., Droguesfrom UCSD [22].1A SEA Swarm operates and moves as a group (swarm)with water current. Each sensor monitors local underwateractivities and reports critical data or events in real-time usingmulti-hop routing to a distant data collection center, e.g.,surface buoys or Autonomous Underwater Vehicles (AUVs).We assume that both surface buoys and AUVs are morecapable than regular sensor nodes in terms of energy, storage,communications and locomotion. They are equipped with GPSand thus can also be used for localization of less capablemobile sensor nodes such as in Dive’N’Rise (DNR) [12],[13]. They are also equipped with wireless communicationdevices for over-the-surface data reporting (e.g., WiFi andSatellite communications). For tactical missions, AUVs maybe equipped with special devices (e.g., weapons to attackenemy submarines).There are several major advantages of SEA swarm architecture.First, mobile sensors provide 4D (space and time)monitoring, thus forming dynamic monitoring coverage. Second,the multitude of sensors helps provide extra control onredundancy and granularity. Third, floating sensors can helpincrease system reusability because we can control the depthof a sensor node. When the battery is low or the mission isover, the sensors resurface and can be recharged and reused.Since high-frequency “radio” signals are quickly absorbedby water, underwater networking must rely on an underwateracoustic channel that has low bandwidth [39] and largepropagation latency. The speed of sound in underwater is fiveorders of magnitude lower than the speed of light (approximately1500m/s vs. 3x108m/s). An acoustic data transmissionconsumes much more energy than terrestrial microwave datacommunications. Moreover, such drastic reduction in communicationbandwidth coupled with high latency makes the wholenetwork vulnerable to congestion due to packet collisions.Consequently, minimizing the number of packet transmissionsis a very important criterion for protocol design.
Download full report
http://citeseerx.ist.psu.edu/viewdoc/dow...1&type=pdf