60. DISTRIBUTED TRACKING FOR UNDERWATER NETWORKED SWARMS

Department: Computer Science & Engineering
Faculty Advisor(s): Ryan Kastner

Primary Student
Name: Karl Magnus Delight
Email: kdelight@ucsd.edu
Phone: 858-534-8908
Grad Year: 2013

Abstract
Location information is crucial to many oceanographic applications. Further, this information is needed in real-time for control and navigation of underwater vehicles and in early warning systems. We formulate the problem of location tracking in mobile underwater systems when vehicles are networked using acoustic modems. For such systems, traditional oceanographic techniques that track the position of each vehicle independently do not scale well. This is because they rely on direct communication with a few reference transponders (often deployed on the surface or sea-floor). In addition, these techniques are less reliable when vehicles experience higher mobility or move in unpredictable ways. To overcome these disadvantages, we propose a collaborative tracking method that estimates the position of the vehicles in real-time and in a completely distributed fashion. Our proposed method leverages from spatial constraints on the relative position of vehicles. These constraints are created by allowing devices to acoustically communicate with each other in order to estimate inter-node distances. Further, information flows both forward and backward in time, thereby improving the accuracy of past estimates, given a delay bound on computing the locations as dictated by any specific application.

« Back to Posters or Search Results