class has moved to Gates 392 as of April 17, 2003
This course is a graduate-level introduction to information processing in sensor networks. The material covered will introduce students to the diverse literature on sensor networks and expose them to the fundamental issues in designing and analyzing sensor network applications. A lab allowing hands-on experience with sensor nodes will be available. Localization and tracking will be used as canonical examples to expose important constraints in scaling and deploying sensor networks. The course will also cover techniques such as data routing, in-network processing, information aggregation and querying, energy-aware processing, etc., and show how these methods can support high-level information processing tasks for a variety of applications.
The course is aimed both at students who wish to do research in the sensor networks area, as well as at students from related disciplines, such as signal processing, networking, databases, algorithms, etc., who wish to understand what new challenges sensor nets pose for their own discipline.
Consent of the instructors.
The class requirements include:
Information processing in sensor networks is an interdisciplinary research area with deep connections to signal processing, networking and protocols, data bases and information management, as well as distributed and on-line algorithms. Because of advances in MEMS micro-sensors, wireless networking, and embedded processing, ad-hoc networks of sensors are becoming increasingly available for commercial and military applications such as environmental monitoring (e.g., traffic, habitat, security), industrial sensing and diagnostics (e.g., factory, appliances), infrastructure maintenance (e.g., power grids, water distribution, waste disposal), and battlefield awareness (e.g., multi-target tracking). There is an active research community pursuing sensor network related research in U.S. and elsewhere, funded in part by DARPA (e.g. SensIT, NEST, PAC/C), ONR, and NSF. The State of California recently established the CITRIS Center at Berkeley, with a primary focus on sensor networks.
However, because this is an emerging research area involving a variety of different technologies, major contributions to sensor networks today are scattered in many specialized conferences on wireless networking, sensors, data fusion, signal/image processing, probabilistic reasoning, and robotics. For example, Infocom, Mobicom, Mobihoc, Fusion, and ICASSP regularly have papers on various topics concerning sensor networks. A practitioner in sensor network research often has to be versed in several disparate research areas before he/she can start to contribute to the sensor network research.
From the computer science point of view, sensor networks become a rich source of problems in communication protocols, sensor tasking and control, sensor fusion, distributed data bases and algorithms, probabilistic reasoning, system/software architecture, design methodologies, and evaluation metrics. This is a vast space of problems to explore, but at its core are a set of key issues about information dissemination, aggregation, and storage from data collected in a distributed fashion over space and time, using sensor nodes dynamically commanded to sense and communicate according to the task at hand.
The aim of this course is to expose students to the fundamental issues and technology constraints of sensor networks, illustrate how techniques from the many areas mentioned above can be integrated into useful systems, and serve as preparation for research in this nascent area.
These pages are maintained by Leonidas Guibas email@example.com.