Main / CoursesNES

Course - Networked Embedded Systems

Spring 2008
Prof. Seapahn Megerian

Draft version - December 10, 2007.

Capsule Statement of Course Content:

  • Study of system design, analysis, implementation, and deployment of large scale networked embedded systems.
  • Main focus is on wirelessly connected autonomous sensing and control networks for a wide range of industrial, consumer, and scientific applications.
  • Discussions on the underlying protocols for ad hoc networking, networked sensing, optimization, self organization, and resource management.

Course Description:

Large scale distributed and networked embedded systems are an emerging class of embedded systems fueled by miniaturization, large scale integration, and wireless communication technologies. The applications of such systems range from distributed control and actuation systems, monitoring and surveillance systems, and enabling technologies for ubiquitous and proactive computing paradigms. While these technologies still face many traditional system building challenges, they pose a number of new issues that are critical in realizing practical solutions. The recent flurry of industrial and research efforts in this domain have resulted in the emergence of a rich body of literature, experimental hardware and software platforms, as well as the proliferation of applications for their use. In this class we will look at a number of system level design issues and the existing and potential approaches for tackling the new engineering problems. The covered issues include, but are not limited to topics such as platforms, ad hoc networking, localization, autonomous sensor calibration, time synchronization, power management, collaborative signal processing and optimization, data fusion, storage, sensor and network coverage, and security. Although the topics of this class by their nature are broad, no specific prior knowledge will be assumed and the class will be fairly self contained. Thus, the exact coverage of the material and their related background will depend on the backgrounds of the students attending the lectures (i.e. class participation is strongly encouraged!).

Prerequisites

Specific computer networking and distributed system knowledge is not required as the course is designed to be self contained. Background and relevant material will be reviewed as deemed necessary by the instructor.

Tentative Course Outline:

  1. Introductions and course overview
  2. Application specific embedded systems
  3. Distributed and networked embedded systems
    1. Pull by existing needs and applications
    2. Push by technological advances
    3. Emerging applications
  4. Platforms
    1. Processing and storage fabrics
    2. Communication subsystems
    3. Sensors and actuators
    4. MEMs
  5. Wireless Communication
    1. Physical layer and Data Link
    2. Medium Access Control
    3. Broadcast and data dissemination
  6. Location awareness and localization
    1. Localization fundamentals
    2. Proposed schemes
    3. Distributed localization
  7. Time synchronization
  8. Ad Hoc Networking
    1. Discovery and routing
    2. Topology formation
    3. Clustering
    4. In-network processing
  9. Sensor and actuators
    1. Types of sensors
    2. Sensor coverage and deployment
    3. Data collection and fusion
    4. Information processing
    5. Sensor calibration
    6. Distributed actuation
  10. Application case studies
    1. Target detection and tracking
    2. Environment monitoring
  11. System planning and Deployment
  12. Reliability and fault tolerance

Grading:

  1. Homework: 20%
  2. Exam: 40%
  3. Project: 40%

Notes:

The material in class will be based on a number of recently published papers and articles as well as online resources. Homeworks, exams, and projects will include quantitative analysis and system design related problems as well as surveys and critiques of the papers assigned in class. Active class participation and discussions will be strongly encouraged. Projects dealing with actual hardware or using data from real systems (e.g. sensors, network or power measurements) will be considered more favorably than simulated ones. Survey paper type projects can be used as a last resort.