Adaptive Systems

CSCE 5013

Class Description

Increasing complexity due to rapid progress in information technology is making systems more and more difficult to integrate and control. Due to the large amount of possible configurations and alternative design decisions, the integration of components from different manufacturers in a working system cannot be done only at design-time anymore. Furthermore the miniaturization of systems makes them more vulnerable to errors that may occur due to physical degradation, cosmic radiation, unpredictable interconnect delay on signals within chips. This increases the risk of failure. Systems must be designed to cope with unexpected run-time environmental changes and interactions. They must be able to organize themselves to adapt to change and avoid non desirable or destructive behaviors. Natural systems have evolved to cope with dynamism, unpredictability, uncertainty, lack of guarantees. Several initiatives (Swarm Intelligence, Organic Computing, Autonomic Computing) were introduced in the past with the goal of designing and building highly reliable and robust systems by borrowing the properties of natural systems. However substantial efforts and competencies from different fields are required in order to make the dream of Autonomic and Organic computing real, in particular in embedded systems.

The goal of this course is the understanding of mechanisms governing adaptive systems as well as the design approaches for those systems. They are usually made upon components limited in their capabilities, but as whole are able to solve a complex problem. They can adapt their behavior to run-time environmental changes. The design of software as well as the hardware, in particular reconfigurable systems, will be discussed in the class. Finally we will discuss applications (Software Defined Radio, Adaptive Antennas, Adaptive Video and Signal Processing, Adaptive Network Protocols, Adaptive Crypto Systems, Adaptive Network Topology) that can benefit from the run-time adaptivity of a system.

 

Grading

  1. 20% Seminar: Each class participant will pick a topic, read a paper that he will present in the class and lead the following discussion
  2. 80% Participation to the class project: All the student will work on a project with the goal of conceptually and technically design an adaptive system. The results of the project will be submitted at a conference for publication and will be used for participation to a student design contest.

Readings

  1. Fault-Tolerant Systems,by Israel Koren and C. Mani Krishna, Morgan Kaufmann. ISBN-10: 0120885255, ISBN-13: 978-0120885251
  2. Adaptive Control, Karl Johan Astrom, Bjorn Wittenmark, Publisher: Prentice Hall; 2 edition, December 31, 1994, ISBN-13: 978-0201558661
  3. Stable Adaptive System, by Kumpati S. Narendra, Anuradha M. Annaswamy, Dover Publications ISBN-10: 0486442268, ISBN-13: 978-0486442266
  4. Conference and position papers. Will be handed out during the semester