PROPOSAL NO.: 0408996 PRINCIPAL INVESTIGATOR: Anderson, James H. INSTITUTION: U of NC Chapel Hill PROPOSAL TITLE: Real-time Fine-grained Adaptivity on Multiproces-sors: Acoustic Tracking as a Test Case
Two trends are evident in recent work on real-time systems. First, multiprocessor de-signs are becoming quite common. Second, many applications now exist that require fine-grained adaptivity, i.e., the ability to react to events within short time scales by ad-justing task parameters. Examples of such applications include systems that track people and machines, many computer-vision systems, and signal-processing applications such as synthetic aperture imaging. The goal of this project is to develop techniques for imple-menting such systems on tightly-coupled multiprocessors. The central thesis of this pro-ject is that this goal can best be met by using multiprocessor fair scheduling algorithms. Such algorithms are designed to mimic the allocations of an ideal fluid scheduler, which can react instantaneously to workload changes, while still ensuring real-time constraints.
Key objectives of this project include the development of (i) multiprocessor scheduling algorithms that facilitate adaptation, (ii) kernel services for synchronization and I/O to be used in conjunction with these algorithms, and (iii) mechanisms for determining how and when to adapt. The resulting platform is to be evaluated by implementing a spread-spectrum acoustic full-body tracking system for virtual environments. The proposed tracking system is a good surrogate for the wide range of highly-adaptive computation-ally-intensive real-time applications that exist. Results of this research will be dissemi-nated through publications. Any software of general utility that is produced will be made publically available on the web. This study will lead to a better understanding of the re-source-allocation issues unique to real-time applications with fine-grained adaptivity re-quirements implemented on multiprocessors.
Dr. Brett D. Fleisch Program Director, CISE/CNS June 25, 2004 .
