[EMBEDDED SYSTEMS]
Topics of interest:
Designing Energy-&-QoS Aware scheduling algorithms for
embedded heterogeneous multiprocessor platforms; Memory management and
allocation for embedded systems; resource management and optimization;
Real-time scheduling algorithms; Application - Body-Area Networks (BANs)
=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
ABOUT MY FUNDED PROJECT:
(Current Extern= al Funding) [Embedded Hybrid Systems] Received a competitive extern= al academic research grant from A*STAR(Agency for Science, Technology and Research) SER= C in November 2005 for ~ USD 182,000.00(~S$301K). (Title: <= i>Design of Adaptive and Hybrid Energy-&-QoS Aware Heterogeneous Multiprocessor Scheduling Strategies for Embedded Systems).
(Note: =
Project
duration is for three years from January 2006 - currently recruiting Resear=
ch
Engineers/Assistants - send in your detailed CV to ele=
bv@nus.edu.sg
with a cover letter)
START DATE – 01/March/2006; END DATE – 28/Februar=
y/2009
I am the Principal Invest= igator of this project looking at issues related to the design of energy and QoS a= ware scheduling algorithms for embedded heterogeneous processors.
BRIEF DESRIPTION OF THE PROJECT (but certainly sufficient to understand wh= at is going on!)
While energy management a= nd QoS control were independently studied in the literature, there has been relati= vely few works that took into consideration both these factors for scheduling and that too under uni-processor platforms. Current= ly existing solutions cannot address the problem of minimizing the energy consumption while satisfying required QoS levels in heterogeneous multiprocessor systems. We intend to explore the design of an adaptive hybr= id scheduling methodologies that integrate both Energy and QoS parameters in designing a scheduler for heterogeneous multiprocessor systems suitable for embedded systems. We attempt to design and validate our schemes in real-life platforms that are useful for mobile terminal applications. Specific list of objectives/deliverables are a listed below.
Our design methodology and solution approach have unique advantages. For real-time mission-critical ta= sk processing demands, it is imperative to have a scheme that adapts to dynamic environments by allowing decisions to be made during run-time. Our schemes indeed will take into consideration all these factors in generating feasibl= e schedules. Secondly, our approach to the problem is novel and is of first kind in considering a combined influence of these two conflicting yet useful metric= s.
The novelty of our approa=
ch is
evident from the fact that there is no work to date that integrates QoS and
Energy parameters and optimizes the performance. Design of hybrid and adapt=
ive
strategies to achieve the above mentioned objectives will be novel
contributions to the literature. The proposed work is of first-kind in its
nature for reconfigurable multiprocessor platforms and the outcomes of this
project are expected to spur further research in this domain. An integrated
approach that schedules the real-time tasks with the consideration of both
energy conservation and QoS provisioning will be extremely useful in many
mobile devices, for instance mobile phones with multimedia streaming
capabilities. The algorithms and approaches that are to be designed in this
project will be tested/applied for Body-Area Networks (BANs).
=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D