This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).
Modern automobiles and flight avionics systems form complex distributed real-time and embedded (DRE) systems. A high-end luxury car can have over 80 Electronic Control Units (ECUs), which are small embedded processors, and multiple networks linking the processors. Furthermore, several hundred software components can be distributed across these multiple networked ECUs. Optimizing the deployment of these software components, by packing the software more tightly onto the processors, can reduce the size of the required underlying infrastructure and have numerous positive side-effects, such as weight and power consumption savings.
Determining how to deploy software to hardware in DRE systems is a challenging problem due to the large number of complex constraints that must be dealt with, such as real-time scheduling constraints, component placement restrictions, and fault-tolerance guarantees. This research effort focuses on developing new hybrid heuristic and meta-heuristic techniques for determining how to deploy software to computational nodes. The algorithms and tools will be made available in open source through the Generic Eclipse Modeling System (www.eclipse.org/gmt/gems), which is distributed by 45 world-wide mirrors, and the ESCHER tool repository. Opportunities for outreach will be sought through existing mechanisms in place at Vanderbilt University, such as the NSF Science and Technology Center called TRUST and Vanderbilt Center for Science Outreach (CSO) to host summer research students. We will continue to support graduate students belonging to underrepresented groups.