Welcome to Nanyan Jiang's Homepage
Home - Projects - Publications - Professionals - Courses - Personal - Links
GridMap Programming System

Programming Sensor-based Scientific and Engineered Applications

Abstract

End-to-end autonomic management through a sensor-driven integration of physical and computational worlds.

The emergence of pervasive computational ecosystems that integrate computing infrastructures with embedded sensors and actuators are giving rise to a new paradigm for monitoring, understanding and managing natural and engineered systems -- one that is information/data-driven and autonomic.

 

 

 

 

 

 

 

Objective

The research objective is to provide programming abstractions and system software support for sensor-driven applications

  • Programming abstractions and runtime systems for integrating sensor systems with computational models and other application components in an end-to-end dynamic sensor-driven scientific and engineered application.

  • Programming abstractions and runtime systems for developing in-network data processing mechanisms.

  • Efficient in-network computation/communication algorithms in resource constrained heterogeneous sensor networks.

  • Support for tradeoffs between data quality, resource consumption and performance.
•
–
–

Conceptual Overview

A conceptual overview of the overall approach is illustrated in the above Figure. The goal of the programming system being developed as part of this research is to provide abstractions and mechanisms to seamlessly access and integrate a wide area sensor data into computational models and support scalable in-network data processing. The underlying approach is to virtualize the physical sensor grid to match the representation of the physical domain used by the models, and dynamically discover and access sensor data independent of any change to the sensor network itself. The sensor network periodically estimates data at the grid points specified by the application using available physical data values. The estimation mechanisms are specified by the applications are and implemented within the sensors network in a decentralized and scalable manner.
 

Publications

Programming Sensor-based, Dynamic Data-driven Scientific Applications, Nanyan Jiang, Manish Parashar, IEEE International Parallel & Distributed Processing Symposium (IPDPS) Ph.D Forum, Miami, FL, April, 2008

Programming Support for Sensor-based Scientific Applications, Nanyan Jiang, Manish Parashar, The Next Generation Software (NGS) Workshop held in conjunction with IPDPS, Miami, FL, April, 2008

 

Related Works

 

 

Last updated 07/25/2008


Meteor is designed and developped at Rutgers University's TASSL lab at the CAIP Center