Digital System Design for the Maryland Smart Dust Project

 


Project Director :

Shuvra S. Bhattacharyya

Associate Investigator :

Chung-Ching Shen

Maryland DSPCAD Research Group, University of Maryland, College Park


Project Description

The Maryland Smart Dust Project was a multi-year, interdisciplinary project at the University of Maryland, College Park. The project was sponsored under the direction of the Laboratory for Physical Sciences. The Maryland DSPCAD Research Group ("DSPCAD Group") was responsible for the Digital System Design thrust of the project. The objective in this thrust was to develop efficient design methods and energy optimization techniques for low power, application-specific wireless sensor network (WSN) systems.

In this project, the DSPCAD Group demonstrated that by applying its dataflow based design methodologies, WSN applications can be optimized and retargeted with high productivity and confidence to different embedded platforms, including microcontroller-, FPGA-, and ASIC-based platforms,. We also
demonstrated a novel application-specific integrated circuit (ASIC) implementation for a WSN-based security application. This ASIC solution provided significant power reduction (approximately a 10X improvement) compared to related COTS (commercial off-the-shelf) platforms.

WSNs often have limited hardware resources and are deployed in areas that are difficult or impossible to access by humans. In such environments, energy efficiency becomes a major concern. Energy-efficient operation for advanced WSN systems becomes especially challenging if the system must perform complex computation and communication tasks. In this project, we also developed a novel methodology called energy-driven partitioning (EDP). Our EDP methodology determines strategic configurations for distributing computational and communication tasks across WSN systems to optimize energy efficiency. By applying our EDP methods, we have demonstrated that the energy consumption of  WSNs can be reduced by 50% or more compared to conventional approaches for WSN application deployment.
 


System Design Flow


MCU-based Platform


FPGA-based Platform


ASIC-based Platform


Publications


Contact Information

For comments on the webpage, please contact Chung-Ching Shen