6th International ICST Symposium on Modeling and Optimization

Research Article

Making Distributed Rate Control using Lyapunov Drifts a Reality in Wireless Sensor Networks

Download441 downloads
  • @INPROCEEDINGS{10.4108/ICST.WIOPT2008.3205,
        author={Avinash Sridharan and Scott Moeller and Bhaskar Krishnamachari},
        title={Making Distributed Rate Control using Lyapunov Drifts a Reality in Wireless Sensor Networks},
        proceedings={6th International ICST Symposium on Modeling and Optimization},
        publisher={IEEE},
        proceedings_a={WIOPT},
        year={2008},
        month={8},
        keywords={Algorithm design and analysis Communication system control Distributed control Multiaccess communication Pressure control Protocols Stability Stochastic processes Testing Wireless sensor networks},
        doi={10.4108/ICST.WIOPT2008.3205}
    }
    
  • Avinash Sridharan
    Scott Moeller
    Bhaskar Krishnamachari
    Year: 2008
    Making Distributed Rate Control using Lyapunov Drifts a Reality in Wireless Sensor Networks
    WIOPT
    IEEE
    DOI: 10.4108/ICST.WIOPT2008.3205
Avinash Sridharan1,*, Scott Moeller1,*, Bhaskar Krishnamachari1,*
  • 1: Ming Hsieh Dept. of Electrical Engineering, University of Southern California, Los Angeles, CA 90089, USA.
*Contact email: asridhar@usc.edu, smoeller@usc.edu, bkrishna@usc.edu

Abstract

We take a top-down approach of formulating the rate control problem, over a collection tree, in a wireless sensor network as a generic convex optimization problem and propose a distributed back pressure algorithm using Lyapunov drift based optimization techniques. Primarily, we show that existing theoretical results in the field of stochastic network optimization can be directly applied to a CSMA based wireless sensor network using our novel receiver capacity model. We back this claim by implementing our algorithm on the Tmote sky class devices. Our experimental evaluation on a 5 node testbed shows that the empirically observed rate allocation on a real sensor network testbed that uses our back pressure algorithm is close to the analytically predicted values, justifying our claims.