Implicit Dynamic Space Time Clustering: A New Clustering Protocol for Sensor Networks
Open Access
Author:
Rish, Anna Claire
Area of Honors:
Computer Science
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
Shashi Phoha, Thesis Supervisor Dr. John Joseph Hannan, Thesis Honors Advisor
Keywords:
sensor networks clustering information fusion routing
Abstract:
Progress in engineering small, inexpensive sensors has allowed the eld of autonomous sensor
networks to grow immensely in recent years. In order for the sensor networks to be useful
in monitoring a large area, the sensors must be able to share and interpret their own data.
Many protocols have been created which dictate clustering and routing routines for these sensor
networks. This paper introduces Implicit Dynamic Space-Time Cluster (IDSTC), a new protocol
which creates clusters and routes based on routing decisions made by GODDeS. GODDeS is a
theoretical framework which provides hop-by-hop routing decisions based on modeling the end-to-
end drop probability for a packet in the network as a Probabilistic Finite State Automata (PFSA).
By relying on a proven-optimal theoretical result to make information routing decisions, IDSTC
guarantees the best possible probability of end-to-end packet success. It also introduces many
new features including network querying, scalability, failure tolerance, and multi-hop clustering.
In this paper, IDSTC is compared to three major protocols from the literature, and our evaluation
shows that it is certainly a protocol worth studying.
i