Graph coloring algorithms for TDMA scheduling in wireless sensor networks

Author

Tiegeng Ren, University of Rhode Island

Date of Award

2007

Degree Type

Dissertation

First Advisor

Timothy Henry

Abstract

Wireless Sensor Networks (WSN) as an emerging technique has received increasing attention in both research and application field. Since these Micro-Electro-Mechanical systems (MEMS) are constrained by power, physical size, and wireless communication range, efficient use of the technique leaves many challenges for researchers. One of the challenges comes from the medium access control (MAC), which determines the physical and network layer communication model within the network. In this research, we focus on MAC protocol, investigate into the mathematical model of the MAC protocols in graph theory, and develop the following algorithms that would improve the performance of data delivery applications in wireless sensor networks. First, algorithms for TDMA (Time-Division Multiple Access) based conflict-free MAC protocols in wireless sensor networks are developed. These algorithms are based on a novel graph coloring technique called the Color Constraint Heuristic (CCH). The algorithms create a near optimal TDMA frame size for the system, which in turn produces a near-optimal periodic data throughput. Both a centralized TDMA slot assignment (CSA) algorithm that uses CCH and assumes knowledge of the entire network topology and a distributed version of the algorithm, Distributed Slot Assignment (DSA) are developed and presented in this work. Second, another TDMA coloring scheme for data aggregation applications in wireless sensor networks is presented. With careful investigation on data aggregation applications in sensor networks, a theoretical lower bound for any TDMA based protocol was found. By an in-depth comparison of node based coloring and edge based coloring, a hybrid coloring algorithm called RAC-CT (Routing Aware Coloring - Conflict Tolerant) is developed. Thirdly, we present a mathematical approach to show that the total number of colors TDMA would use is bounded by a 13 approximation of ω(G) and also it converges to 2Δ(G) in a grid random in any type of sensor networks deployment. Analytical and experimental results show that our algorithms are not only affordable in implementation, but also, yield near optimal performance for data query applications in wireless sensor networks.

Recommended Citation

Ren, Tiegeng, "Graph coloring algorithms for TDMA scheduling in wireless sensor networks" (2007). Open Access Dissertations. Paper 2213.
https://digitalcommons.uri.edu/oa_diss/2213