A hybrid method of CSMA/CA and TDMA for real-time data aggregation in wireless sensor networks

  • Authors:

  • Qin Liu;Yanan Chang;Xiaohua Jia

  • Affiliations:

  • School of Computer, Wuhan University, Wuhan, Hubei Province, China;School of Computer, Wuhan University, Wuhan, Hubei Province, China;Department of Computer Science, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong

  • Venue:
  • Computer Communications
  • Year:
  • 2013

Quantified Score

Hi-index 0.24

Visualization

Abstract

We study the real-time data aggregation in contention-based wireless sensor networks that use CSMA/CA MAC layer protocols as defined in IEEE 802.15.4 or IEEE 802.11 standard. The problem is, for a given data aggregation tree and a delay bound, to maximize the average transmission success probability of all sensor nodes within the delay bound. In CSMA/CA protocols, the success probability and the expected transmission delay are highly sensitive to node interference, while the node interference is often very high in the large scale sensor networks. We propose a hybrid method that combines the CSMA/CA protocol with TDMA scheduling of transmissions. We divide the child nodes of a parent into groups and schedule the groups into different ''time-frames'' for transmission. Within the group, the nodes still use the CSMA/CA protocol to compete for data transmission. By doing so, we divide a large collision domain (i.e., all child nodes competing to transmit to their parent) into several small collision domains (i.e., a group of nodes competing for transmission), and the success probability can thus be significantly improved. On the other hand, the ''time-frame'' used in our method is much larger than the timeslot used in pure TDMA protocols. It only requires loose synchronization of clocks, which is suitable for low-cost sensor networks. We transform our objective of maximizing the average success probability into minimizing the average node interference. We then convert our problem to the maximum weight k-cut problem, which is NP-hard. We propose two efficient heuristic algorithms to solve the problem. Simulation results have shown that our proposed method can improve the success probability significantly.