Studies On Adaptive Medium Access Control Protocol For Wireless Sensor Networks

Wireless sensor networks (WSNs), which appear along with the development ofembedded system, wireless communication, networks and micro-electro-mechanical systemhave recently attracted the interests of researchers in these years. WSNs are dynamicallyself-organized and self-configured, with nodes in the networks automatically establishing anad hoc network and maintaining the mesh connectivity. With the advantages of low cost andeasy to be deployed, WSNs can be widely used for various purposes.Medium access control (MAC) protocols which control how neighboring nodes access ashared wireless channel are fundamental protocols for WSNs. Given that most node energy isconsumed by radio units, large energy gains can be achieved in MAC layer which operatesradios directly. Thus, proper design of the MAC protocol is critical to the lifetime of WSNs.Focusing on energy efficiency and adaptability to dynamic traffics, this thesis studies andpenetrates MAC protocols for WSNs. After modeling the behavior of MAC protocols anddeeply analysing the inbeing factors which influence the protocols performances, this thesispresents alternative concepts to improve protocols performances and does some usefulattempts as well. The major contributions of this thesis include:1) Classifies and summarizes the existed protocols according to the methods they accessthe channel; provides a taxonomy of current work based on protocol mechanism and thechronology of their appearance; models the performances of some classic protocols in lowtraffic conditions and details both their advantages and disadvantages; makes a comprehensive comparison between these classic protocols based on the radio CC2430.2) Sets up a Markov chain based model for analyzing synchronous protocols. A Markovchain to describe the backoff period is proposed at first. After that a Markov chain to describethe node states in active period is proposed immediately. Network performances includingthroughput, latency and energy consumption are deduced by these two Markov chains.Simulations based on NS2prove the accuracy of this model.3) Sets up a Markov chain based model for analyzing asynchronous protocols. Thismodel improves existing work by introducing the assumption that a node may transmit morethan one packet in a cycle and the hidden terminals may influence transmission. A Markovchain to describe the length of node's buffer queue is proposed at first. Network performancesare deduced after figuring out the transition probability for the queue length, the probabilitythat nodes access the channel and the probability that several packets can be transmitted in acycle. NS2based simulations show that this model is more accurate than existing work.4) Designs an adaptive asynchronous protocol based on greedy listening, namedAA-MAC. The optimal greedy listening time is deduced in the constraint of highest energyefficiency. Simulation results show that this protocol is more efficient than traditionalasynchronous protocols and it decreases energy consumption by70%compared to X-MAC.However, AA-MAC only performances better in density and high traffic condition networks.5) Considering AA-MAC only fits for density and high traffic networks, designs anadaptive polling interval and short-preamble-based asynchronous protocol, named AX-MAC.Nodes dynamically adjust their polling intervals according to network traffic conditions inthis protocol. Threshold parameters used to determine traffic conditions and adjust pollingintervals are analyzed based on a Markov chain. Simulation results show that AX-MAC canperform better than traditional protocols by choosing appropriate threshold parameters invarious conditions.6) Considering the asynchronous mechanism is not efficiency in high traffic condistions, designs a multi-channel based hybrid protocol, named MCH-MAC. In this protocol, nodeschoose synchronous channel polling mode in low traffic conditions and choose TDMA modein high traffic conditions. An unbalanced slot assignment algorithm is designed to solve thefunneling problem in WSNs. Meanwhile, a flavor multi-channel based strategy to improvethe utility of TDMA slots is also designed. Implementations on CC2430and TinyOS showthat MCH-MAC performs better than Tree-MAC by increase20%in terms of networkthroughput.