Optimising lower layers of the protocol stack to improve communication performance in a wireless temperature sensor network

Abstract

The function of wireless sensor networks is to monitor events or gather information and report the information to a sink node, a central location or a base station. It is a requirement that the information is transmitted through the network efficiently. Wireless communication is the main activity that consumes energy in wireless sensor networks through idle listening, overhearing, interference and collision. It becomes essential to limit energy usage while maintaining communication between the sensor nodes and the sink node as the nodes die after the battery has been exhausted. Thus, conserving energy in a wireless sensor network is of utmost importance. Numerous methods to decrease energy expenditure and extend the lifetime of the network have been proposed. Researchers have devised methods to efficiently utilise the limited energy available for wireless sensor networks by optimising the design parameters and protocols. Cross-layer optimisation is an approach that has been employed to improve wireless communication. The essence of cross-layer scheme is to optimise the exchange and control of data between two or more layers to improve efficiency. The number of transmissions is therefore a vital element in evaluating overall energy usage. In this dissertation, a Markov Chain model was employed to analyse the tuning of two layers of the protocol stack, namely the Physical Layer (PHY) and Media Access Control layer (MAC), to find possible energy gains. The study was conducted utilising the IEEE 802.11 channel, SensorMAC (SMAC) and Slotted-Aloha (S-Aloha) medium access protocols in a star topology Wireless Temperature Sensor Network (WTSN). The research explored the prospective energy gains that could be realised through optimizing the Forward Error Correction (FEC) rate. Different Reed Solomon codes were analysed to explore the effect of protocol tuning on energy efficiency, namely transmission power, modulation method, and channel access. The case where no FEC code was used and analysed as the control condition. A MATLAB simulation model was used to identify the statistics of collisions, overall packets transmitted, as well as the total number of slots used during the transmission phase. The bit error probability results computed analytically were utilised in the simulation model to measure the probability of successful transmitting data in the physical layer. The analytical values and the simulation results were compared to corroborate the correctness of the models. The results indicate that energy gains can be accomplished by the suggested layer tuning approach.