Performance Evaluation of a Wireless Protocol for Mesh Networking Under the Influence of Broadband Electromagnetic Noise

Migrating from a wired to a wireless implementation for communication system used in industrial applications is a logical move to avoid the many shortcomings associated with wires. When operated under harsh environments, those wires can break and could cause not only damage to the system but also endanger human lives. However, it is not well documented how well a wireless protocol can work under such harsh industrial environments. This thesis attempts to answer this research question in the point of view of gauging the performance of a wireless protocol under the influence of electromagnetic noise. More specifically, the type of noise signal that is the focus of this investigation is the random, pulsed type (e.g., discharges caused by sparking) that creates a hyperbolic broadband disturbance in the frequency domain. Consequently, a fractal noise model is used to study noise of this nature. The steps toward achieving this goal include: requirements gathering, wireless technology selection; noise modelling and synthesis; real noise capture and analysis to validate the chosen noise model; high-frequency fractal noise emulation in hardware; the use of a novel noise injection method for empirical work; and the conducting of a controlled synthetic noise-to-wireless node performance evaluation to obtain performance measure in the form of packet error rate (PER). Performance data in terms of PER versus signal-to-noise ratio (SNR) for various nodes separation have been collected. There were three significant findings: the obtained performance curves follow the standard'S' trend; for a specific desired reliability (denoted by a certain PER), the SNR at the transmitter needs to be boosted as the correlation of the noise being present increases; and the maximum distance between nodes separation for a certain reliability to be achieved depends exponentially with the transmitter's SNR. The relationship in the third finding assists in placement of wireless nodes, which in turn can determine the minimum amount of wireless nodes required for an industrial system to reach the desired system reliability, thus boasting network cost saving.