Strategy for wireless transmission disturbances detection and identification in industrial wireless sensor networks: Dissertation

This doctoral thesis presents a novel method for detecting and identifying radio channel disturbances. Wireless transmission in factory environments occasionally becomes degraded due to harsh radio signal propagation conditions. The most disruptive disturbances are caused by other wireless devices (co-channel and adjacent interference), environmental characteristics such as metal constructions and heavy equipment (signal fading), and electrical and mechanical equipment (noise). We have performed extensive measurements in different industrial environments to study the effects of different radio channel disturbances on signal propagation. The obtained measurement results led us to develop signal analysis methods and a disturbance detector based on a classifier. The classifier was tested in a real industrial environment; these tests showed an encouraging detection performance. Signal analysis is performed both in time and frequency domains. In the time domain, the probability density function (PDF) method is used. It is based on detecting the PDF shape variations calculated from the received signal and their correlation to the environmental changes in the radio channel. In the frequency domain, spectrogram analysis is performed. Here, we calculate the spectrogram of each received data packet, and by applying image analysis tools on the spectrograms, we obtain the required information on the received signal and on other radio transmissions which may have disturbed our radio signal transmission. The recognition and identification of the radio channel disturbances is based on the fusion of the signal magnitude analysis (PDF) and spectrogram analysis with a classifier. We also discuss the development process of a classifier-based tool towards an embedded solution. The classifier was developed and extensively tested using USPR N210, ETTUS SDR (Software Defined Radio). The classifier proved to be a feasible solution for detecting and identifying reliability flaws of wireless transmission in the industrial environment, and will be further developed into a portable, small-sized SDR-based tool.