Wireless Sensor Network and Radio Wave Propagation in Harsh Environments

Abstract: This licentiate thesis cover two important subjects regarding the application of wireless sensor networks (WSNs). Both subjects are related to propagation mechanisms. The first subject is what the radio channel looks like and how it behaves. In this case three different, and extreme, environments are measured and characterized - a train, a half scale jet engine fan and a full scale military jet engine. The train environment is characterized by measure the path loss and fading over distance. For the case with the jet engines, difficulties were found to measure the path loss over distance, so in this case stationary antennas were used, but with the engine running. Each of these shows an extreme type of fading, also known as Rayleigh fading. For the case of jet engine at full speed (about 10 000 rpm), it was shown that the time between two consecutive fading dips where only 290 μs, which is about twice the length of a data package from the WSN involved in this project. The Rayleigh distributed amplitude fading occurs when there is a multipath environment, the radio waves propagate several different paths between the transmitter and receiver, which causes a superposition at the receiver. When having Rayleigh fading, the performance of the radio link is greatly reduced. When applying a WSN in this type of environment, the use of several antennas will improve the received power of the signal. This is done by adding extra antennas to a wireless system and in a clever way combine the signals, or select one of two signals from the antennas. In a book chapter and in a paper presented in the thesis, a new and low energy type of diversity is described. The performance of this new type of diversity is shown by having a two branch diversity and discretely shifting the phase of each branch before combining them. In this case, four relative phase shifts are performed during each symbol received. When performing a combining like this, the energy is saved by not having any decision circuitry. By using this type of diversity and using a 90 % signal reliability and an ideal environment, the diversity gain is 5.5 dB for an averaging detector and 10.3 dB for a peak detector.  The drawback with this system is that it is only limited for systems using simple types of amplitude (ASK) or frequency (FSK) modulation.