Measurement, characterization, and modelling of the indoor radio propagation channel

by Morrison, Gerald Dale

The motivation for studying the indoor radio propagation environment is spurred on by recent developments in the wireless industsy as a whole. New wireless systems will need to be portable, light weight, handle a large number of users, and provide a wide variety of services. It is expected that over the next several years indoor wireless will be heavily used by providers of such services as Intemet-enabled ce11 phones and persona1 digital assistants, paging systems, wireless local area networks, persona1 communications networks, and cordless phones. To achieve acceptable and cost effective wireless performance, new systems will become increasinglymore complex and will continue to push the envelope for symbol transmission rate. To be successfbl, system designers will have to contend with the radon and complicated nature of indoor radio transmission. This will require an intimate knowledge of how the indoor radio propagation channel behaves and this necessity leads to characterizationand modelling. Accurate characterization and good modelling can yield information that is invaluable to the radio systerns designer. As a result, characterization and modelling of the indoor radio environment has received a lot of attention in recent years. The work presented in this thesis is in three parts. First a technique is examined for measuring the fi-equencyresponse characteristics of the channel. The more familiar impulse response estimate is then available by transformation. Typically, Fourier methods can be employed. Second, the radio channel is characterized by extracting interesting information directly fiom the measured data. Third, statistical models are constructed base on mode1 parameters derived fiom the measured data. This thesis shows five different models for the indoor radio propagation channel. This is unusual since, traditionally, previous modelling efforts described in the Iiterature show one model in a single dissertation or paper. Therefore, this work represents the first time a cornparison of competing models is set forîh. As well, an explanation of the relative merits and assurnptions of each model is given.. Furthermore, with the exception of the Peak Extractor in Chapter Six, the models are al1 newly applied techniques in the field. The addition of newer modelling techniques demonstrates that the traditional Peak Extractor model does not model the radio channel accwately. In previous modelling work, contributors have not regenerated the channel fiom the theoretical model for coniparison to the data fiom which the model was denved. As a result, the models can only be assumed to be working as expected. In this work, every time a model is constmcted, the mathematical representation of the model is used to regenerate data and then that result is cornpared to the original empirical measurement to test for validity. Therefore, it can be stated with certainty as to how well the model represents the actual data. The information reported in this thesis provides charactenzation information and models that will assist communication system engineers and manufacturers in obtaining a better understanding of the complex indoor radio environment.