Analysis and characterization of uncertainty components in the measurement of radio-electric signals and parameters

Resumo

Accurate measurements of radio-electric signals and parameters are fundamental to assure the correct performance of any wireless communication system. That kind of measurements encompass the characterization of the antennas used for signal transmission and reception, the sounding of the radio channel, that acts as propagation media, and the assessment of the electromagnetic field generated by the transmitting antennas. For a measurement to be complete the associated uncertainty has to be computed. That involves identifying the possible contributors to the uncertainty, determining their relation with the measured magnitude and quantifying and minimizing (if possible) their effect. Regarding the uncertainty of the measurements mentioned above, there still exists contributions to be done, even in those cases that have been studied for years, as the characterization of antenna parameters. This thesis aims to cover some of the existing gaps by providing more knowledge about some of the uncertainty sources in the measurement of radio-electric signals and some tips on how manage them. The analysis comprises the use of formulation, simulations and /or measurements that will be applied to different study cases. Ambient temperature and humidity are possible sources of uncertainty that all radio-electric measurements have in common. The attenuation effect of water vapour in earth-space communications has being studied for years and there still is recent research regarding its impact in terrestrial communications. However, this attenuation also affects the assessment of human exposure to the fields generated by the antennas and the evaluation of the parameters of such antennas. This effect has not been analysed in depth yet. Besides that, ambient conditions do not only degrade the transmitted/received signal, they may also induce thermal effects in the instrumentation or in the constructive material of antennas. In this thesis, we provide all the steps for a complete evaluation of the uncertainties that the ambient conditions cause in electromagnetic measurements. This evaluation is required to establish the impact in the overall uncertainty budget, and to decide if it is negligible or if we should apply ambient control mechanisms, and how to do it properly. A practical approach is followed by analysing some actual measurement cases, including outdoor and indoor scenarios, and thus with different specific ambient conditions. Through this analysis, we can state the relevance of this source of uncertainty in each case and provide recommendations on its control. The main conclusion achieved through this analysis is that its significance will depend on the test to be performed, on the measurement scenario and on the control we can apply on the temperature and humidity fluctuations. As key elements of a wireless communication systems, the uncertainty in the characterization of antenna parameters has been studied for decades. And it is still a matter of concern nowadays due to changes in technology that results in complex antenna design at higher frequencies, involving new instrumentation and measurements methods and/or facilities. The latter play a key role in the measurement uncertainty of antenna parameters. Using the minimization of the measurement uncertainty as the driving force during the design of a facility will result in more accurate results. How to do that is part of the work presented here. We provide a spherical far field anechoic chamber design methodology focused on reducing the uncertainty contribution of the facility in the measurement results. The control of the uncertainties derived from phase tapper and from mutual coupling and how to proper select the absorber are some of the critical points presented. Relevant information on enclosure and antennas supports and advice on the measurement instrumentation are also provided. For a better understanding, the methodology is illustrated with the design and construction of a rectangular anechoic chamber for antenna characterization at millimetre wave frequencies. As mentioned above, the assessment of human exposure to electromagnetic fields is another radio-electric measurement process linked to wireless communication systems. The arrival of the latest mobile network generations, LTE and 5G, involves the deployment of new base stations. The field level associated to their emissions must comply with the legislation on the matter, and that requires to carry out accurate measurements. The modulation schemas in the new generation allow high flexibility, which is translated into large fluctuations of the transmitted signal. As part of the research presented here, the feasibility of using the traditional exposure assessment based in broadband field probes is analysed. A key point is to quantify the uncertainties in the measured field level caused by the fluctuations in the LTE and 5G signal as the user load changes. Results show that uncertainties associated to an unknown user load are much higher than those coming from other contributors. As a consequence, a revision of the current legislation is necessary, establishing reduced decision levels able to assure exposure compliance. This will limit the amount of scenarios where the assessment can be performed with broadband field probes, increasing the situations where methods based in frequency or code selective measurements need to be used. Channel sounding is basic to understand how the waves are affected by the propagation media and then to establish the requisites necessary to have an operative communication link. This kind of measurements used to characterise the parameters of wideband propagation channels, do not have traditionally involved an uncertainty analysis. The main concern has always been the performance parameters of the sounder. However, the instrumentation can be an important source of error. When performing frequency measurements with a sounder based in a vector network analyser, the uncertainties of the measured transmission coefficients are propagated to the estimation in channel parameters. The dynamic range of the system is another source of inaccuracies. Processing decisions, as the number of samples to be averaged and the selection of the threshold to filter the noise present in the measured signal, also influence the estimation of channel parameters, and thus inducing certain level of uncertainty in the results. In this thesis, the combined influence in channel parameters estimation of the uncertainties from the measurement system and of those coming from the processing decision is studied. The uncertainty analysis of the RMS delay, path loss, first arriving ray amplitude and delay is done through simulations based in Monte Carlo method, where three different frequency responses, generated from a propagation model, are used as inputs. Simulations show that the uncertainties of the transmission coefficients add fake components to the impulse responses of the input signals, also altering the power of the true response components. This induces uncertainties in the channel parameter estimations that increase with the phase uncertainty of the transmission coefficient. The growing rate of this error will depend on the magnitude of the parameters and on the number of components in the original signal, and the final values of uncertainty on the relative amplitudes of the signal components. The selection of the noise threshold, together with the averaging of samples, appears as the processing decision of most influence for reducing the final uncertainty of the channel parameters. Another important outcome is that a minimum signal to noise relation is required if we want to keep the uncertainty to remain under a certain value.