| Femtocell is one of the key indoor technologies in the fifth generation(5G) mobile communication. Under the femtocell scenario, due to the small radius and the irregular shapes of the cell, the propagation environments are becoming more and more complex, therefore it is necessary to further study the propagation characteristics, i.e., path loss, multipath effect, wall insertion loss. In this dissertation, the frequency domain measurements are extensively carried out under several typical environments: office, meeting room, corridor and stairwell, and a large number of measurement data are accumulated. The propagation parameters under indoor environment, such as the path loss, the multipath propagation and the wall insertion loss are analyzed and modeled.First of all, a novel path loss model under the indoor office, meeting room, corridor and stair environment with antenna-height dependency at 2.6GHz band is proposed in this dissertation. In order to simulate the attenuation caused by the variation of antenna-height, a new height attenuation factor(HAF), is added into the usual local path loss model. Especially, under the stair environment, a floor attenuation factor is added into the model for describing the multifloor attenuation. The path loss exponent, the antenna-height dependent factor HAF and the floor attenuation factor FAF are obtained by using the least square method. The results indicate that the HAF can be expressed as the quadratic function of the logartithm of the antenna height. Besides, the influence of antenna-height’s variation on the shadowing effects is studied and the empirical expression of standard deviation about antenna-height is obtained. The proposed model is verified by employing plenty measurement results under the similar structure and different scenarios. The results present that the proposed model is correct and effective.Next, under the office, meeting room, corridor and stair environment, the property of the multipath propagation is studied and a novel tapped delay line model is proposed. Through accumulating the massive measurement data, the statistical characteristics of each tap’s amplitude are obtained. It is found that the Nakagami distribution fits the statistical distribution of the model taps’ amplitude. The m parameter is larger than 1 for each tap. Not only the m parameter under the corridor and the stair is obviously larger than that under the office and the meeting room, but also the number of the taps under the stair and corridor environment is larger than that under the office and the meeting room environment. The results indicate that the wave guide effect under the corridor and stair environment with narrow structure is much more remarkable than that uner the wide office environment. Moreover, enough measured and modeled results show that the m parameter is linear dependent on the tap delay.Then, the property of the wall insertion loss under indoor environment at 3-6GHz is studied and a novel frequency-polarization-dependent wall insertion loss model is proposed. A linear, frequency-dependent polarization gain factor is introduced to describe the attenuation more precisely. The plenty measurement data is fitted by the minimum mean square error(MMSE) method. The results indicate that the wall insertion loss exponent is linear, frequency-dependent and increases with the frequency. Furthermore, the modeling results are analyzed for three different antennas’ polarization setup cases: V-V, V-H, H-H. It is found that the polarization gain factor is also linear-frequency dependent. The wall insertion loss for the V-V case is higher than that for the H-V and the H-H ones. Meanwhile, the attenuation in the the H-H polarization is more obvious than that in the H-V one.Finally, based on the singular value decomposition of the channel matrix, a novel stochastic multiple-input multiple-output(MIMO) channel model is proposed in this study. Under the framework of the proposed model, each of the right singular vectors can be modelled as the product of a stochastic scalar and a non-random vector, as is each of the left singular vectors. The non-random vectors, defined as the eigenmodes of the transmitter and receiver respectively, can be easily extracted from the measurements, so are the singular values of the channel matrix. The implications of the proposed models parameters that provide further insight into the MIMO channel are interpreted and a way of exploiting the parameters is given. To validate the proposed model, MIMO channel measurement is carried out under different indoor environments and the meaqsured channel capacity is analysed. It is shown that the proposed model provides a better fit to the measurement results than the other popular stochastic channel models.The proposed wireless propagation and channel models in this dissertation can be used in the link budget calculation, the indoor wireless coverage estimation/planning, the link level simulation and the interference alignment for the Femtocell communication system design and evaluation. |
