This thesis presents a design and implementation of a long-range communication subsystem for a UAV and a ground control station. The subsystem is a low-cost alternative employing a line of sight, local communication network for optimal communications between a low-altitude UAV and a portable ground control station. In this thesis, real world experiments are conducted to model the time-variant wireless channel between a low-altitude micro-UAV and a portable ground control station operating in an urban environment. The large-scale and small-scale fading coefficients are calculated and analyzed for this dynamic channel. The channel properties, along with the fading distribution parameters, are computed and analyzed for two most popular antenna configurations for UAV systems (Yagi to omnidirectional and omnidirectional to omnidirectional). For the Yagi-to-omnidirectional link, the effects of three major impacting factors i.e. propagation distance, antenna gains in specific spherical angles and polarization mismatch factor on the overall fading distribution is investigated. Through regression analysis, a multiple-regression model is derived that estimates the instantaneous fading parameter, given these channel conditions. For this model, a modified particle-swarm optimization algorithm is designed and implemented to estimate the underlying model coefficients, given the instantaneous fading information. The implementation of this algorithm, along with the regression model, demonstrates that a sufficient approximation of the fading parameter can be provided for any given wireless channel when the impacting factors and instantaneous fading information is available. / February 2016

The demands of multimedia services from mobile user equipment (UE) for achieving high data rate, high capacity and reliable communication in modern mobile wireless systems are continually ever-growing. As a consequence, several technologies, such as the Universal Mobile Telecommunications System (UMTS) and the 3rd Generation Partnership Project (3GPP), have been used to meet these challenges. However, due to the channel fading and the Doppler shifts caused by user mobility, a common problem in wireless systems, additional technologies are needed to combat multipath propagation fading and Doppler shifts. Time-variant channel estimation is one such crucial technique used to improve the performance of the modern wireless systems with Doppler spread and multipath spread. One of vital parts of the mobile wireless channel is channel estimation, which is a method used to significantly improve the performance of the system, especially for 4G and Long Term Evolution (LTE) systems. Channel estimation is done by estimating the time-varying channel frequency response for the OFDM symbols. Time-variant channel estimation using Discrete Prolate Spheroidal Sequences (DPSS) technique is a useful channel estimation technique in mobile wireless communication for accurately estimating transmitted information. The main advantage of DPSS or Slepian basis expansion is allowing more accurate representation of high mobility mobile wireless channels with low complexity. Systems such as the fourth generation cellular wireless standards (4G), which was recently introduced in Sweden and other countries together with the Long Term Evolution, can use channel estimation techniques for providing the high data rate in modern mobile wireless communication systems. The main goal of this thesis is to test the recently proposed method, time-variant channel estimation using Discrete Prolate Spheroidal Sequences (DPSS) to model the WINNER phase II channel model. The time-variant sub-carrier coefficients are expanded in terms of orthogonal DPS sequences, referred to as Slepian basis expansions. Both Slepian basis expansions and DPS sequences span the low-dimensional subspace of time-limited and band-limited sequences as Slepian showed. Testing is done by using just two system parameters, the maximum Doppler frequency Dmax v and K, the number of basis functions of length N = 256. The main focus of this thesis is to investigate the Power spectrum and channel gain caused by Doppler spread of the WINNER II channel model together with linear fitting of curves for both the Slepian and Fourier basis expansion models. In addition, it investigates the Mean Square Error (MSE) using the Least Squares (LS) method. The investigation was carried out by simulation in Matlab, which shows that the spectrum of the maximum velocity of the user in mobile wireless channel is upper bounded by the maximum normalized one-sided Doppler frequency. Matlab simulations support the values of the results. The value of maximum Doppler bandwidth vDmax  of the WINNER model is exactly the same value as DPS sequences. In addition to the Power spectrum of the WINNER model, the fitting of Slepian basis expansion performs better in the WINNER model than that of the Fourier basis expansion.