In recent years, Multiple Input Multiple Output (MIMO) systems have gained importance due to the improvements on the performance of radio systems. Channel parameter estimation is an important factor in the design and optimization of MIMO systems. In this thesis, channel parameters such as delay, angles (azimuth and elevation) of arrival (AoA) and departure (AoD), Doppler frequency and polarization are estimated from measurement data using Space Alternating Generalized Expectation-Maximization (SAGE) algorithm. One of the focuses of this thesis is to reduce the computational complexity of the algorithm by using Particle Swarm Optimization (PSO) technique. Additionally, a new initialization procedure is proposed to get better estimates and to improve the processing time of the algorithm. Moreover, performance of the SAGE algorithm and improvements on the algorithm are tested via extensive measurement data. It is found that SAGE algorithm is a powerful tool for channel estimation and it can further be improved by the aforementioned propositions.