A Computationally Efficient Least Squares Channel Estimation Method for MIMO-OFDM Systems

The 5th generation of cellular system is expected to incur a huge traffic rise which would necessitate the adoption of an estimation method that is efficient but at the same time practical through easy implementation. Some of the most popular methods used in cellular communication for channel estimation are the Least Squares (LS) algorithm and the Minimum Mean Square Error (MMSE) algorithm. Both of them has their own merits and limitations. While LS estimation is simple to adopt and resource-friendly, its performance is par to MMSE, which requires channel statistics and is thus more impractical for the industry. In this paper, an efficient LS estimation method is proposed by minimising relative error or difference between each estimated channel coefficient from its actual value, which is often overlooked when considering overall error. It’s in turn, reduces the error per bit and eventually induces faster processing of data. Results on the proposed algorithm are demonstrated via bit error rate and mean square error comparison.