Channel Estimation Based On Superimposed Pilot And Compressive Sensing For MIMO-OFDM Systems

With the extensive growth of user requirements and the unceasing progress of communication technology, it’s significant to research in the fourth generation(4G) wireless communication systems, which can provide more stable, high-speed, ubiquitous and diverse communication service. As the key technique of4G, multiple input multiple output and orthogonal frequency division multiplexing (MIMO-OFDM) system can improve channel capacity and transmission rate, and enhance noise margin because of time, space and frequency diversity. In the MIMO-OFDM system, the high-performance is ensured by precoding, adaptive modulation, coherent demodulation, power controlling and time coding, and these technique requires accurate channel state information (CSI). So it is important to study the channel estimation technique to acquire the CSI quickly and reliably in the MIMO-OFDM system, which decides the quality of communication in large degree.The traditional way of channel estimation is divided into three kinds:blind channel estimation, pilots-aided channel estimation and superimposed pilot channel estimation. Blind estimation is realized by analyzing large amount of data, acquiring priori knowledge needed, and estimating the channel. It doesn’t need the pilots for saving the bandwidth and resource, but the algorithm is complex, convergent slowly. It performs badly in real time system. Pilots-aided estimation has been applied wider, and is realized by inserting the pilots to the to-be-sent. But the method occupies the bandwidth, and reduces the spectrum efficiency and the data rate. Therefore, in research of channel estimation, we pay much attention on how to reduce the number of pilots, on the premise that receiving terminal could obtain qualified CSI. In this respect, superimposed pilot in channel estimation has great advantage for not consuming extra sub-carriers or time slots. So the method stands out of the others in spectrum efficiency and data transmission rate.In MIMO-OFDM systems, because of delay spread, the wireless channel has multi-path effect. Then the wireless channel is sparse in time-domain naturally. Conventional methods of channel estimation don’t make full use of sparsity of the channel, so it’s wasting of resources. In recent years, Candes and Donoho introduced the principle and methodology of compressive sensing (CS), a new method of data acquisition, which breaks through the limit of Nyquist sampling rate. Many researchers exploit CS into wireless channel estimation.In this paper, we propose a new method of combining CS with superimposed pilot, for efficient channel estimation to improve spectrum efficiency and data transmission rate further more. The quality of the superimposed pilot is decided by the interference between the data and the pilots, moreover the high accuracy of compressive sensing can reduce interference and shorten the length of pilots. On the basis of the superimposed pilot algorithm, use the reconstruction algorithm of compressive sensing to acquire a rough estimate of channel, which replaces the related arithmetic. We analyze the model of the OFDM system and MIMO-OFDM system. Through theoretic analyses and computer simulations, we conclude that the mean square error (MSE) and symbol error rate (SER) performance of channel estimation has been improved obviously with the combined method. At the expense of a bit of computational complexity, the new method uses fewer pilots to obtain more accurate result, and then enhances the spectrum efficiency and data transmission rate.