Research On Channel Estimation Technology Based On Compressed Sensing

Compressed sensing is a compressive sampling and reconstruction program proposed for sparse signals. It has been widely used by digital signal processing, medical imaging and other fields for its low sampling rate and high accuracy of signal reconstruction. Among them, channel estimation combined with compressed sensing has draw great attention of scholars in the field of wireless communications.For wireless communication systems, receiver needs channel information for signal detection. Therefore, reliable channel estimation is very important. Pilot-based channel estimation uses the pilot information inserted in transmitted data to estimation wireless channel. There are several commonly used channel estimation methods, including Least Square(LS) estimation, Minimum Mean Square Error(MMSE) estimation and Wiener filtering estimation. After using these methods to estimate pilot channel, the whole channel can be obtained by adopting reasonable interpolation algorithm. In addition, iterative turbo channel estimation combines channel estimation and Turbo equalization organically. Receiver with this technique can achieve adequate information exchange between decoder and detector, then make the updated channel estimation more accurate through multiple iterations.Recently, among the transmit technology of wireless communication systems, Single-Carrier Frequency Domain Equalization( S C-FDE) and Orthogonal Frequency Division Multiplexing(OFDM) are becoming powerful competitor. SC-FDE has similar performance and complexity with OFDM, while overcoming the deficiencies of OFDM, such as sensitiveness of frequency deviation and high Peak-to-Average Power Ratio(PAPR). In future wireless mobile communication, SC-FDE has broad prospect of development. On the other hand, OFDM systems combined with Multiple Input Multiple Output(MIMO) is widely used because of its high-speed transmission rate. As a new type of MIMO transmit technology, Spatial Modulation(SM) is also a hot researching branch. SM chooses only one active transmit antenna to transmit the corresponding data per time slot, others transmit 0.This paper focuses on applying compressed sensing on channel estimation in wireless systems above, which specifically includes:Chapter I introduces several pilot-based channel estimation methods and the development of compressed sensing, then summarizes the main contents and structure arrangement of this paper.Chapter II researches the theory of compressed sensing from two aspects, respectively, the design of measurement matrix and reconstruction algorithm. In reconstruction algorithm, greedy algorithm is discussed in detail.Chapter III analyzes the multipath effect on the characteristics of the time-domain wireless channel. Then the system equation and block transmission structure of SC-FDE system is introduced. On the premise that the length of the guard period of piot block is regular, this chapter discusses the variation of system equation on pilot block and different channel estimation methods when the length of pilot block is decreased, and then proposes compressed sensing based channel estimation method for SC-FDE system and verify the reasonableness of the measurement matrix. Simulations and analysis of the method compared with several commonly used channel estimation methods are carried out at last.Chapter IV proposes the idea of compressed sensing based decision feedback channel estimation method for both OFDM and SM-OFDM system. In OFDM, the specific steps of the method is discussed. For different number condition of data symbols inserted between pilots, the estimation performance is verified. Then the model of SM-OFDM system is given, and the proposed method is modified to fit SM-OFDM system. The proposed method combined with Turbo equalizaiton technology is proved to obtain a good estimation performance and significantly reduce the number of pilot symbols of the system.Chapter V summarizes the whole work of this paper.