Channel Estimation Of UWB System Based On Compressed Sensing

Channel estimation technique is the core of the channel state information acquisition technology,according to a certain channel estimation criterion to obtain the impact of the system to the input signal transmission process,and achieve the goal of improving system performance further.And the biggest advantage of Compressed Sensing(CS) theory is the low rate of sampling(far less than the Nyquist rate) to realize signal reconstruction, and signal reconstruction is the key step in the process of channel estimation.The CS theory need to meet channel and signal sparse characteristics,and based on the validation of Ultra-wide Bandwidth(UWB) system channel meet the nature, we can use CS to complete the channel estimation process,so as to realize the excellent performance of channel system.First of all, this paper introduces the compressed sensing theory, and clarify it as the theoretical support of channel estimation, and analyzes the framework and its characteristics of UWB system, and validates the system is adaptive for CS theory, then describes the process of signal transmission and signal estimation in UWB channel model.Secondly, the paper puts forward a UWB systems channel estimation scheme based on CS.Aiming at the shortage of the traditional channel estimation scheme based on pilot, this article is based on compressed sensing estimation algorithm proposed an application of ultra-wideband systems channel estimation scheme, the simulation results show that the system can achieve better system performance and don’t have to pay more observation data, and also verifies that we can realize the recovery of signal with lower than the Nyquist sampling rate,and the simulation also analyzed the influence factors of CS recovery algorithm.Finally, the paper proposed a UWB systems channel estimation scheme based on SN-CS.Aiming at the structure and acquisition of the observation matrix in recovery algorithm,we proposed a new SN-CS(Compressed Sensing based on Sub-Nyquist) algorithm after learning existing gaussian random matrix and Bernoulli matrix as perception matrix.The new algorithm based on the broadband modulator architecture toimplement analog-to-digital conversion with a low sampling speed, and achieved observation matrix through hardware method directly by the corresponding fast Fourier transform coefficient to construct the sensor, It can more close to the channel characteristics.At the same time,this paper,applies the new algorithm of ultra-wideband system and puts forward four kinds of the actual sampling of the structure of the frame and then implement the simulation, and compares the scheme with Nyquist sampling recovery performance for the energy gain and the mean square error, the results show that the scheme designed of the structure of can achieve almost as well as the Nyquist recovery performance, but only needs far less than the Nyquist sampling points.