| Ultra-wideband (UWB) communications, with its high data rate, lower-power consumption, anti-multipath interference, simple structure, good security and many other advantages, has become a hot research in a short-range wireless communication technology. However, the extremely high bandwidth of received UWB signals requires high-speed analog-to-digital converters (ADC) in the UWB digital receiver, usually take up to 10GHz or more, it is difficult to achieve such high sampling rate in the current level of technology.The novel theory of compressed sensing (CS) offers an effective approach to solve the unachievable problem of high sampling rate. CS theory has show that a parse signal can be recovered, with high probability, from a set of random linear measurement using nonlinear reconstruction algorithms. Due to IR-UWB received signal with sparsity, CS theory is suitable for IR-UWB communication system.This paper focuses on the goal by casting the problem of UWB channel estimation and detection into the framework of CS, by introducing the basic principles of UWB communication systems, channel model and the compressed sensing theory.First, basing on the sparsity of the IR-UWB received signal in the time domain, the approach directly reconstructs the original received signal with CS theory. It shows that the feasibility of CS-UWB system, but the reconstruction performance is not efficient and the accuracy of the recovered IR-UWB signal is unacceptable with few random projections.Furthermore, UWB channel in general are rich in multipath diversity motivating the other method, multipath diversity. Since CS theory relies on the fact that the underlying signal is sparse, nonlinear measurement and signal reconstruction, it constructs a dictionary that closely matches the information-carrying pulse-shape and the nonlinear measurement matrix,chooses the orthogonal matching pursuit (OMP) as CS reconstruction algorithm. Then it has simulated the multipath diversity IR-UWB received signal reconstruction and channel estimation. The results show that CS multipath diversity approach requires fewer measurements with a high probability of successful reconstruction IR-UWB received signal, and the relative err of reconstruction is much smaller than the time sparse one. In signal detection, CS-UWB, in the medium SNR (15-25dB), is superior to the correlator-based detection, and a few samples can be accurately reconstructed of the original signal, reduces ADC resources.On this basis, this research has carried on the statistics the IR-UWB multipath channel characteristics, as the prior constraint to improve the CS reconstruction algorithm and its convergence. Compared the estimation performance between the improved algorithm and the OMP, the former is higher 3-4 dB than the latter one in the same mean squared error (MSE), and lower bit error rate (BER). |
