Research Of Advanced Signal Processing Technologies And Its’ Applications In The Channel Emulator

With the improvement of demand for communication, wireless communi-cation equipment is more and more rich, and its update rate is becoming more and more fast. Wireless devices require a lot of testing to ensure that its per-formance meet the requirements before marketing them. In order to effectively shorten the wireless product development cycle and reduce the cost of test, to simulate real radio channel propagation environment in the laboratory, the chan-nel emulator emerges as required, and has a pivotal role in the entire product development process.However with the development of wireless communication technology, multiple input multiple output (MIMO) technology has been widely used, mak-ing the channel emulator is facing many challenges. First, the signal bandwidth becomes wider to carry more information, resulting the sampling rate based on the Nyquist theorem may beyond capacity of existing devices. Second, in order to obtain high delay resolution, the channel emulator is required to be able to simulate the minimal multipath delay interval under low clock frequency; Final, simulating complex MIMO channel model requires a large amount of compu-tation, but resources are limited, so how to reduce the simulation complexity to make the emulator to optimize performance under the limited hardware re-sources is important.In this paper, the multiple antenna broadband channel simulation technol-ogy in the new generation of wireless communication system is discussed, and several key technologies of broadband MIMO radio channel emulator are sys-tematically made in-depth research. The main works and innovations of this paper are as follows: (1) With the increase of signal bandwidth, sampling rate and processing speed required by the signal processing framework based on Nyquist theorem are becoming more and more high, even beyond the capacity of existing device. Analog-to-information converter (AIC) based on compressed sensing (CS) can sample the broadband analog signal at a very low rate. This paper analyzes and studies AIC to make it easier for hardware implementation. First, analyze the architecture of AIC and improve the measurement matrix, using cyclic matrix instead of Bernoulli random matrices and defining the generalized cyclic matrix as a measurement matrix, to reduce the system complexity. Second, the struc-ture of AIC is improves and a parallel multi-rate Compressed sampling (PMCS) architecture is proposed, where broadband signal is pre-modulated, integrated and sent into parallel analog to digital converters (ADCs) which have different sampling rates. After a series of pretreatment to the subsamples, the appropriate recover algorithm is used to reconstruct the original signal in high probability. We describe in detail the PMCS structure, analyse the effects of sample paths, sampling rate and other factors on system performance, and compare it with the parallel compressed sampling (PCS) architecture and parallel segmented com-pressed sampling (PSCS) architecture.(2) MIMO system performance is closely related to the wireless channel multipath environment. In order to better simulate the wireless channel with rich scatterers, the multipath delay resolution characterized by channel model becomes higher, so the channel emulator is required to achieve high delay reso-lution simulation, that is, can obtain arbitrary fractional sample delay of broad-band signal. For the exiting fractional sample delay methods, such as time do-main interpolation method and fractional delay filter method, we analyse these methods and summarizes the advantages and disadvantages of their own. How-ever these existing algorithms are without considering the inherent character of the signal, so a new method based on signal sparse is proposed to implement fractional sample delay. First, we analyse the sparse feature of broadband signal and construct a suitable sparse matrix. Then through theoretical derivation, the sample delay can be mapped into the delay and sampling of spare basis. Final, the feasibility of the algorithm is verified by simulation experiments, and the results are compared with existing methods. It is also pointed out that the frac-tional sample delay is no longer a difficult problem for compressed sampling.(3) In order to simulate the MIMO channel with high antenna dimension and multipath delay resolution on the hardware platform with the limited re-sources, the implementation of the wireless channel fading needs to be effec-tively designed and optimized. In this paper, we study the time domain tapped delay line model and analyze the various factors influencing the system com-plexity. Then we transform the time domain fading to frequency domain mod-ulation and compare the computational complexity of time domain method and frequency domain method. For the implementation of channel emulator in fre-quency domain, we propose an optimized implementation architecture of fast Fourier transform (FFT), which can be called as parallel pipeline architecture, where only one butterfly unit is used in each computing stage. Then we opti-mize the butterfly operation with adopting CORDIC algorithm to convert mul-tiplication to iterative addition, and analyse the complexity of time domain method, usual frequency domain method, and CORDIC based frequency do-main method.In this paper, the design and implement of a new generation of broadband MIMO channel emulator is studied. We make reaches on various technical problems and puts forward the corresponding solutions. These works provide research foundations for the research and development of broadband MIMO channel simulator.