LTE系统中关键算法的研究
发布时间：2018-09-03 11:57
【摘要】：LTE系统是最有发展前景的数字通信系统之一,具有高数据速率、分组传送、延迟降低、广域覆盖和向下兼容的技术优势。目前LTE系统已经达到上行50Mbps,下行100Mbps的传输速率,这就要求接收机具备低时延的同时,还要能快速处理高速率数据。但是,传统的单核处理器的处理能力,已经不能满足LTE系统对高速率数据处理的要求。多核处理器具有高性能、低功耗、易于编程等诸多优势,研究基于多核处理结构的LTE并行化实现方案,已经成为亟待解决的问题。首先,本文研究了典型信号处理算法的并行化实现。针对复数矩阵求逆,本文采用了并行的方法来实现。整体使用流水线的结构方式,增加了模块的吞吐量,计算量是传统方法的三分之一,同时流水线的阶数是传统方法的二分之一。在时延和资源消耗,以及计算精度上,并行的方法相对于传统串行方法都与较大提升。并且,针对LTE系统中常见的DFT/IDFT模块,本文采用了并行化的互质因子算法(PFA)和Cooley-Tukey算法相结合的整体算法方法来实现,在一定程度上减少了操作量和资源消耗。其次,本文研究了LTE系统加扰算法的并行化实现技术通过回顾常用的并行加扰方法,在此基础上提出了一种新的并行加扰方案：基于稀疏矩阵运算的并行加扰方法,这种方案相对传统的加扰方法在运算量上具有显著的优势。该方法为在多核平台上实现任意位宽扰码的加扰提供了新的思路。最后,本文研究了低密度奇偶校验码(LDPC)的并行化编码技术。首先回顾了LDPC码的发展历程和现状,然后介绍了LDPC码检验矩阵的构造、编译码原理。针对常用的基于全下三角形式的编码算法,分析了该方法对于LDPC码检验矩阵构造的局限性。针对这一问题给出一种基于稀疏矩阵相乘的LDPC编码方式,建议算法具有高效、通用、受LDPC码检验矩阵构造约束小的优势。仿真结果表明,建议方法在运算时间上相对传统方法具有较大的优势。本文提出的并行化方案具有效率高、速度快、实现复杂度低的优势,一系列研究成果可广泛应用于各种高速无线通信系统中。丰富了下一代无线通信LTE系统的实现方案,为LTE系统收发机的并行化处理提供了参考。
[Abstract]:LTE system is one of the most promising digital communication systems. It has the advantages of high data rate, packet transmission, reduced delay, wide area coverage and downward compatibility. At present, the LTE system has reached the uplink 50Mbps.Downlink 100Mbps transmission rate, which requires the receiver to have low delay, but also to process high-speed data quickly. However, the processing capability of traditional single-core processors can not meet the requirements of high-rate data processing in LTE systems. Multi-core processor has many advantages, such as high performance, low power consumption, easy programming and so on. The research of LTE parallelization based on multi-core processing architecture has become an urgent problem. Firstly, the parallel implementation of typical signal processing algorithms is studied. In order to solve the inverse of complex matrix, the parallel method is used in this paper. With the overall pipeline structure, the throughput of the module is increased, and the computation amount is 1/3 of the traditional method, and the order of the pipeline is 1/2 of that of the traditional method. In terms of time delay, resource consumption, and computational accuracy, the parallel method is much better than the traditional serial method. Moreover, for the common DFT/IDFT module in LTE system, this paper uses the parallel mutual prime factor algorithm (PFA) and the Cooley-Tukey algorithm to realize the whole algorithm, which can reduce the operation amount and the resource consumption to a certain extent. Secondly, this paper studies the parallelization of scrambling algorithm in LTE system. Based on the review of common parallel scrambling methods, a new parallel scrambling scheme is proposed: the parallel scrambling method based on sparse matrix operation. Compared with the traditional scrambling method, this scheme has significant advantages in computation. This method provides a new idea for the scrambling of arbitrary bit width scrambling codes on multi-core platform. Finally, the parallel coding technology of low density parity check code (LDPC) is studied in this paper. This paper reviews the development and present situation of LDPC codes, and then introduces the construction of LDPC code verification matrix and the principle of encoding and decoding. Aiming at the common coding algorithms based on the full lower triangular form, the limitation of this method to the construction of LDPC code check matrix is analyzed. In order to solve this problem, a method of LDPC coding based on sparse matrix multiplication is proposed. It is suggested that the algorithm has the advantages of high efficiency, generality and small constraint of constructing LDPC code check matrix. The simulation results show that the proposed method has a great advantage over the traditional method in computing time. The proposed parallelization scheme has the advantages of high efficiency, high speed and low implementation complexity. A series of research results can be widely used in various high-speed wireless communication systems. It enriches the implementation scheme of next generation wireless communication LTE system and provides a reference for parallel processing of LTE transceiver.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.5
，
本文编号：2219910
[Abstract]:LTE system is one of the most promising digital communication systems. It has the advantages of high data rate, packet transmission, reduced delay, wide area coverage and downward compatibility. At present, the LTE system has reached the uplink 50Mbps.Downlink 100Mbps transmission rate, which requires the receiver to have low delay, but also to process high-speed data quickly. However, the processing capability of traditional single-core processors can not meet the requirements of high-rate data processing in LTE systems. Multi-core processor has many advantages, such as high performance, low power consumption, easy programming and so on. The research of LTE parallelization based on multi-core processing architecture has become an urgent problem. Firstly, the parallel implementation of typical signal processing algorithms is studied. In order to solve the inverse of complex matrix, the parallel method is used in this paper. With the overall pipeline structure, the throughput of the module is increased, and the computation amount is 1/3 of the traditional method, and the order of the pipeline is 1/2 of that of the traditional method. In terms of time delay, resource consumption, and computational accuracy, the parallel method is much better than the traditional serial method. Moreover, for the common DFT/IDFT module in LTE system, this paper uses the parallel mutual prime factor algorithm (PFA) and the Cooley-Tukey algorithm to realize the whole algorithm, which can reduce the operation amount and the resource consumption to a certain extent. Secondly, this paper studies the parallelization of scrambling algorithm in LTE system. Based on the review of common parallel scrambling methods, a new parallel scrambling scheme is proposed: the parallel scrambling method based on sparse matrix operation. Compared with the traditional scrambling method, this scheme has significant advantages in computation. This method provides a new idea for the scrambling of arbitrary bit width scrambling codes on multi-core platform. Finally, the parallel coding technology of low density parity check code (LDPC) is studied in this paper. This paper reviews the development and present situation of LDPC codes, and then introduces the construction of LDPC code verification matrix and the principle of encoding and decoding. Aiming at the common coding algorithms based on the full lower triangular form, the limitation of this method to the construction of LDPC code check matrix is analyzed. In order to solve this problem, a method of LDPC coding based on sparse matrix multiplication is proposed. It is suggested that the algorithm has the advantages of high efficiency, generality and small constraint of constructing LDPC code check matrix. The simulation results show that the proposed method has a great advantage over the traditional method in computing time. The proposed parallelization scheme has the advantages of high efficiency, high speed and low implementation complexity. A series of research results can be widely used in various high-speed wireless communication systems. It enriches the implementation scheme of next generation wireless communication LTE system and provides a reference for parallel processing of LTE transceiver.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.5
，
本文编号：2219910
本文链接：https://www.wllwen.com/kejilunwen/wltx/2219910.html
