【作者】 王萍；

【导师】 宋荣方；

【作者基本信息】 南京邮电大学 ， 通信与信息系统， 2012， 硕士

【摘要】 利用软件无线电来实现无线通信系统是未来发展的重要方向,而虚拟无线电是软件无线电发展的最新趋势,它以通用计算机的计算能力代替专用器件来完成数字信号处理过程。本文研究的虚拟无线电WiMAX系统是实现绿色宽带无线通信的重要基础探索。用通用计算机实现通信系统的瓶颈是CPU信号处理的实时性,因此需要对一些算法复杂、耗时较大的模块进行优化。WiMAX系统信道编码采用卷积Turbo码,相比于Turbo码具有编码效率高、性能好、译码延时小等优点。但在译码过程中采用的Max_Log_MAP算法较为复杂,限制了译码速率的提高。本文主要研究的是基于虚拟无线电的WiMAX系统信道编译码技术,首先分析了卷积Turbo码的编译码原理及分量译码器算法,并对不同序列长度、不同调制方式、不同码率以及不同迭代次数下的性能进行了仿真。然后对衰落信道下WiMAX系统上行链路卷积Turbo码译码迭代次数进行了优化,在系统性能可以接受的情况下提出了一种联合自适应的优化方案,使得平均迭代次数远远低于固定的最大迭代次数,从而提高了译码速率。最后对基于虚拟无线电的Max_Log_MAP算法进行了并行优化,在对译码模块进行定点化后,采用CPU内核指令SIMD(单指令多数据流)并行化处理,最终将译码速率提高了约3.05倍,达到了优化的目的。更多还原

【Abstract】 The use of software radio for wireless communication systems is an important direction for future development. Virtual radio is the latest trend in software radio development, which is based on general computer instead of application-specific devices to realize digital signal processing. The virtual radio WiMAX system studied in this thesis is an important foundation to explore the green broadband wireless communications. The bottleneck of communication system based on general computer is the real-time performance of CPU signal processing. Therefore, complex algorithm and time-consuming module need to be optimized. The WiMAX system channel coding uses Convolution Turbo Codes, which has high coding efficiency, good performance and low decoding delay compared with Turbo Codes. But the Max_Log_MAP algorithm is relatively complex in the decoding process, which limits the improvement of decoding rate.This thesis mainly studies the WiMAX system channel codec technology based on virtual radio. First, the codec principles and component decoder algorithm of Convolutional Turbo Codes are analyzed, and the performance on different sequence length, different modulation modes, different coding rate and different iteration number is simulated. Then the decoding iteration number is optimized on the uplink of WiMAX system in fading channel. This thesis presents a combined adaptive optimization scheme under the conditions that system performance is acceptable, making the average iteration number far below the fixed maximum iteration number, thereby improving the decoding rate. Finally, parallel optimization is carried out for the Max_Log_MAP algorithm based on virtual radio. After the fixed-point quantization, CPU core instruction SIMD(Single Instruction Multiple Data) is used to do parallel processing, that eventually makes decoding rate increase by about 3.05 times, achieving the purpose of optimization.更多还原