摘要
针对目前基于CPU架构的软件GNSS中频信号模拟器计算能力有限,不能实时生成多通道数、高采样率信号的问题,在分析了软件模拟器的数学模型之后,提出了一种利用图形处理单元(GPU)架构实现的改进方案,算法建立在信号分段并行计算的基础上,有效将GPU中大量的计算单元用于信号合成过程,可实现多种导航信号的仿真。测试结果验证了改进后的信号模拟器的可用性与实时性,可以为GNSS软件接收机的算法验证和导航系统信号设计提供更好的测试评估支持。
The calculating speed of traditional GNSS IF signal simulator based on CPU architecture is limited, it can not generate multi-channel, high sampling rate signals in real time. To solve this problem, this paper proposed an improved implementation base on graphics processing unit （GPU） architecture. The parallel algorithm is based on signal segmentation, which can make effective use of the large number of computing units in GPU for signal synthesis, and it can simulate a variety of navigation signals. The test results verify the availability and real-time performance of improved signal simulator, and it can provide a good test platform for the receiver algorithms and new navigation signals design.
出处
《计算机仿真》
CSCD
北大核心
2013年第1期120-123,249,共5页
Computer Simulation
关键词
全球导航卫星系统
信号仿真
统一计算架构
并行计算
Global navigation satellite systems（GNSS）
Signal simulation
CUDA
Parallel computing
