摘要
现有的实时GNSS软件接收机不能满足干扰监测型接收机的带宽需求。CPU计算能力不能满足宽带软件接收机的实时计算需求,提出了利用GPU加速接收机的实现方法。通过分析基带信号处理模块的各个子模块的计算量,对计算量大的模块用GPU进行合理的优化,提升了基带信号处理能力。将结果与CPU软件接收机的结果进行了对比,发现信号处理时间大幅度减少,满足了实时性的要求。此外,提出了GPU编程方面的一般性建议,有利于提高信号处理速度。
The exsiting real-time GNSS software receiver can not meet the requirement of bandwidth in interference monitoring application. The computing capacity of CPU cannot meet the real-time computation need of broadband GNSS software receiver,so the approach of acceleration in receiver using GPU is proposed in this paper. By analyzing the computation of every sub-module in baseband signal processing,it focuses on time-consuming parts,tries to optimize them reasonable and improve their realtime performance. Then it compares the results of CPU version with the results of GPU version,and finds the signal processing time reduces dramatically and finally meets the requirements of real-time processing. Besides,some advice on GPU programming are given so as to achieve better performance.
出处
《信息技术》
2017年第1期121-126,130,共7页
Information Technology
关键词
干扰监测
GNSS软件接收机
实时
基带信号处理
GPU
interference monitoring
GNSS software receiver
real-time
baseband signal processing
GPU
