基于GPU的列车卫星定位接收机捕获引擎设计  

Design of acquisition engine for train satellite positioning receiver based on GPU

作　　者：黄颂 巴晓辉 蔡伯根[2,3,4] 姜维 王剑 HUANG Song;BA Xiaohui;CAI Baigen;JIANG Wei;WANG Jian(School of Electronics and Information Engineering,Beijing Jiaotong University,Beijing 100044,China;Frontiers Science Center for Smart High-speed Railway System,Beijing Jiaotong University,Beijing 100044,China;Beijing Engineering Research Center of EMC and GNSS Technology for Rail Transportation,Beijing Jiaotong University,Beijing 100044,China;School of Computer Science and Technology,Beijing Jiaotong University,Beijing 100044,China)

机构地区：[1]北京交通大学电子信息工程学院,北京100044 [2]北京交通大学智慧高铁系统前沿科学中心,北京100044 [3]北京市轨道交通电磁兼容与卫星导航工程技术研究中心,北京100044 [4]北京交通大学计算机与信息技术学院,北京100044

出　　处：《铁道科学与工程学报》2023年第8期3093-3104,共12页Journal of Railway Science and Engineering

基　　金：国家自然科学基金联合基金资助项目(U1934222);北京交通大学人才基金资助项目(2022XKRC003)。

摘　　要：下一代列控系统中提出了应用卫星导航提供位置服务的需求,相应的卫星导航硬件接收机价格高昂,不能兼具场景测试的灵活性和设备的便携性。传统软件接收机可根据需求更改基带算法,灵活性高,但由于串行捕获的原因导致其耗时较长,影响用户首次定位的时间。本文针对铁路沿线开阔和存在遮挡的强弱卫星信号场景,设计了基于GPU的双模捕获引擎,分析了信号捕获速度以及灵敏度。首先,介绍了基于FFT的并行码相位捕获算法;其次,利用GPU与CPU异构并行计算平台,设计了强弱信号的并行捕获架构;然后,对捕获的各个模块进行并行性设计,通过对载波多普勒、多颗星并行实现强星的搜索,对于弱信号捕获,提出一种间断性补偿码多普勒的全比特法来提高捕获速度。最后,使用现场轨道坐标测量数据仿真生成了不同载噪比的中频数据来验证并行捕获的正确性,并对捕获耗时、灵敏度进行了分析与对比。仿真结果表明:5 MHz采样率下,冷启动搜索12颗GPS/BDS强星信号只需32.5 ms;对25 dBHz的弱信号,GPS与BDS的捕获概率分别达到90%和80%,捕获时间减少了29.3%,在保证灵敏度的同提升了捕获速度。对2种铁路场景设计了相应的并行捕获架构,在铁路沿线捕获强弱卫星信号的同时能够降低列车首次定位时间,为列车专用软件接收机捕获引擎的设计提供了参考。In the next generation train control system,it has been proposed that satellite navigation be used to provide location services.The corresponding satellite navigation hardware receiver is expensive,which cannot balance the features of flexibility and portability.The traditional software receiver can change the baseband algorithm according to demand,providing high flexibility.However,it takes a long time due to serial acquisition,which affects the user's first positioning time.In this paper,a dual-mode acquisition engine based on GPU was designed for open and occluding scenes along the railway,which includes both strong and weak satellite signals.The signal acquisition speed and sensitivity were analyzed.First,a parallel code phase acquisition algorithm based on FFT was introduced.Second,a heterogeneous parallel computing platform consisting of GPUs and CPU,the parallel acquisition architecture for both strong and weak signals.Then,the parallel design of each module was carried out,and the search for a strong signal was realized through carrier doppler and multiple satellites in parallel.For weak signal acquisition,an intermittent compensation code doppler full-bit method was proposed to improve the acquisition speed.Finally,intermediate frequency data with different CNRs were generated by simulating field track coordinate measurements to verify the correctness of parallel acquisition.The time and sensitivity of acquisition were analyzed and compared.The results show that with a 5 MHz sampling rate,it only takes 32.5 ms to search for 12 strong GPS/BDS satellites in cold start.For the weak signal of 25 dBHz,the acquisition probability of GPS and BDS can reach 90%and 80%respectively,and the acquisition time is reduced by 29.3%,which improves the acquisition speed while ensuring the sensitivity.This paper designs the corresponding parallel acquisition architecture for the two railway scenarios,which can reduce the first positioning time of the train while acquiring the strong and weak satellite signal along the railw

关 键 词：列车定位 双模软件接收机 捕获 CUDA 并行计算 

分 类 号：U284[交通运输工程—交通信息工程及控制]