有限资源下车载北斗双频软件接收机基带信号实时处理算法研究

Satellite positioning receiver is essential for vehicle navigation system. Software receiver is easy to improve positioning performance by using various optimized algorithm compared with common hardware receiver, which also has obvious advantages in signal fusion with machine vision, radar and other sensors used for unmanned vehicles. However, large amounts of data is needed for complex computation at high speed, which makes current software receiver do real-time processing only by general CPU of desktop PC. Therefore, the practicality of software receiver is restricted in vehicle-mounted navigation system with limited resources. Now Beidou2 satellite navigation system has entered an era of multiple-frequency application. According to the signal structure of Beidou system, at least three algorithms are proposed and applied to baseband signal processing of in-car Beidou dual-frequency software receiver in this project, namely, the B1-aided fast acquisition strategy for B1/B2 dual-frequency signal based on Parallel Code-phase Search acquisition algorithm under assembled FFT, the pure-software fast parallel correlator technology of B1/B2 signal which is suitable for embedded platform and the high-accuracy tracking algorithm of B1/B2 signal based on double-loop coupling. The study on the above algorithms lays a theoretical foundation for the real-time processing of base-band signal in Beidou software receivers running with limited sources.

卫星定位接收机是车辆导航系统的必备装置。相比于常用的硬件接收机，卫星定位软件接收机便于利用各种优化算法提高其定位性能，同时在与用于智能车辆的机器视觉、雷达等其他传感器的信号融合方面也具有明显的优势。但由于需要对大量的数据进行快速复杂运算，这使得目前的软件接收机只能利用台式PC机的通用CPU进行实时处理，这成为软件接收机在资源有限的车载导航系统中走向实用化的瓶颈。目前我国北斗二代卫星定位系统已进入多频应用的时代。本项目针对北斗系统信号结构的特点，将基于组合FFT并行码相位搜索算法的B1辅助B2双频信号快速捕获策略、适合嵌入式实现的B1/B2信号纯软件快速并行相关器技术、基于双环耦合的北斗B1/B2信号高精度跟踪算法等，应用于车载北斗双频软件接收机基带信号处理中，为有限资源下的北斗双频软件接收基带信号实时处理奠定理论基础。

与硬件接收机相比，软件接收机更有利于无人驾驶车辆实现多传感器融合高精度导航。同一颗卫星播发的多频点卫星信号经由相同的路径传播到用户接收机，不同频点卫星信号中夹带的多普勒频移具有固定的比例关系，利用该比例关系可以提高双频软件接收机的捕获和跟踪性能。针对目前软件接收机中存在算法实时性差、硬件资源耗费较大等问题，本项目首先根据北斗B1/B2信号结构，设计了双频仿真信号发生器；其次提出了基于分解FFT的并行捕获算法，该算法相较于传统并行码捕获算法将计算量减少了80%；然后又提出了基于组合FFT并行码相位搜索的B1辅助B2双频信号快速捕获方法，相较于传统双频辅助捕获算法，该算法将计算量减少了约25%；最后提出了双频耦合跟踪算法，该算法可以使得B2信号在跟踪环路带宽减少的情况下，仍具有良好的动态特性。本项目以DSP+FPGA平台为基础，研究用于无人驾驶导航的车载北斗软件接收机，有利于北斗软件接收机在资源有限的嵌入式平台上实现，可为无人驾驶导航多传感器信息融合提供软件和硬件基础。