The Design And Implement Of GPS Software Receiver Based On GPU

This design of the GPU-based GPS software receiver departure from thetraditional hardware platform, so that the function of the software tomaximize.Processing the signal’s capture and track in the GPU,to process the GPSsignal acquisition and tracking’s large number of relevant aspects of computingthrough the GPU parallel processing.The results of the tracking loop import to CPU toimplement the positioning, implementthe final completion of the full functionality ofGPS software receiver.Compared to CPU-based software receiver which has a higherprocessing speed, real-time has been very good improvement, after which the study ofreal-time software receiver has a very important significance.This article is the use ofX5650(Nehalem)6-core CPU plus NVIDIA’s Tesla K20c graphics to achieve a rateof62M samples GPS IF digital signal’s capture,tracking and positioning solution.This paper briefly describes the significance of GPU-based,theoverview of thedevelopment of software receiver and the GPU-based software receiver, after theresearch status, complete the following three-part work.(1) The basic theory and CUDA platform software receiver has been studied onthe CPU to achieve a GPS positioning solution.Including GPS L1signal’s structureand generation method, the relevant characteristics of the pseudo-code,calculate theposition and velocity of the satellite, sub-frame synchronization, parity, calculate theuser’s location and speed, coordinate system conversion, etc.,using real data to thealgorithm simulation to verify the correctness of the algorithm,andimplement theCPU-based positioning using the C language.(2)Theresearch, simulation and GPU implementation GPS L1signal foracquisition and tracking. Includingthe signal’s sliding correlation acquisitionalgorithm, FFT algorithm, the basic phase-locked loop, the carrier tracking loop, codetracking loop, etc..Using Matlab simulation algorithm correctness, and according tothe algorithm characteristics, the paper is designed based parallel processing on GPUplatform, and using CUDA C to achieve.(3)According to the characteristics of GPS software receiver to study the CPU+GPU heterogeneous platforms operating mode and method. Finally,tthe time performance and positioning resultsof the system were tested.This design of the GPU-based software GPS receiver can greatly shorten thetime signal processing,do some help for the studies of the real-time processingsoftware receivers, provides a new platform for the research and development of newalgorithms for the satellite navigation receiver.