Research On The Synchronization Techniques Of High Sensitivity GNSS Receivers

With the development of location service based on GNSS (Global Navigation Satellite System), the users gradually move from the open outdoor space to the urban areas and indoor environments. In these conditions, GNSS signal power will be degraded badly by all the penetration loss. Traditional receiver techniques can not provide continuous and stable positioning. Thus, how to improve the performance of receivers in weak signal environments becomes a research hotspot during the progress of GNSS. This paper aims at improving the performance of GNSS baseband signal synchronization and carries out a deep research on the acquisition and tracking algorithms of high sensitivity GNSS signal in weak signal environments such as urban areas and indoors. The main innovation points and research contents are(1) Research on detection methods of signal parameters is performed on the basis of analysis of GNSS signal model and acquisition algorithms. The probability distributions of detection variables under threshold crossing detection, peak-to-mean ratio detection and peak-to-peak detection methods are derived. Detection performance of each method is analyzed from both theoretical derivation and simulation, including cell false alarm probability and cell detection probability. The drawbacks of peak-to-peak and peak-to-mean methods are summarized. Based on the research on searching strategies of parameter plane,all system detection probabilities, system false alarm probabilities and system miss detection probabilities of serial search, hybrid search and maximum search are derived. Theoretical and simulation results both prove that the maximum search method performs best in weak signal environments.(2) In order to eliminate the processing loss during acquisition due to the bit transition of navigation message, mimic circular operation is proposed based on the analysis of correlation splitting problem in detail. From the derivation of noise probability distribution in the calculation results, it is observed that the mimic circular operation will not reduce the detection probability while solving the problem of bit transition.(3) In order to shorten acquisition time and improve acquisition speed, fast Fourier transform is applied to mimic circular operation and parallel calculation is obtained.Combining the traditional time parallel acquisition, the modified time parallel acquisition with transition detection (TPATD) is proposed. Simulation results illustrate that the TPATD algorithm achieves higher sensitivity and more suitable to weak GNSS signal acquisition.(4) The key point to increase the receiver sensitivity is to get the required processing gain by extending signal integration time. This thesis analyzes coherent integration and non-coherent integration theoretically. The relationship between integration gain and square loss is derived. Also relationship between acquisition probability and detection threshold is obtained. Then combining design methods of coherent and non-coherent integration in high sensitivity acquisition are presented. Research on the high sensitivity acquisition methods of GPS L1 C/A signal, Galileo El signal and BDS B1 D1 signal is performed in the following sections.(5) Based on hypothesis of navigation message of Galileo El signal and combined with TPATD, TPATD and navigation data transition hypothesis based acquisition (TNTA) is proposed and acquisition sensitivity is improved by integration time extension. In order to reduce the calculation time, FFT and WH transform are applied to coherent integration on the basis of research on characteristics of m sequence and then parallel navigation data searching based coherent integration (PNSCI) is proposed. Compared to the typical energy averaging algorithm, simulation results prove higher sensitivity and reliability of proposed algorithms.(6) The relationship between navigation bit duration and PRN code period in GPS L1 C/A signal is analyzed in detail. Under the assumptions of navigation bit boundaries,navigation data transition hypothesis based full-bit coherent integration (NTFB) is proposed.The proposed algorithm can extend the coherent integration of GPS L1 C/A signal to 20ms.And simulation results show that the algorithm will not increase computation burden in comparison with conventional half bit method while assuring acquisition sensitivity.(7) The NH code characteristics in D1 signal of BDS B1 band are analyzed in detail. By definition of characteristic length and characteristic sequence, navigation data and NH subsequence hypothesis based acquisition (NNSA) is proposed. This method detects all the possible NH code subsequences in parallel by FFT/IFFT calculation after NH code sequence extension.(8) Research of design methods of GNSS tracking loop is performed. The algorithms and corresponding characteristics of loop discriminators are illustrated. The relationship between discriminating curves and pre-integration time is analyzed by simulation. On the basis of NCO models, the design method of optimal digital filter suitable to high sensitivity tracking loop is proposed. And the filter parameters are presented under phase acceleration inputs.Kalman filter is applied to carrier tracking loop and design method of high sensitivity loop filter based on EKF (Extended Kalman Filter) is derived. Simulation results prove that EKF performs best. But the computation burden makes it more suitable to pure software receiver.The optimal filter is more suitable to assisted tracking loops as it needs the prior knowledge of receiver dynamics.(9) On the basis of the previous research, acquisition and tracking of BDS B1 D1 signal under 34dB-Hz carrier to noise ratio are presented as example. Module design, Verilog programing and realization of acquisition and tracking stage are achieved on Xilinx V7 FPGA platform. Design processes of averaging sampling, correlation branch and coherent/noncoherent integration in acquisition are described in detail. Also design details of pre-integration, discriminators and loop filters in tracking loop are illustrated. The hardware resources are saved by techniques such as averaging sampling and time division multiplex of calculation resources. All the simulation results, FPGA platform test results and design summary report prove the validity and efficiency of the design.