Research On Weak Signal Tracking In Satellite Navigation Systems

With the construction and operation of many new Global Navigation SatelliteSystem(GNSS), the applications of GNSS are increasingly popular in people’s dailylife. However, there are greater attenuation of the GNSS signals arriving at receiverin the complex environments such as urban canyons, indoors and forest, and most oftraditional receivers suffer degraded performance notably, even does not work. Thus,the traditional GNSS receivers can not meet the application requirements in these hashenvironments. Compared with the traditional receivers, high sensitivity receivers havehigher acquisition sensitivity and tracking sensitivity, it can be used in weak signal en-vironment. Nowadays, with the popularity of personal handheld navigation terminals,high sensitivity receivers have a broad market prospect. Therefore, the design of theGNSS receiver is facing the challenges and opportunities.The key techniques of high sensitivity receiver include weak signal acquisitiontechnique and tracking technique. Compared with foreign advanced countries, the de-velopment of high sensitivity receiver is still in its infancy and has a gap in our country.At present, most of the research of high sensitivity receiver is more focused on weaksignal acquisition algorithms than weak signal tracking algorithms. In order to im-prove the tracking sensitivity, the thesis investigated deeply the weak signal trackingtechnique in satellite navigation system. The main contents of the thesis are as follows:(1) Tracking sensitivity can be improved by increasing coherent integration time,so the effect of coherent integration time on the traditional digital tracking loop is stud-ied in this thesis. First, the thesis analyses traditional digital tracking loop’s stabilityand characteristic of closed-loop under the different normalized noise bandwidth indetail. With the increase of normalized noise bandwidth, the low-pass characteristicof traditional digital tracking loop will gradually deteriorate and result in instability ofthe loop finally. Then, the thesis analyses the deviation of the actual loop noise band-width from the desired noise bandwidth in different normalized noise bandwidth. Withthe increase of coherent integration time, the equivalent noise bandwidth of the loopis deviated from desired noise bandwidth gradually. The results of the analysis showsthat the traditional digital tracking loop is not suitable for weak signal tracking.(2) Due to the traditional digital tracking loop has limitation on coherent inte-gration time extending, the thesis designs an optimized digital tracking loop based onthe parameterized root method. What’s more, the thesis compares the closed-loopcharacteristic and tracking jitter between optimized digital tracking loop and tradition-al digital tracking loop under the different normalized noise bandwidth through theoryanalysis and simulation. The research results show that the optimized tracking loophas better low-pass characteristic and noise performance. Despite all this, the low-pass characteristic and noise performance of the optimized tracking loop will degradeby increasing coherent integration time.(3) The navigation data transition limits the extending of coherent integration timealso. In order to handle the data transition, the thesis provides analysis on three looptracking architectures such as the coherent architecture based on sign detection, the co-herent architecture based on energy detection and the non-coherent architecture. Thethesis verifies the availability of the three tracking architectures and compares the per-formance difference on tracking jitter between three tracking structure through sim-ulation and experiment. The research results show that three tracking architectureswithout any prior information of navigation data can effectively weaken the effect ofnavigation data. In terms of the performance of tracking jitter, the coherent architecturebased on energy detection is best and the non-coherent architecture is worst. However,in terms of designing complexity, the coherent architecture based on sign detection issimplest and the coherent architecture based on energy detection is most complex.(4) Due to the performance degradation for optimized digital tracking loop by in-creasing coherent integration time, so, in order to make the performance of the trackingloop better, the thesis designs an improved tracking loop in which a moving average fil-ter is introduced. Therefore, the thesis proposes an iterative algorithm which is basedon parameterized root method. The algorithm is available for determining the filtercoefficients of the tracking loop based on moving average filter. According to the lin-ear model of the tracking loop based on moving average filter, the thesis analyses theclosed-loop response of the loop, and verifies tracking jitter performance and avail-ability of the loop by simulation. The results of the analysis and simulation show thatthe noise and dynamic performance of the loop based on moving average filter is betterthan the optimized digital tracking loop.(5) Considering the output of phase discriminator and the output of frequencydiscriminator as measurement residual of Kalman filter respectively, the thesis de-signs third-order phase-locked loop and second-order frequency-locked loop whichare based on Kalman filter, and builds a specific linear model in z domain for third-order phase-locked loop and second-order frequency-locked loop through mathematicsderivation. By using the linear model in z domain, the thesis designs Kalman track-ing loop that is three-order phase-locked loop aided by second-order frequency-lockedloop, verifies the performance of the loop by simulation, and proposes the designmethod for Kalman tracking loop based on moving average filter according to Kalmanfilter and the characteristic the tracking loop based on moving average filter. The re-sults of simulation show the Kalman frequency-locked loop based on moving averagefilter can extend the pull-in range of the frequency discriminator, and the Kalman track-ing loop based on moving average filter can track the weak signal whose carrier to noisedensity ratio equals 10 d B-Hz when data bits don’t exist.On the other side,the Kalmantracking loop based on moving average filter can track the weak signal whose carrierto noise density ratio equals 15.5d B-Hz when data bits exist.