Research Of GNSS Signal Tracking Algorithm Under Spinning Condition And Complemented By GPU

GNSS is widely used in military applications for its low cost and high precision of long term. Many vehicles, such as artillery shells and spacecraft, spin rapidly during at least a portion of their flight. Spinning will cause the Doppler of receiving signal changing periodically, and have a bad effect on the performance of receive, thus normal receiver with traditional loop parameter will not ensure robust work. So the study of PLL(phase lock loop) and FLL's(frequency lock loop) performance under spinning conditions, and design of advanced algorithm have fundamental significance. This thesis aims on the performance of tracking loop under spinning conditions, and the main work of this thesis includes the following contents.The amplitude, Doppler and phase of GNSS signal under spinning conditions will change largely. Based on physical laws and the change of angle of incidence to antenna, combining with antenna pattern, this thesis models the receiving GNSS signal under spinning conditions, and point out that the Doppler has a rule of sinusoid. Spinning has a bigger effect on carrier than code, so it doesn't need to consider spinning's effect on code tracking.The sine changed Doppler and phase leads to lose effectiveness of traditional method of analyzing tracking loop. Based on the high pass filter property of PLL error transfer function, this thesis builds the relationship of input phase amplitude and tracking error amplitude, come up with a method to analyze tracking loop under spinning conditions, and verify it, which proves its rationality adequately.By the above method, this thesis analysis the sensitivity of PLL and FLL and the biggest spinning rate of tracking loop, gives the best noise band of PLL by simulation, and point out that the noise band of FLL cannot be optimized. This thesis also compares different kind of carrier loop with each other, and points out that the performance of 2-order loop and 3-order is related to the spinning rate when using the same noise band. It's better to use 2-order PLL when the rate is low and to use 3-order PLL when the rate is high.The real-time performance is an important direction of research of GNSS software receiver. Recurring to the powerful parallel computation ability of GPU,this thesis designs the entire architecture of real-time software receiver based on CPU+GPU, and gives the realization detail of GNSS signal tracking. The reliability of this software receiver is validated by raw data from spinning weapons. The test of real-time indicate that when the sampling rate is 5MHz,tracking 12 satellites together,the average time of processing 1ms data is about 0.109 ms,which means it meets real-time requirements.