GPS Signal Degradation Analysis Using a Simulator

Transcription

1 GPS Signal Degradation Analysis Using a Simulator G. MacGougan, G. Lachapelle, M.E. Cannon, G. Jee Department of Geomatics Engineering, University of Calgary M. Vinnins, Defence Research Establishment Ottawa Author s Biographies Glenn MacGougan holds a BSc in Geomatics Engineering from the University of Calgary where he is currently pursuing a MSc. Gérard Lachapelle is Professor and Head of the Department of Geomatics Engineering at the University of Calgary, where he also holds a CRC/iCORE Chair in Wireless Location since early He has been involved with GPS R&D since Elizabeth Cannon is Professor and NSERC/Petro-Canada Chair for Women in Science and Engineering in the Department of Geomatics Engineering, the University of Calgary. She has been involved in GPS R&D since Gyu-In Jee is Professor in the Department of Electronics Engineering, Konkuk University, Korea, where he has been involved in GPS and navigation research for many years. He is currently on sabbatical leave at the University of Calgary. Michael Vinnins heads the Navigation Group at the Defence Research Establishment Ottawa, Department of National Defence. ABSTRACT Procedures to test selected receiver performance using a high fidelity GSS 4760 simulator are developed and applied to two receivers, namely a P-Code PLGR-96 and a NovAtel C/A Code OEM4 receivers. The latter is equipped with a Pulse Aperture Correlator to enhance performance under multipath. The characteristics tested and reported herein are signal power tracking threshold, signal variations, high dynamic effects and multipath effects. The performance measured are presented, compared and analysed. INTRODUCTION The objective of this paper is to analyse the following performance of the PLGR-96 and the NovAtel OEM4 receivers using a GPS simulator: - signal power tracking threshold - signal variations - high dynamic effects - multipath effects Performance will be assessed in term of the impact of the above on receiver derived position, velocity and C/N o. The multipath test also included a NovAtel OEM3, which uses a standard Narrow Correlator spacing technique. EQUIPMENT The PLGR-96 (Precision Lightweight GPS Receiver 96) is a 5-channel L1 receiver that operates in C/A code during acquisition and then switches to P/Y code. The unit tested switched to the P-Code as it did not contain a Y-Code encryption key. The receiver outputs position, velocity and C/N o data at a rate of 1 Hz. Raw data is not available, and the analysis was therefore performed in the position domain only. The maximum prescribed dynamics is 1,200 m/s in velocity and 9 g in acceleration. The time-tofirst-fix (TTFF) is 90 s, while the reacquisition Annual Meeting, The Institute of Navigation, Albuquerque, N.M., June 10-13,

2 time is 60 s. The single point positioning accuracy is given as 16 m SEP (Spherical Error Probable) but since it is a L1-only receiver, the accuracy will also depend on the level of ionospheric activity. The NovAtel OEM4 is a 24-channel L1/L2 C/A- Code receiver equipped with a Pulse Aperture Correlator to further improve performance under multipath conditions. Only the 12 L1 channels were used for the simulations reported herein. The maximum prescribed dynamics is 515 m/s in velocity and 10 g in acceleration. The TTFF is 60 s and reacquisition time, < 6 s. The simulator used was a GSS STR-4760 unit comprising of two synchronous hardware units, each with 16 L1 or 8 L1/L2 channels. The STR is capable of real-time and scriptable remote control and can reproduce a wide range of multipath, signal blockage and receiver dynamic scenarios. The signal level specifications are given in Table 1. Table 1: GSS STR-4760 Simulator Signal Level Specifications multipath errors were input. Position errors are therefore be mostly due to receiver noise, amplified by satellite geometry. TEST RESULTS Signal Tracking Threshold During the test, equal power on each channel was applied. The receivers were allowed to acquire the full navigation message during a 20- minute warm-up period. The signal power was decreased on all channels by 0.2 db per minute until no satellites were tracked, as shown in Figure 2. Figure 2: Signal Power Applied During Signal Tracking Threshold Test The tracking results for theplgr-96 and OEM4 are shown in Figure 3 and 4, respectively. The blue line shows the number of satellites theoretically available while the green and red lines indicate the numbers of satellites tracked by the OEM4 and PLGR-96, respectively. The two receivers lost all satellites after a 9 and 5 db signal drop, respectively. Figure 1: Parallel Receiver Testing with the GSS STR-4760 The STR-4760 has a high degree of fidelity. This was verified by repeating each test a second time to ensure repetitiveness. Shown in Figure 1is the STR-4760 during the parallel testing of the receivers. During the simulations, no orbital errors were input. Minimal atmospheric errors were assumed. Unless otherwise stated, no Figure 3: Tracking Threshold Test Result Annual Meeting, The Institute of Navigation, Albuquerque, N.M., June 10-13,

3 The C/N o results are shown in Figure 4. The tracking threshold corresponding to the loss of all satellites is -24 db-hz for the PLGR-96, and - 30 db-hz for the OEM4. An offset of 3 db between the two receivers C/N o, corresponding to the difference in signal P and C/A codes, was expected. Instead, an offset of 9 db is observed. This is likely due to the different methods used to measure the C/N o in each receiver. (Note that during the initial acquisition period, the PLGR-96 had difficulty in getting positions. This is due to a difficulty in reading the simulated almanacs, a phenomenon observed with other receivers when using the simulator). The OEM4 tracked 15 to 20 minutes longer than the PLGR-96, which corresponded to a relative signal strength change of 4 db. Figure 4: C/No Test Result What typically happens when the signal strength drops below an acceptable floor level is illustrated in Figure 5, in this case for PRN #16 with the OEM4 receiver. At a specific epoch all satellites are lost. Within minutes, the tracking loops somewhat regain tracking of the signal for brief intermittent periods and the signal is then lost for good. Not shown here is the fact that the measurements are deemed unusable for navigation during the above intermittent period. Figure 6: Position Errors as a Function of Signal Strength A signal strength variation test, with the signal strength changing randomly within +/-1Hz along the decreasing signal strength ramp, was also performed. The results are similar to those obtained with the linear signal strength test performed above. High Dynamic Tests These tests were performed to assess receiver behavior under up to 5 g s acceleration and 300 m/s constant velocity in the horizontal components. The test trajectory simulated is shown in Figure 7, and the acceleration and jerk, in Figure 8. Figure 5: Tracking Behavior at Signal Level Threshold The position errors resulting from the above simulation are shown in Figure 6. The position component accuracies are good until the signal strength drops below a certain level. They then degrade until the signals are deemed unusable and the receivers go into failed navigation mode. Figure 7: Test Trajectory Annual Meeting, The Institute of Navigation, Albuquerque, N.M., June 10-13,

4 The position errors that occurred during the high dynamics test are shown in Figure 10 and reach 12 m and 6 m with the PLGR-96 and OEM4 receivers, respectively. There is a direct correlation between position errors and accelerations, as shown in Figure 11. This phenomenon, which is a function of the receivers tracking loops, was observed previously (e.g. Cannon et al 1997, Hebert et al 1997,). Figure 10: Position Errors during High Dynamics Test Figure 8: Test Dynamics The measured Doppler shift with the OEM4 on PRN #3 during a portion of the test during which the dynamic variations were high is shown in Figure 9. A comparison with the reference values generated by the simulator shows that the Doppler shift variations measured by the receiver were correct. Figure 11: Correlation Between Position Errors and Accelerations Latitude Component Figure 9: Measured Doppler Shift OEM4 Likewise, the horizontal velocity errors during the high dynamics tests are shown in Figure 12. Errors of up to 18 m/s and 3 m/s occur with the PLGR-96 and the OEM4, respectively. This indicates the use of a higher order tracking loop in the OEM4. An analysis, not shown here, reveals a high correlation between velocity error and jerk in the case of the PLGR-96. During Annual Meeting, The Institute of Navigation, Albuquerque, N.M., June 10-13,

5 periods of constant velocity (100 m/s), the RMS error was found to be 2 cm/s and 6 mm/s for the PLGR-96 and the OEM4, respectively. Table 2: Static Multipath Parameters Figure 12: Velocity Errors during High Dynamics Test Static Multipath Test This multipath test included a 3 rd receiver, namely a NovAtel OEM3, equipped with a standard Narrow Correlator Spacing technology. The simulated multipath was a function of satellite elevation and azimuth and was induced 20 minutes after receiver warm-up. The multipath pattern files defining a ground reflector are given in Table 2. The signal attenuation varies between 46 and 18 db, while the corresponding multipath delay pattern ranges from 20 ns to 999 ns. A sample multipath signal generated for one satellite (PRN 17) is shown in Figure 13. A few satellite dropouts occurred as a result of the induced multipath. However this did not significantly affect the satellite geometry (DOP). The 2-D radial errors that resulted from the simulated multipath are shown in Figure 14. The effect on the P-code PLGR-96 is generally below the 2-m level. In the case of the Pulse Aperture Correlator C/A code OEM4 receiver, the effect is generally below the 5-m level. In the case of the Narrow Correlator spacing OEM3 however, the effect exceeds 8 metres at several epochs. The use of a standard wide Correlator C/A receiver would obviously have resulted in much larger effects. Annual Meeting, The Institute of Navigation, Albuquerque, N.M., June 10-13,

6 Figure 13: Sample Multipath Signal PRN 17 Figure 14: 2-D Radial Errors Resulting from Simulated Multipath CONCLUSIONS The use of simulations with a high fidelity simulator was shown to be very useful in testing receiver performance. The tests reported herein shown reasonably good performance for the PLGR-96. Many more tests are underway to further test the unit, namely multipath effects in kinematic mode, signal cross-correlation effects and the effect of evil waveforms. REFERENCES Cannon, M.E., G. Lachapelle, M. Szarmes, J. Hebert, J. Keith, and S. Jokerst (1997) DGPS Kinematic Carrier Phase Signal Simulation Analysis for Precise Velocity and Position Determination. Navigation, The Institute of Navigation, Alexandria, VA, 44, 2, Hebert, J., J. Keith, S. Ryan, M. Szarmes, G. Lachapelle, and M.E. Cannon (1997) GPS Carrier Phase Signal Simulation Analysis for Aircraft Velocity Determination. Proc. of 53rd Annual Meeting of The Institute of Navigation (Albuquereque, N.M., June 30 - July 2), Annual Meeting, The Institute of Navigation, Albuquerque, N.M., June 10-13,