Simulated GNSS Signal Strength Variation Under Leaky Feeder Segments in a Tunnel—A Case Study

In this article, we studied rapid, centimeter-scale oscillations in received GNSS signal strength under a leaky feeder in a tunnel. GPS L1 signals were generated with a commercial GNSS simulator and injected into both ends of a radiating cable, after which the carrier-to-noise ratio was recorded with a mobile receiver moving incrementally along a rail. We developed a simplified, tunable propagation model that represents 1) signal attenuation and delay inside the leaky feeder, 2) radiation from multiple slot locations, and 3) reception directly below the cable, and ran it in two versions, with and without a term representing a wave reflected from the far end of the cable. Recorded and simulated signals were aligned and compared on three quantities: correlation, oscillation spacing, and oscillation amplitude. At 50 m from the injection point, the version with the reflected wave agreed with the measurement significantly better than the one without it (bandpassed correlation 0.88 against 0.63), in two independent sessions, reproducing the measured 10.3 cm spacing to within 0.1 cm and the amplitude to within about 20% with nothing adjusted. At the other locations the difference between the two versions was not significant, but none favored the version without the reflected wave. The results indicate that a standing wave can account for part of the observed oscillation and that a simple model reproduces the measured spacing and amplitude, supporting future work on GNSS-based positioning and augmentation in tunnels.