DME/TACAN Interference and its Mitigation in L5/E5 Bands

The Galileo E5a/E5b signals and the GPS L5 signal lie within the Aeronautical Radionavigation Services (ARNS) band. They suffer interference from the services in this frequency band, in particular, high power pulsed signals from Distance Measuring Equipment (DME) and Tactical Air Navigation (TACAN) systems. The pulsed interference degrades received Signal to Interference and Noise Ratio (SINR), lowers the acquisition sensitivity and even causes the tracking loops to diverge. To maintain system accuracy and integrity, interference mitigation is beneficial and necessary. In this paper, the Stanford GNSS Monitor System (SGMS) is used to investigate the DME/TACAN signal environment at Stanford, CA, USA. The DME/TACAN beacons of six nearby airports, Woodside, SJC, SFO, Sausalito, OAK, Moffet, are observed. The TACAN signals are characterized by a 15 Hz sinusoidal envelope with north reference pulse code patterns and another 135 Hz modulation with reference pulse group patterns. Current DME/TACAN interference techniques can be categorized as time-domain approach and frequencydomain approach. ‘Pulse blanking’ is the time-domain method. It zeroes out the portion where the amplitude of the complex I/Q signal exceeds a certain threshold level related to the noise. Pulse blanking is simple to implement, can be executed in real time without extra delay and only functions when the interference exists. However, when blanking the interference pulses, it also zeroes out the signals over that time slot. If the pulses are extremely dense in time, all received signals including both DME/TACAN pulses and GNSS signals will be blanked. The tracking will fail due to the unavailability of the signal. Moreover, because of the Gaussian pulse tailing effect, pulse blanking cannot completely suppress the interference. ‘Notch filtering’ mitigates the pulse interference in the frequency domain, where the DME/TACAN signals appear as narrow-band frequency tones. If the signal spectral density at certain frequencies is above the noise spectral density, these frequency components will be filtered out. Notch filtering can thoroughly suppress the DME/TACAN interference, including the central part of the Gaussian pulse and the tails. It also preserves the energy of the signal superposed with the interference pulses in the time domain. However, it not only filters interference, but also removes the signal energy at the DME/TACAN frequencies. Even during the time period when there are no DME/TACAN pulses, the E5 signal at these frequencies is still suppressed. If there are multiple DME/TACAN transponders nearby, the filter design will be complicated due to multiple notches in the filter. ‘Hybrid blanking’ exploits the advantages of both pulse blanking and notch filtering. In the time domain, if an interference pulse is detected, it triggers the notch filtering of a slice of 12 μsec data centered at the estimated pulse position. Filtering is only implemented when DME/TACAN pulses exist. It overcomes the disadvantage of regular notch filtering, which always filters out the corresponding frequency components of the signal even when there is no interference. For the slices of data that are covered by DME/TACAN pulses, hybrid blanking preserves most of the signal energy, and thus overcomes the disadvantage of time-domain pulse blanking. The filter design is simple, as there is a high chance of pulses from one certain transponder within the 12 μsec time window, thus there is only one notch in the fitler To evaluate these three methods, signals from the GIOVE-A test satellite are collected. The interferencemitigated signals are acquired with a multi-signal all-inview GNSS software receiver. The correlation peak to next peak ratio (CPPR) and the correlation peak to mean peak ratio (CPPM) are chosen as the figures of merit for evaluation.