Application of Bayesian Filtering for Multipath Mitigation in LDACS1-based APNT Applications

Kurzfassung

Civil avionic systems have been undergoing a major transformation during the past few years. To keep up with the growing demand for higher capacity and efficiency, the analog aeronautical communication systems have to be exchanged with a more efficient digital transmission. Moreover, the civil air traffic navigation systems need to be improved to provide more precise and reliable navigation information for the aircraft to improve the operation efficiency. The L-band digital aeronautical communication system – type 1 (LDACS1) is one of the most promising candidates for the future air traffic management data link. The LDACS1 signal has 500 kHz effective bandwidth and shall be allocated in the lower part of L-band (960 - 1164 MHz). One of the key advantages of LDACS1 is its spectral deployment: it is foreseen to allocate the LDACS1 channels in spectral gaps between two adjacent channels distance measuring equipment (DME) – a currently used radio navigation system. The civil aeronautics relies heavily on DME and VHF omnidirectional radio range (VOR). These legacy systems were established over several decades ago. They have limited performance that does not meet the future demands. Hence, the future navigation system is planned to rely on global navigation satellite systems (GNSS). The fields of GNSS navigation have been improved greatly during the last decades, and validations of flight procedures using GPS have already been done. However, integrity, continuity and availability of the navigation information are also important factors. These topics are major concerns for the use of GNSS navigation because of the large distance between the satellites and the aircrafts. The received signal can be interrupted by intentional or unintentional interference. Consequently, the GNSS failures might temporary occur. Therefore, there is a need for a parallel backup navigational infrastructure. This backup is commonly referred to as alternative positioning, navigation and timing (APNT). At the moment, several APNT solutions are being considered. One possible solution is to intensify the use of the DME system by increasing the density of the DME stations. However, this solution has some major disadvantages. First, the additional infrastructure might involve costly installation. Moreover, large portion of the L-band spectrum will be assigned to an old and spectrally inefficient technology. This gives difficulty for the frequency allocation of the mid- and long-term future communications in L-band. To support efficient spectrum usage, we propose to consider the future communication system LDACS1 as APNT solution. LDACS1 is a ground-based cellular communications system. Using ground stations (GS) as pseudolites, an airplane can exploit trilateration to estimate its position. The LDACS1 system is not optimized for ranging, yet it can be well exploited for navigation with only a few minor modifications. The deployment of LDACS1 GSs is planned for the future aeronautical communication systems and no additional infrastructure is needed for enabling navigational functionality. The LDACS1 system is also planned to be able to work in parallel with DME, and therefore it is possible to combine the range calculation from both systems in the situations where not a sufficient number of LDACS1 stations is visible. To assess the usability of the LDACS1 system as APNT solution, two measurement campaigns were conducted by DLR in November 2012 and 2013. The analysis of the measurement data gave the conclusion that LDACS1 signals offer an excellent ranging source. However, the results also indicate that multipath propagation can have a significant impact to the ranging performance, especially at low altitudes [1]. With the signal bandwidth of 500 kHz, the multipath in the range of few hundreds of meters are not easily resolvable. It is well-known that in such situation the ranging performance of the conventional correlation-based technique such as delayed lock loop (DLL) are largely biased. Unresolvable multipath environments cannot be easily avoided and would be a major problem for any low bandwidth ground-based APNT solutions (DME, LDACS1). Multipath mitigation is a very challenging task as the propagation channel is dynamic with the number of paths varies with time. A new algorithm that can perform robust range estimation under such dynamic multipath channels is needed to ensure good accuracy and integrity of the LDACS1-based APNT system. One of the considered algorithms for multipath mitigation in LDACS1-based APNT is particle filtering, which is a variant of sequential Bayesian filter implemented by Monte Carlo methods. For this algorithm, the underlying process model can be non-linear/non-Gaussian and is especially designed for dynamic channel conditions. The parameters estimation is based on a posterior density, and it uses a movement model to incorporate the temporal correlation of the change of the estimated parameters. It was demonstrated in [2] that this method can be efficiently used for the pseudo-range estimation of LOS component in dense and dynamic multipath environments in the GNSS receiver platform. The algorithm can correct the range error bias introduced by unresolvable multipath and give a much better ranging performance compared to the conventional DLL receiver. The fundamental drawback of the particle filtering is the complexity but it has been shown in [2] that the computation effort can be reduced greatly through the use of two-fold Bayesian filtering methods involving Kalman filter, Grid-based method and the particle filter. Here we propose to exploit this technique for LDACS1 ranging applications and demonstrate the algorithm performance on measured signals. The final paper starts with a brief description of the LDACS1 signal model, measurement campaign setup and the channel conditions. The detailed descriptions of the applied algorithm based on sequential Bayesian filtering and the application in the LDACS1 platform are then discussed. The analysis of the measurement data is focused on the dynamic multipath scenarios. Finally the obtained results and the future investigation directions are discussed.