Research on GNSS Multipath Detection and Mitigation Methods for Low-Altitude Navigation

Abstract

In dense urban and complex low-altitude environments, GNSS signals are blocked, attenuated, and reflected, producing large numbers of multipath and non-line-of-sight (NLOS) gross errors, while low-altitude economy applications such as drone logistics and urban air mobility (UAM) demand navigation that is both highly accurate and highly reliable. Since the Required Navigation Performance (RNP) framework was introduced in civil aviation, integrity has ranked alongside accuracy, continuity, and availability as a fundamental requirement, and it remains the core challenge in extending satellite navigation to urban and low-altitude scenarios, where few satellites are directly visible, the affected satellites often exceed half of those in view, and the classical single-fault, Gaussian, and sparsity assumptions no longer hold.

This talk addresses the problem along three threads. For detection, an AE-LSTM representation-learning approach combined with deep unsupervised clustering achieves GNSS multipath detection under limited-labelling conditions. For mitigation, inertial-navigation kinematic constraints enhance the sparse structure and preserve positioning robustness even when more than half of the satellites are affected, with real-vehicle tests showing a significant reduction in 3D positioning error. For fusion, when satellites are completely lost, a dual-domain deep learning network predicts and compensates the inertial dead-reckoning to maintain continuous positioning during GNSS signal outages.

Speaker

Dr Xiaopeng Hou, PhD, is an Assistant Researcher at the School of Aviation of Beihang University (BUAA), currently conducting research at the State Key Laboratory of CNS/ATM. He received his BSc and PhD degree, all in Information and Communication Engineering, from the School of Electronic and Information Engineering, Beihang University. His research focuses on high-precision dynamic RTK positioning algorithms, pan-source resilient fusion navigation, and integrity monitoring for aviation Ground-Based Augmentation Systems (GBAS).

He has published more than 10 SCI/EI papers in domestic and international academic journals and holds 3 granted national invention patents. He has been deeply involved in several major national programmes, including the National Key R&D Program, NSFC Key/General/Youth grants, the BeiDou major special projects, and civil-aircraft special projects. Representing China's BeiDou, he has actively participated in the International Civil Aviation Organization (ICAO) drafting of Dual-Frequency Multi-Constellation (DFMC) GBAS Standards and Recommended Practices (SARPs), advancing the international standardisation of BeiDou.