One of the most used Position, Navigation and Timing (PNT) technology of the 21st century is Global Navigation Satellite Systems (GNSS). GNSS signals are affected by urban canyons that limit line-of-sight and reduce satellite availability to receivers. Smart cities are expected to adopt autonomous Unmanned Aerial Vehicles (UAV) operations for critical missions such as transportation of organs which are time-sensitive. Therefore, higher accuracy for position and velocity information is required. This paper investigates the use of Gated Recurrent Units (GRU) as a suitable technique that can memorize previous information in conjunction with the inputs (consisting of attitude, change in attitude, and change in velocity) to reduce position and velocity error when GNSS is not available. The fusion approach is developed and tested using Spirent's SimGEN GSS7000 hardware simulator which simulates GNSS signals and Spirent's SimSENSOR software that simulates accelerometer and gyroscope stochastic and deterministic errors. GNSS outage is varied between 1 and 20 seconds randomly to affect predicted position and velocity. The data is collected and used to train the GRU to predict the position and velocity error measured by the Inertial Measurement Unit (IMU). From the performance evaluation, a 60% reduction in Root Mean Squared Error (RMSE) is observed compared to Recurrent Neural Networks (RNN). Comparing 95 th percentile with Inertial Navigation System (INS), RNN, and GRU, an 80% reduction is observed between INS and RNN. Furthermore, a 35% drop in the 95 th percentile is observed between RNN and GRU.

One of the most used Position, Navigation, and Timing (PNT) technology of the 21st century is Global Navigation Satellite Systems (GNSS). GNSS signals are affected by urban canyons that limit Line-Of-Sight (LOS) and increase position ambiguity. Smart cities are expected to adopt autonomous Unmanned Aerial Vehicles (UAV) operations for critical missions such as the transportation of organs that are time-sensitive. Therefore, techniques to mitigate Non-Line-Of-Sight (NLOS) interference are required for improved positioning accuracy. This paper proposes a Gated Recurrent Unit-based (GRU) multipath detection algorithm that uses pseudorange, ephemerides, Doppler shift, Carrier-To-Noise Ratio (C/N0), and elevation data from each satellite to determine whether multipath is present. Signals from the satellite classified as multipath are then flagged and ignored for Position, Velocity, and Timing (PVT) calculations until they are deemed as LOS. The classification algorithm is developed and tested on Spirent GSS7000 to generate GNSS Radio Frequency (RF). OKTAL-SE Sim3D is used to simulate urban canyon environments in which signals propagate from the satellite to the receiver. RF signals are then transmitted to a Ublox F9P GNSS receiver that can receive GPS and GLONASS signals which are processed to output PVT information. The data collected is used to train the GRU to classify received signals as no multipath or multipath.From performance evaluation, GRU outperforms decision tree, K-Nearest Neighbor (KNN) classifiers, and Support Vector Machines (SVM). Furthermore, comparing GRU with SVM, a 50% increase in accuracy is observed with a 95% error of 0.85 m for GRU compared to 1.78 m for SVM.

As a result of the increasing usage of UAVs (Unmanned Air Vehicles) in urban environments for UAM (Urban Air Mobility) applications, the preciseness and reliability of PNT (Positioning, Navigation and Timing) systems have critical importance for mission safety and success. With its high accuracy and global coverage, GNSS (Global Navigation Satellite System) is the primary PNT source for UAM applications. However, GNSS is highly vulnerable to Non-Line-of-Sight (NLoS) blockages and multipath (MP) reflections, which are quite common, especially in urban areas. This study proposes a machine learning-based NLoS/MP detection and exclusion algorithm using GNSS observables to enhance position estimations at the receiver level. By using the ensemble machine learning algorithm with the proposed method, overall 93.2% NLoS/MP detection accuracy was obtained, and 29.8% accuracy enhancement was achieved by excluding these detected signals.1 Graduate student, School of Aerospace, Transport and Manufacturing (SATM) 2 PhD candidate, School of Aerospace, Transport and Manufacturing (SATM) 3 Senior Lecturer, Centre for Autonomous and Cyberphysical Systems

Urban air mobility is a growing market, which will bring new ways to travel and to deliver items covering urban and suburban areas, at relatively low altitudes. To guarantee a safe and robust navigation, Unmanned Aerial Vehicles should be able to overcome all the navigational constraints. The paper is analyzing a novel sensor fusion framework with the aim to obtain a stable flight in a degraded GNSS environment. The sensor fusion framework is combining data coming from a GNSS receiver, an IMU and an optical camera under a loosely coupled scheme. A Federated Filter approach is implemented with the integration of two GRUs blocks. The first GRU is used to increase the accuracy in time of the INS, giving as output a more reliable position that it is fused, with the position information coming from, the GNSS receiver, and the developed Visual Odometry algorithm. Further, a master GRU block is used to select the best position information. The data is collected using a hardware in the loop setup, using AirSim, Pixhawk and Spirent GSS7000 hardware. As validation, the framework is tested, on a virtual UAV, performing a delivery mission on Cranfield university campus. Results showed that the developed fusion framework, can be used for short GNSS outages.

Uncertainty-based sensor management for positioning is an essential component in safe drone operations inside urban environments with large urban valleys. These canyons significantly restrict the Line-Of-Sight signal conditions required for accurate positioning using Global Navigation Satellite Systems (GNSS). Therefore, sensor fusion solutions need to be in place which can take advantage of alternative Positioning, Navigation and Timing (PNT) sensors such as accelerometers or gyroscopes to complement GNSS information. Recent stateof-art research has focused on Machine Learning (ML) techniques such as Support Vector Machines (SVM) that utilize statistical learning to provide an output for a given input. However, understanding the uncertainty of these predictions made by Deep Learning (DL) models can help improve integrity of fusion systems. Therefore, there is a need for a DL model that can also provide uncertainty-related information as the output. This paper proposes a Bayesian-LSTM Neural Network (BLSTMNN) that is used to fuse GNSS and Inertial Measurement Unit (IMU) data. Furthermore, Protection Level (PL) is estimated based on the uncertainty distribution given by the system. To test the algorithm, Hardware-In-the-Loop (HIL) simulation has been performed, utilizing Spirent's GSS7000 simulator and OKTAL-SE Sim3D to simulate GNSS propagation in urban canyons. SimSENSOR is used to simulate the accelerometer and gyroscope. Results show that Bayesian-LSTM provides the best fusion performance compared to GNSS alone, and GNSS/IMU fusion using EKF and SVM. Furthermore, regarding uncertainty estimates, the proposed algorithm can estimate the positioning boundaries correctly, with an error rate of 0.4% and with an accuracy of 99.6%.