Accurate and reliable localization is of great importance for autonomous vehicles (AV). Mainstream localization approaches in autonomous vehicles (AV) are limited by the reliability of onboard sensors, which could be vulnerable to sensor failure, such as signal outages of the camera and signal spoofing of the global navigation satellite systems (GNSS). Different from these active or passive sensors, the vehicle dynamic model (VDM), which is the application of physical laws to a vehicle in motion, is environmentally independent and is capable of providing vehicle motion estimation continuously. However, the performance of the VDM-based motion estimation is dominated by the accuracy of the system dynamics model. To tackle this issue, this study proposes a sensor-free localization method VDM-SI by integrating system identification into the design of vehicle dynamic models (VDM). A system identification process based on low-order process models is proposed to identify the system dynamics of the AV, where the identified system responses are taken as the control input of VDM to estimate the vehicular positioning. The localization experiments in two scenarios show that the mean absolute translation error of VDM-SI can be reduced by 70% compared to conventional VDM methods. In addition, VDM-SI is experimentally proven to improve the localization performance of sensor fusion-based localization systems with high noise levels. Furthermore, in the application of re-localization after sensors fail and recover, VDM-SI shows strength in enhancing the security of AVs in extreme conditions.

A sharp yet conservative overbound for heavy-tailed error distributions is essential in integrity monitoring applications due to availability and continuity constraints. This paper proposes the Principal Gaussian overbound (PGO) for heavy-tailed error distributions by leveraging the characteristics of the Gaussian mixture model. The overbounding property of the PGO is proved to be preserved through convolution, which makes it possible to derive pseudorange-level requirements from the position domain integrity requirements. Experimental results on two datasets show that the PGO provides the most competitive bounding performance for heavy-tailed differential global navigation satellite system (DGNSS) pseudorange errors when compared to the two-step Gaussian overbound and Gaussian-Pareto overbound, yielding a sharp bound in both the core and tail parts of the error distribution. The proposed method reduces the mean vertical protection level (VPL) by more than 78% compared to the two-step Gaussian overbounding method on the urban dataset. In addition, the mean computation time of VPL is only 0.08 _S with fifteen measurements by employing fast Fourier transforms, suggesting the substantial potential of the PGO in GNSS applications with strict integrity and real-time requirements. Furthermore, the feasibility of the PGO in fault detection is discussed.

Global navigation satellite system (GNSS) positioning is essential for achieving absolute vehicular positioning in urban scenarios; however, it suffers from limited measurement redundancy and substantial faults caused by complex urban environments. In this work, we propose the subspace-based adaptive error modeling and fault detection and exclusion (FDE) method for pseudorange-based GNSS positioning in urban canyons, which integrates the adaptive error modeling into the FDE process and the positioning-solving process. Notably, we divide the pseudorange measurement space into subspaces regarding elevation angle and carrier-to-noise ratio (C/N0), each of which maintains a Gaussian mixture model (GMM) to adaptively characterize measurement error profiles. Results show that the proposed method has the ability to detect environmental changes. In addition, the proposed method outperforms the conventional FDE method with Gaussian assumptions, reducing the mean positioning error by 16 % and 9% in slightly and medium urbanized datasets, respectively. The impacts of step size (elevation angle and C/N0) and time window of the proposed method are discussed through controlled experiments.

In the low-altitude economy, ensuring the safe and agile flight of unmanned aerial vehicles (UAVs) in dynamic obstacle environments is essential for expanding interactive applications like parcel delivery. While deep reinforcement learning (DRL) shows promise for UAV motion planning and control, its trial-and-error exploration often struggles to ensure both agility and safety, especially under uncertain observational noise. Therefore, this paper proposes a deep confidence-enhanced reachability policy optimization (DCRPO) framework. By integrating safe DRL with nonlinear model predictive control (NMPC), DCRPO achieves high-level safety decisions, complex real-time joint planning and control for UAVs. Furthermore, we develop a deep confidence-enhanced reachability guarantee that constructs a set of stochastically forward-reachable planned trajectories under uncertainty, enabling robust safety collision probability certifications. This safe reachability mechanism adaptively selects belief space actions from planned actions to interact with the environment, further enhancing safety and reducing training time. In extensive experiments of UAVs traversing a fast-moving rectangular gate, the proposed method outperforms other state-of-the-art baseline methods under varying environments in terms of operational robustness. Furthermore, the proposed method significantly reduces overall collision violations and training time, greatly improving both training safety and efficiency. The demonstration video (https://youtu.be/7xkp9U7FSJg) and the source code (https://github.com/ZyyFLY/DCRPO) are also provided.

Fast and accurate fault detection and isolation for multiple faults is crucial for satellite navigation systems. However, conventional deletion-based greedy search methods suffer from swamping effects, i.e., wrongly excluding healthy measurements, which leads to degradation in positioning performance after executing the isolation. This study proposes an incrementally expanding algorithm to isolatemultiple faulty measurements in the multi-constellation global navigation satellite system positioning. The proposed algorithm is designed to find the most consistent set by incrementally expanding the minimum basic set with fault-free assumption. In each iteration, the no-fault hypothesis testing is conducted on the ordered studentized and jackknife residuals, enabling the correction of the fault-free assumption made in constructing the minimum basic set. The isolation performance and its impacts on positioning accuracy are evaluated in a worldwide simulation. The proposed method shows a 26% reduction in the swamping event rate and a 75% reduction in the mean post-isolation positioning error, compared to the deletion-based greedy search method. Through Monte Carlo simulations, the stability of the proposed method regarding the number of faults and the fault magnitude is demonstrated. An application to the real-world dataset with artificially injected bias is also studied, showing a reduced post-isolation positioning error.