Accurate orbit prediction (OP) of space debris is vital in Space Situation Awareness (SSA) related tasks, like space collision warnings. However, due to the sparse and low precision observations, unknown geometrical and physical features of debris, and effects of incomplete force models, OP based on the orbital mechanics theory or physics-based OP of space debris suffers from rapid error growth over a long duration, limiting the period of validity of debris OP for precise space applications. Considering that the tracking arcs of a debris object over a single station often share a similar temporal and spatial distribution in the inertial space, the resultant OP errors possibly have a coherent relationship with the temporal and spatial distribution of tracking arcs. This paper proposes a machine learning (ML)-based approach to model the underlying pattern of debris OP errors from historical observations and apply it to modify the future physics-based OP results. The approach includes three steps: constructing a historical OP error set, training a ML model to fit the historical OP error set, and correcting the future physics-based OP with ML-predicted orbital errors. The ensemble learning algorithm of boosting tree (BT) is studied as the primary ML method for the error modeling and predicting process. Experiments with three low-Earth-orbit (LEO) objects, tracked by a single radar station, demonstrate that the trained ML models can capture more than 80% of the underlying pattern of the historical OP errors. More importantly, the errors of physics-based OP over the future 7 days reduce from thousands of meters to hundreds or even tens of meters through the error correction with the learned error pattern, achieving at least 50% accuracy improvement. Such dramatic OP improvements show the promising potential of ML for enhanced SSA capability.

Satellite remote sensing systems require accurate, autonomous and real-time orbit determinations (RTOD) for geo-referencing. Onboard Global Positioning System (GPS) has widely been used to undertake such tasks. In this paper, a novel RTOD method achieving decimeter precision using GPS carrier phases, required by China’s HY2A and ZY3 missions, is presented. A key to the algorithm success is the introduction of a new parameter, termed pseudo-ambiguity. This parameter combines the phase ambiguity, the orbit, and clock offset errors of the GPS broadcast ephemeris together to absorb a large part of the combined error. Based on the analysis of the characteristics of the orbit and clock offset errors, the pseudo-ambiguity can be modeled as a random walk, and estimated in an extended Kalman filter. Experiments of processing real data from HY2A and ZY3, simulating onboard operational scenarios of these two missions, are performed using the developed software SATODS. Results have demonstrated that the position and velocity accuracy (3D RMS) of 0.2–0.4 m and 0.2–0.4 mm/s, respectively, are achieved using dual-frequency carrier phases for HY2A, and slightly worse results for ZY3. These results show it is feasible to obtain orbit accuracy at decimeter level of 3–5 dm for position and 0.3–0.5 mm/s for velocity with this RTOD method.

Achieving continuous and high-precision positioning services via smartphone under a Global Navigation Satellite System (GNSS)-degraded environment is urgently demanded by the mass market. In 2018, Xiaomi launched the world’s first dual-frequency GNSS smartphone, Xiaomi Mi 8. The newly added L5/E5 signals are more precise and less prone to distortions from multipath reflections. This paper discusses the characteristics of raw dual-frequency GNSS observations from Xiaomi Mi 8 in urban environments; they are characterized by high pseudorange noise and frequent signal interruption. The traditional dual-frequency ionosphere-free combination is not suitable for Xiaomi Mi 8 raw GNSS data processing, since the noise of the combined measurements is much larger than the influence of the ionospheric delay. Therefore, in order to reasonably utilize the high precision carrier phase observations, a time differenced positioning filter is presented in this paper to deliver continuous and smooth navigation results in urban environments. The filter first estimates the inter-epoch position variation (IEPV) with time differenced uncombined L1/E1 and L5/E5 carrier phase observations and constructs the state equation with IEPV to accurately describe the user’s movement. Secondly, the observation equations are formed with uncombined L1/E1 and L5/E5 pseudorange observations. Then, kinematic experiments in open-sky and GNSS-degraded environments are carried out, and the proposed filter is assessed in terms of the positioning accuracy and solution availability. The result in an open-sky environment shows that, assisted with L5/E5 observations, the root mean square (RMS) of the stand-alone horizontal and vertical positioning errors are about 1.22 m and 1.94 m, respectively, with a 97.8% navigation availability. Encouragingly, even in a GNSS-degraded environment, smooth navigation services with accuracies of 1.61 m and 2.16 m in the horizontal and vertical directions are obtained by using multi-GNSS and L5/E5 observations.

Accurate vehicle ego-localization is key for autonomous vehicles to complete high-level navigation tasks. The state-of-the-art localization methods adopt visual and light detection and ranging (LiDAR) simultaneous localization and mapping (SLAM) to estimate the position of the vehicle. However, both of them may suffer from error accumulation due to long-term running without loop optimization or prior constraints. Actually, the vehicle cannot always return to the revisited location, which will cause errors to accumulate in Global Navigation Satellite System (GNSS)-challenged environments. To solve this problem, we proposed a novel localization method with prior dense visual point cloud map constraints generated by a stereo camera. Firstly, the semi-global-block-matching (SGBM) algorithm is adopted to estimate the visual point cloud of each frame and stereo visual odometry is used to provide the initial position for the current visual point cloud. Secondly, multiple filtering and adaptive prior map segmentation are performed on the prior dense visual point cloud map for fast matching and localization. Then, the current visual point cloud is matched with the candidate sub-map by normal distribution transformation (NDT). Finally, the matching result is used to update pose prediction based on the last frame for accurate localization. Comprehensive experiments were undertaken to validate the proposed method, showing that the root mean square errors (RMSEs) of translation and rotation are less than 5.59 m and 0.08°, respectively.

The continual miniaturization of mass-market sensors built in mobile intelligent terminals has inspired the development of accurate and continuous navigation solution for portable devices. With the release of Global Navigation Satellite System (GNSS) observations from the Android Nougat system, smartphones can provide pseudorange, Doppler, and carrier phase observations of GNSS. However, it is still a challenge to achieve the seamless positioning of consumer applications, especially in environments where GNSS signals suffer from a low signal-to-noise ratio and severe multipath. This paper introduces a dedicated android smartphone application called Walker that integrates the GNSS navigation solution and MEMS (micro-electromechanical systems) sensors to enable continuous and precise pedestrian navigation. Firstly, we introduce the generation of GNSS and MEMS observations, in addition to the architecture of Walker application. Then the core algorithm in Walker is given, including the time-differenced carrier phase improved GNSS single-point positioning and the integration of GNSS and Pedestrian Dead Reckoning (PDR). Finally, the Walker application is tested and the observations of GNSS and MEMS are assessed. The static experiment shows that, with GNSS observations, the RMS (root mean square) values of east, north, and up positioning error are 0.49 m, 0.37 m, and 1.01 m, respectively. Furthermore, the kinematic experiment verifies that the proposed method is capable of obtaining accuracy within 1–3 m for smooth and continuous navigation.