The quality of Automatic Identification System (AIS) data is of fundamental importance for maritime situational awareness and navigation risk assessment. To improve operational efficiency, a deep learning method based on Bi-directional Long Short-Term Memory Recurrent Neural Networks (BLSTM-RNNs) is proposed and applied in AIS trajectory data restoration. Case studies have been conducted in two distinct reaches of the Yangtze River and the capability of the proposed method has been evaluated. Comparisons have been made between the BLSTM-RNNs-based method and the linear method and classic Artificial Neural Networks. Satisfactory results have been obtained by all methods in straight waterways while the BLSTM-RNNs-based method is superior in meandering waterways. Owing to the bi-directional prediction nature of the proposed method, ship trajectory restoration is favourable for complicated geometry and multiple missing points cases. The residual error of the proposed model is computed through Euclidean distance which decreases to an order of 10 m. It is considered that the present study could provide an alternative method for improving AIS data quality, thus ensuring its completeness and reliability.

This paper reviews currently existing fault-tolerant navigation system architectures and data fusion methods used in the design and development of integrated aircraft navigation systems and also compares their advantages and disadvantages. Four fault-tolerant navigation system architectures are reviewed and the associated Kalman filter architectures and algorithms are discussed. These techniques have been used in most integrated aircraft navigation systems. The aim of this review paper is to provide a guide for navigation system designers to develop future aircraft multisensor navigation systems.

The paper presents the continuation of the author's research on Evolutionary Sets of Safe Ship Trajectories (ESoSST) methodology. In an earlier paper (Szlapczynski, 2011) the author described the foundations of this methodology, which used Evolutionary Algorithms (EA) to search for an optimal set of safe trajectories for all the ships involved in an encounter. The methodology was originally designed for open waters or restricted waters when only the standard Convention on the International Regulations for Preventing Collisions at Sea (COLREGS, 1972) rules apply. However, within Traffic Separation Schemes (TSS), where additionally Rule 10 of COLREGS applies, the problem is much more complex and a new solution is needed. This paper introduces the extended ESoSST methodology, with a focus on changes that have to be made to obey Rule 10 and fully support TSS. These changes include detecting and penalizing TSS violations, as well as the pre-processing phase (generating the initial population, which includes predefined TSS-compliant tracks). The methodology has been designed for possible application in Vessel Traffic Service (VTS) centres. Its new mechanisms are presented with details. The examples are included of the results of the computer simulation tests carried out for the Gulf of Gdansk TSS to illustrate the methodology's effectiveness and functional scope.

The Automatic Identification System (AIS) is an efficient tool to exchange positioning data among participating naval units and land control centres. It was developed primarily as an advanced tool for assistance to sailors during navigation and for the safety of the life at sea. Maritime security has become a major concern for all coastal nations, especially after September 11, 2001. The fundamental requirement is maritime domain awareness via identification, tracking and monitoring of vessels within their waters and this is exactly what an AIS could bring. This paper will be focused on how the AIS-derived information could be used for coastal security, maritime traffic management, vessel tracking and monitoring with the help of GIS technology. The AIS data used in this paper was collected by the Canadian national aerial surveillance program.

As Earth-based radio tracking navigation is severely limited because of communications constraints and low relative navigation accuracy, autonomous optical navigation capabilities are essential for both robotic and manned deep-space exploration missions. Image processing is considered one of the key technologies for autonomous optical navigation to extract high-precision navigation observables from a raw image. New image processing algorithms for deep-space autonomous optical navigation are developed in this paper. First, multiple image pre-processing and the Canny edge detection algorithm are adopted to identify the edges of target celestial bodies and simultaneously remove the potential false edges. Secondly, two new limb profile fitting algorithms are proposed based on the Least Squares method and the Levenberg-Marquardt algorithm, respectively, with the assumption that the perspective projection of a target celestial body on the image plane will form an ellipse. Next, the line-of-sight (LOS) vector from the spacecraft to the centroid of the observed object is obtained. This is taken as the navigation measurement observable and input to the navigation filter algorithm. Finally, the image processing algorithms developed in this paper are validated using both synthetic simulated images and real flight images from the MESSENGER mission.

The integration of GPS and INS observations has been extensively investigated in recent years. Current systems are commonly based on the integration of INS data and the double differenced GPS measurements from two GPS receivers in which one is used as a reference receiver set up at a precisely surveyed control point and another is as the rover receiver whose position is to be determined. The requirement of a base receiver is to eliminate the significant GPS measurement errors related to GPS satellites, signal transmission and GPS receivers by double differencing measurements from the two receivers. With the advent of precise satellite orbit and clock products, the un-differenced GPS measurements from a single GPS receiver can be applied to output accurate position solutions at centimetre level using a positioning technology known as precise point positioning (PPP). This then opens an opportunity for the integration of un-differenced GPS measurements with INS for precise position and attitude determination. In this paper, a tightly coupled un-differenced GPS/INS system will be developed and described. The mathematical models for both INS and un-differenced GPS measurements will be introduced. The methods for mitigating GPS measurement errors will also be presented. A field test has been conducted and the results indicate that the integration of un-differenced GPS and INS observations can provide position and velocity solutions comparable with current double difference GPS/INS integration systems.

This paper comprehensively investigates the fault exclusion problem in multi-constellation Global Navigation Satellite Systems (GNSS). In future GNSS, the heightened likelihood of fault detection events will cause more interruptions in the continuity of the navigation operation. The main contribution of this paper is to establish the theoretical basis to quantify the contributions of fault events on continuity risk, therefore allowing us to assess the desired exclusion function performance based on specific continuity requirements. Accordingly, a new real-time exclusion algorithm is developed, for which the upper bounds on integrity risks are rigorously derived. Using the new method, performance is comprehensively investigated for two important civil aircraft navigation operations using various numbers of constellations. We show that high service availability can be achieved for both operations.

The paper, an edited and shortened version of the original, discusses a new approach to the display of collision threat parameter areas, obtained as a result of theoretical considerations. It can be applied to radar-computer anti-collision systems as well as manual radar plots.

With the development of digital techniques, and electronic digital computers in particular, radar-computer anti-collision systems are now aboard ships. Digital processing of signals and data in such systems can be divided into three stages as follows:

(a) Primary processing, which includes an analysis of radar signals within one rotation of the radar antenna. A result of this processing would be the position coordinates of detected objects. For relatively simple algorithms and high speed of analysis, special devices are usually employed.

(b) A secondary processing, which includes a correlation of the information obtained in the primary processing stage; the objects' position coordinates in consecutive antenna rotations; and an estimate of position, motion and approach parameters. For secondary processing devices general-purpose mini- and microcomputers are usually applied.

(c) A graphical display of the calculation.

As part of an investigation into the difference of absolute sea-level across the English Channel, a detailed series of measurements of currents at all depths was recently made in the Strait of Dover. The main object of the measurements was to relate the residual non-tidal flow of water to the voltage on a cross-channel cable generated by electromagnetic effects. Over 6000 readings of current strength and direction were made during ideally calm conditions. This paper describes the results of analysis of these measurements, which covers vertical current profiles and tidal components, besides the non-tidal flow. Distinction is made between mean current measured at a fixed depth or at the surface, and the mean volume flux of water which involves tidal variation of depth. A method of tidal analysis based on Fourier sums, which is convenient for use with an electronic computer, is outlined in the Appendix.

For multi-Global Navigation Satellite System (GNSS) constellations, the Geometric Dilution of Precision (GDOP) is an important parameter utilised for the selection of satellites. This paper has derived new formulae to describe the change of GDOP. The result shows that, for GNSS single point positioning solutions, if one more satellite belonging to the existing tracked multi-GNSS constellation used in the single point positioning solution is added, the GDOP always decreases with the number of the added satellites. On the other hand, when the constellation of the added satellite is not from the tracked existing constellations, the different numbers of the added satellites have different influences on the change of GDOP. Generally, adding one satellite from another constellation into the existing multi-GNSS constellations will increase the GDOP, but adding two satellites will decrease the GDOP compared with adding one from another constellation. Additionally, the GDOP also increases in the cases of adding two satellites from two different constellations into the tracked existing constellations.

This paper presents results from a doctoral dissertation (reference 12), on the problem of scheduling with Maximum Position Shift (MPS) constraints. The problem typically arises in a queueing system, when the service rate depends on the sequence in which customers are handled by the server. Such a queueing system is the airport runway system, where it has been shown (references 3, 10, 12) that rearranging the sequence of landings and takeoffs affects the runway capacity – that is, the rate of runway operations. An optimal Dynamic Programming (DP) scheduling algorithm is developed that takes as input the set of currently known flights, their desired time of operation, basis of the First Come First Served (FCFS) sequence, and a matrix of Air Traffic Control (ATC) minimum time separations, in order to produce an optimal schedule defined as runway and threshold time assignments maximizing runway capacity.

For multi-Autonomous Underwater Vehicle (multi-AUV) system task assignment and path planning, a novel Glasius Bio-inspired Self-Organising Map (GBSOM) neural networks algorithm is proposed to solve relevant problems in a Three-Dimensional (3D) grid map. Firstly, a 3D Glasius Bio-inspired Neural Network (GBNN) model is established to represent the 3D underwater working environment. Using this model, the strength of neural activity is calculated at each node within the GBNN. Secondly, a Self-Organising Map (SOM) neural network is used to assign the targets to a set of AUVs and determine the order of the AUVs to access the target point. Finally, according to the magnitude of the neuron activity in the GBNN, the next AUV target point can be autonomously planned when the task assignment is completed. By repeating the above three steps, access to all target points is completed. Simulation and comparison studies are presented to demonstrate that the proposed algorithm can overcome the speed jump problem of SOM algorithms and path planning in the 3D underwater environments with static or dynamic obstacles.

In this paper, back-propagation (BP) neural networks (NN) are applied to the GPS satellite Geometric Dilution of Precision (GDOP) approximation. The methods using BPNN are general enough to be applicable regardless of the number of satellite signals being processed by the receiver. BPNN is employed to learn the functional relationships firstly, between the entries of a measurement matrix and the eigenvalues and thus generate GDOP, and secondly, between the entries of a measurement matrix and the GDOP, both without inverting a matrix. Consequently, two sets of entries and two sets of output variables, respectively, are used that in total yield four types of mapping architectures. Simulation results from these four architectures are presented. The performance and computational benefit of neural network-based GDOP approximation are explored.

The information on the Closest Point of Approach (CPA) of another vessel to own ship is required in a potential collision situation as it helps determines the risk to each vessel. CPA is usually calculated based on the speed and direction of the approaching ship neglecting the Change Of Speed (COS) and the Rate Of Turn (ROT). This will make the CPA less useful. To improve the CPA calculation, Automatic Identification System (AIS) information containing the Speed Over Ground (SOG), Course Over Ground (COG), COS and ROT is used. Firstly, a model using these four factors is built to predict ship positions better. Secondly, a three-step CPA searching method is developed. The developed CPA calculation method can assist in informing the navigation decisions and reducing unnecessary manoeuvres. Through the analysis of a real collision scenario, this paper shows that the proposed method can help identify and warn of anomalous ship behaviours in a realistic time frame.

Marine accident analysis is a sophisticated and complex official interpretation that requires a professional and fair judgment. For accidents such as a collision, contact and other incidents in the maritime field, the judgment mechanism of the courts depends on the decision process of the field experts. Field experts define the “Fraction Defectives” (FDs) of the vessels for the intended case based on the existing evidence and navigational specifications. However, evaluation of human judgment can be limited and problematic in analysing many aspects of a case. In this paper, a pairwise comparison method is used to simplify and clarify the judgment process. We aim to assess marine accidents in a stepwise approach that is inaccurately carried out in a holistic perspective by the field experts. A real accident that occurred in the past is simulated in front of the field experts. After conducting an expert consultation, FDs are re-calculated to compare both the regular judgment and the results in our suggested Generic Fuzzy Analytical Hierarchy Process (GF-AHP) approach.

A review of existing knowledge regarding how to manage ice during offshore work in cold waters has been carried out. The objective has been to contribute to increased safety and efficiency in future projects by learning from world-wide experiences. It was found that offshore drilling has been carried out in a wide range of ice conditions and at water depths spanning from five metres to more than a thousand metres. Icebergs in open waters have been handled safely over several years and the possibility of detecting icebergs is considered good. With respect to icebergs frozen in sea ice, both detection and deflection is considered difficult and the technology for doing so is not proven. Good ice management systems are considered to represent the main factors for operating successfully in the covered waters. Future work should focus on how to include the effect of ice management statistically in a design process.

This paper focuses on depth trajectory tracking control for a Remotely Operated Vehicle (ROV) with dead-zone nonlinearity and saturation nonlinearity of thruster; an adaptive sliding mode control method based on neural network is proposed. Through the analysis of dead-zone nonlinearity and saturation nonlinearity of thruster, the depth trajectory tracking control system model of a ROV which uses thruster control signals as system input has been established. According to the principle of sliding mode control, an adaptive sliding mode depth trajectory tracking controller is built by using three-layer feed-forward neural network for online identification of unknown items. The selection method and update laws of the control parameters are also given. The uniform ultimate boundedness of trajectory tracking error is analysed by Lyapunov theorem. Finally, the effectiveness of the proposed method is illustrated by simulations.

This study examines maritime routes between ports along the Atlantic coast of the US, utilising Automated Identification System (AIS) data for the years 2010 through 2012. The delineation of vessel routes conducted in this study was motivated by development planned for offshore Wind Energy Areas (WEAs) along the Atlantic coast of the US and the need to evaluate the effect of these development areas on commercial shipping. To this end, available AIS data were processed to generate commercial vessel tracks for individual vessels, though cargo vessels are the focus in this study. The individual vessel tracks were sampled at transects placed along the Atlantic coast. The transect samples were analysed and partitioned by voyages between Atlantic ports to facilitate computation of vessel routes between ports. The route boundary analysis utilised a definition from UK guidance in which routes' boundaries encompassed 95% of the vessel traffic between ports. In addition to delineating route boundaries, we found multi-modal transverse distributions of vessels for well-travelled routes, which indicated preference for lanes of travel within the delineated routes.

A new nonlinear robust filter is proposed in this paper to deal with the outliers of an integrated Global Positioning System/Strapdown Inertial Navigation System (GPS/SINS) navigation system. The influence of different design parameters for an H∞ cubature Kalman filter is analysed. It is found that when the design parameter is small, the robustness of the filter is stronger. However, the design parameter is easily out of step in the Riccati equation and the filter easily diverges. In this respect, a singular value decomposition algorithm is employed to replace the Cholesky decomposition in the robust cubature Kalman filter. With large conditions for the design parameter, the new filter is more robust. The test results demonstrate that the proposed filter algorithm is more reliable and effective in dealing with the outliers in the data sets produced by the integrated GPS/SINS system.

Traditionally, GNSS receiver autonomous integrity monitoring (RAIM) has been based upon single epoch solutions. RAIM can be improved considerably when available dynamic information is fused together with the GNSS range measurements in a Kalman filter. However, while the Kalman filtering technique is widely accepted to provide optimal estimates for the navigation parameters of a dynamic platform, assuming the state and observation models are correct, it is still susceptible to unmodelled errors. Furthermore, significant deviations from the assumed models for dynamic systems may also occur. It is therefore necessary that the state estimation procedure is complemented with effective and reliable integrity measures capable of identifying both measurement and modelling errors. Within this paper, fundamental equations required for the effective detection and identification of outliers in a kinematic GNSS positioning and navigation system are described together with the reliability and separability measures. These quality measures are implemented using a Kalman filtering procedure formulated with Gauss-Markov models where the state estimates are derived from least squares principles. Detailed simulations and analyses have been performed to assess the impact of the dynamic information on GNSS RAIM with respect to outlier detection and identification, reliability and separability. The ability of the RAIM algorithms to detect and identify dynamic modelling errors is also investigated.