Efficient container handling and early damage detection are critical for minimizing operational delays, reducing costs, and ensuring safety in global maritime logistics.This work presents a deep learning-based methodology for real-time container tracking and automated damage detection during crane unloading operations at container terminals.We develop and deploy two specialized YOLOv12-based object detection models: one for identifying containers in motion and another for detecting structural damages such as bents, dents, and holes.Our models are trained and evaluated on a real-world dataset curated from video feeds captured at the EUROGATE Container Terminal in Limassol, Cyprus.The system is designed for robust performance under realistic terminal conditions, including variable lighting and motion.Our models achieve high detection accuracy, with a mAP50 of 0.99 for container detection and 0.75 for damage detection, substantially outperforming existing benchmarks.These results highlight the practical potential of our method for improving efficiency and safety in automated maritime logistics.

The global navigation satellite system (GNSS) is a widely used positioning technology that provides heading for moving objects with speed greater than 1 m/s; however, it cannot determine heading for a stationary receiver. The pedestrian dead reckoning (PDR) is a good substitute, especially in the absence of GNSS, due to its independence of external measurements. Pedestrian location and trajectory cannot be effectively estimated by a PDR system without an accurate and precise heading estimation. Heading is a vital part that has a direct effect on the PDR system’s overall accuracy and performance. The complex nature of human movement and the variety of phone orientations and placements make pedestrian heading estimation a persistent challenge. Techniques introduced in the literature often suffer from drift and inaccuracies, particularly when using gyroscope measurements, which might become biased over time. This study presents a novel heading estimation method that utilizes optimum axes mapping through the fusion of accelerometer and magnetometer data. Data is gathered from ten different realistic phone placements, providing a rich dataset that considers a wide range of input patterns. It is evident from the results of the experimental evaluation that the proposed heading estimation outperforms other well-known heading estimation solutions proposed in the literature, including a commercial application. The heading is estimated in a linear constant manner without fluctuations, thus avoiding future drift. Furthermore, the root mean square error (RMSE) is less than 14.67° for all experimental settings, with the lowest value of 1.75° achieved when placing the phone in the left front pants pocket with the screen facing outwards.

This study investigates the combined berth allocation problem (BAP) and quay crane allocation problem (QCAP) while considering a multi-quay setting. First, a mixed integer linear programming mathematical model is developed based on various constraints and real port settings. Then, the multi-quay combined BAP and QCAP is solved using both the exact method and a metaheuristic optimization method, namely, the cuckoo search algorithm (CSA). This analysis pertains to a one-week planning scenario, utilizing data from a real port. The results of the comparative analysis show that the proposed CSA can provide a near-optimal solution (< 1.02% from the optimal) at a fraction of the computational time (10 times faster), as compared to the exact solution. This makes it suitable for solving larger instances of the combined BAP and QCAP for bigger terminals and extended planning horizons.

The Internet of things (IoT) ecosystem provides a platform for the connectivity of interrelated smart devices to automate manual processes and reduce labor costs.IoT has brought significant benefits to all industries, including maritime, as various objects (e.g., ports, ships, agents, etc.) are connected to gather and share information within the maritime ecosystem.The innovative technological aspects of IoT are promoting the effective collaboration between the research community and the maritime industry, for enhancing the performance of maritime transportation systems.Therefore, this study discusses recent advances delivered by the IoT and other emerging technologies, like machine learning (ML) and computer vision (CV), for smart maritime transportation systems (SMTSs).In particular, this paper presents two specific use cases of SMTSs, namely, predictive maintenance and container damage/seal inspection.Moreover, the key benefits of integrating IoT with ML and CV are highlighted for the above-mentioned use cases.Finally, a discussion is presented to highlight key opportunities along with foreseeable future challenges in adopting these new technologies by the maritime industry.

The Automatic Identification System (AIS) provides information for tracking and monitoring vessel activity in real time. The vessel traffic data from AIS includes position coordinates in latitude and longitude, speed and course over ground, the vessel's unique identification number, and many more. In this work, we investigate the use of AIS data for estimating sea surface current velocities in the Eastern Mediterranean sea. Specifically, we apply the dead reckoning technique to compute the difference between the projected position and the true position of a vessel over time, which is mainly attributed on the force of sea surface currents. The estimated sea surface current velocities and directions are compared with the ones provided by the Copernicus Marine ocean product system. The analysis reveals that the dead reckoning technique can be used reliably for estimating sea currents at a very fine granularity, especially in high-traffic and coastal areas, where there is an increased complexity of obtaining accurate results from other sources.

Automatic identification system (AIS) data provides a wealth of information regarding vessel traffic and is used for a variety of applications such as collision detection and avoidance, route prediction and optimization, search and rescue operations, etc. However, several challenges exist when working with AIS data including huge volume and velocity (as AIS signals are sent by vessels every few seconds), message duplication, various types of data irregularities, as well as the need for real-time processing and analysis. This paper presents a new framework for collecting, processing, storing, and analyzing AIS data in real time plus a set of algorithms for doing so in an efficient and scalable way. At the same time, a set of intelligent services are provided as building blocks for improving and creating new AIS data driven applications. This framework has been operational for the past few years in Cyprus, and has collected and processed around one billion AIS messages from the Eastern Mediterranean Sea.

The port call process encapsulates a visitation cycle of a ship to a port and can generate a wealth of data. The real time analysis of port call data can be used to find bottlenecks in the port call process, establish targets based on key performance indicators (KPIs), and to understand how shipping traffic impacts a port's efficiency. This demonstration will showcase a new Power BI interactive report powered by a multidimensional OLAP cube for very fast performance, which is built on top of a data warehouse collecting information from various sources in real time. The report currently visualizes several KPIs and other types of information that can be filtered per port, time-period, vessel type, origin or destination ports, and various other categories to help manage arrivals, departures, and port operations.

Proper utilization of port resources and efficient berth planning play a crucial role in minimizing port congestion and overall handling costs. Therefore, this study focuses on efficient berth planning in maritime container terminals composed of multiple quays. In particular, this study addresses the Multi-Quay Berth Allocation Problem (MQ-BAP), where a continuous berthing layout is considered along with dynamic ship arrivals and practical constraints such as safety time windows and safety distances between ships. Since MQ-BAP is an NP-hard problem, this study proposes a metaheuristic-based approach, the Cuckoo Search Algorithm (CSA) for solving the problem. A comparative study is also performed using real data instances collected from the Port of Limassol, Cyprus, against a genetic algorithm solution proposed in the recent literature, as well as the optimal exact solution implemented using MILP. The results of the experiments show the effectiveness of our proposed CSA approach in handling real-world berth allocation in ports with multiple quays while also considering practical constraints.

Over the last couple of decades, demand for seaborne containerized trade has increased significantly and it is expected to continue growing over the coming years. As an important node in the maritime industry, a maritime container terminal (MCT) should be able to tackle the growing demand for sea trade. Due to the increased number of ships that can arrive simultaneously at an MCT combined with inefficient berth allocation procedures, there are often undesirable situations when the ships have to stay in waiting queues and delay both their berthing and departure. In order to improve port efficiency in terms of reducing the total handling cost and late departures, this study investigates the dynamic and continuous berth allocation problem (DC-BAP), where vessels are assigned dynamically as they arrive at their berth locations assuming a continuous berth layout. First, the DC-BAP is formulated as a mixed-integer linear programming (MILP) model. Since BAP is an NP-hard problem and cannot be solved by mathematical approaches in a reasonable time, this study adopts the recently developed metaheuristic cuckoo search algorithm (CSA) to solve the DC-BAP. For validating the performance of the proposed CSA method, we use a benchmark case study and a genetic algorithm solution proposed in recent literature as well as compare our results against the optimal MILP solution. From the simulation results, it becomes evident that the newly proposed algorithm has higher efficiency over counterparts in terms of optimal berth allocation within reasonable computation time.

The recent emergence of Internet-of-Things (IoT) technologies in mission-critical applications in the maritime industry has led to the introduction of the Internet-of-Ships (IoS) paradigm. IoS is a novel application domain of IoT that refers to the network of smart interconnected maritime objects, which can be any physical device or infrastructure associated with a ship, a port, or the transportation itself, with the goal of significantly boosting the shipping industry toward improved safety, efficiency, and environmental sustainability. In this article, we provide a comprehensive survey of the IoS paradigm, its architecture, its key elements, and its main characteristics. Furthermore, we review the state of the art for its emerging applications, including safety enhancements, route planning and optimization, collaborative decision making, automatic fault detection and preemptive maintenance, cargo tracking, environmental monitoring, energy-efficient operations, and automatic berthing. Finally, the presented open challenges and future opportunities for research in the areas of satellite communications, security, privacy, maritime data collection, data management, and analytics, provide a road map toward optimized maritime operations and autonomous shipping.

Intelligent buildings are typically endowed with sensing devices that are able to measure the concentration of specific contaminants in relevant zones. The collected measurements are subsequently processed by intelligent algorithms in order to enable the prompt detection and isolation of contaminant sources inside the building. Unfortunately, in real-world conditions, these sensing devices may suffer from faults affecting the sensors or the embedded electronics. Such faults, generally result in perturbed or missed data in the acquired data-stream, that can induce false alarms (or possibly missed alarms) and compromise the contaminant detection and isolation ability. This paper proposes a three-layer cognitive monitoring system for the detection and isolation of both contaminants and sensor faults in intelligent buildings. The first two layers are designed for the prompt detection of small variations in the concentration of a specific contaminant, while reducing the possible occurrence of false alarms. At the third layer, a cognitive mechanism employing a propagation model for the contaminant, which is based on the airflows between the building zones, allows to isolate the source zone and discriminate between sensor faults and the presence of a contaminant source. The proposed method is validated using a realistic 14-zone building scenario.

Intelligent buildings are responsible for ensuring the indoor air quality for their occupants under normal operation as well as under possibly harmful contaminant events due to accidental or malicious actions. An emerging environmental control application is monitoring the intelligent buildings against the presence of such events, by incorporating various sensing technologies and distributed detection and isolation algorithms. The needed simplicity, the improved scalability and fault tolerance are some of the main reasons for choosing distributed approaches over centralized ones. Hence, the effective partitioning of buildings into smaller sections for contaminant detection and isolation approaches is of great importance. In this paper, we present an exact Mixed Integer Linear Programming (MILP) formulation for partitioning the building into smaller sections. The building is transformed into a graph which is partitioned into subgraphs indicating the groups of zones in each section while ensuring (i) maximum decoupling between the various subgraphs, (ii) strong connectivity between the zones of a subgraph and (iii) control of the number of allocated zones in each subgraph. The main contribution of this work is the automatic partitioning of the building into sections, which enables the distributed simulation, modeling, analysis and management of the intelligent building in real time, while ensuring the effective detection and isolation of contaminants in the building interior.

This paper investigates the use of a Wireless Sensor Network for localizing and tracking multiple event sources (targets) using only binary data. Due to the simple nature of the sensor nodes, sensing can be tampered (accidentally or maliciously), resulting in a significant number of sensor nodes reporting erroneous observations. Therefore, it is essential that any event tracking algorithm used in Wireless Sensor Networks (WSNs) exhibits fault tolerant behavior in order to tolerate misbehaving nodes. The main contribution of this paper is the development and analysis of a low-complexity, distributed, real-time algorithm that uses the binary observations of the sensors for identifying, localizing, and tracking multiple targets in a fault tolerant way. Specifically, our results indicate that the proposed algorithm retains its performance in tracking accuracy in the presence of noise and faults, even when a large percentage of sensor nodes (25 percent) report erroneous observations.

The provision of accurate and reliable localization and tracking information for a target moving inside a binary Wireless Sensor Network (WSN) is quite challenging, especially when sensor failures due to hardware and/or software malfunctions or adversary attacks are considered. Most tracking algorithms assume fault-free scenarios and exploit all binary sensor observations, thus their accuracy may degrade when faults are present in the field. Spatiotemporal information available while the target is traversing the sensor field can be used not only for tracking the target, but also for detecting certain types of faults that appear highly correlated both in time and space. Our main contribution is ftTRACK, a target tracking architecture that is resilient to sensor faults and consists of three main components, namely the sensor health-state estimator, a fault-tolerant localization algorithm, and a location smoothing component. The key idea in the ftTRACK architecture lies in the sensor health-state estimator that leverages spatiotemporal information from previous estimation steps to intelligently choose which sensors to employ in the localization and tracking tasks. Simulation results indicate that ftTRACK maintains a high level of tracking accuracy, even when a large number of sensors fail.

Tracking of a source (target) which is moving inside a binary Wireless Sensor Network (WSN) is a challenging problem particularly when sensors may fail either due to hardware and/or software malfunctions, energy depletion or adversary attacks. Using information from failed sensors to locate and track a target may lead to high estimation errors, therefore, there is a need to develop fault tolerant localization algorithms which perform well even when a percentage of the sensors report erroneous observations. Alternatively, one may fuse information from neighboring sensors in order to determine the health state of each sensor, and subsequently use only healthy sensors in the localization and tracking process. Our contribution is the development of an architecture which combines the sensor health state estimation together with fault tolerant localization algorithms that leads to more robust target tracking in binary WSNs. Simulation results indicate that the proposed approach is resilient to various types of faults.

We consider the problem of target (event source) tracking using a binary Wireless Sensor Network (WSN). For this problem, a WSN consisting of sensors that can detect the presence of a target in an area around them, should fuse the information received by the individual sensors in order to localize and track the target. This is a challenging problem particularly when sensors may fail either due to hardware and/or software malfunctions, energy depletion or adversary attacks. Using information from failed sensors during target tracking may lead to high estimation errors. Since failure of individual sensors is unavoidable, there is a need to estimate the health state of each sensor in order to ignore those sensors that are considered as faulty. The contribution of this work is the investigation of three different algorithms for estimating the sensors' health state simultaneously with target tracking.

This paper investigates the use of a sensor network for localizing and tracking a moving target using only binary data. Due to the simple nature of the sensor nodes, sensing can be tampered (accidentally or maliciously), resulting in a significant number of sensor nodes reporting erroneous observations. Therefore, it is essential that any event tracking algorithm used in Wireless Sensor Networks (WSNs) exhibits fault tolerant behavior in order to tolerate a number of misbehaving nodes. SNAP (Subtract on Negative Add on Positive), is a simple event localization algorithm designed for WSNs applications that exhibits this fault tolerant behavior. The main contribution of this paper is to combine the decentralized implementation of SNAP with Kalman Filter techniques for tracking the movement of a target. This efficient tracking procedure provides fairly accurate results and turns out to be fault tolerant even when a large percentage of the sensor nodes report erroneous observations.

The increasing demand for indoor location-based services has motivated the development of positioning methods that exploit the existing wireless network infrastructure. Accuracy is an important requirement, however fault tolerance is also highly desirable in case of failures or malicious attacks. We investigate the fault tolerance of fingerprint-based methods under a variety of fault or attack scenarios. We study the Subtract on Negative Add on Positive (SNAP) algorithm and modify it appropriately for the WLAN setup. Our results indicate that SNAP achieves adequate accuracy with very low computational complexity and exhibits smoother performance degradation in the presence of faults compared to other methods.

The provision of reliable location estimates in case of unpredicted failures or malicious attacks, which inject faults and compromise the performance of the positioning system, is very important. Thus, our main interest is on the fault tolerance of WLAN fingerprint-based methods, rather than the absolute positioning error in the fault-free case. We study the Nearest Neighbor method and as a first step we develop a robust detection scheme to accurately detect faults. We incorporate this into a hybrid positioning method that switches to a modified distance metric, instead of the Euclidean, if faults are present. Experimental results indicate that the proposed approach exhibits higher resilience to faults compared to other positioning methods.

Accurate and reliable location estimates using wireless networks are important for enabling indoor location oriented services and applications, such as in-building guidance and asset tracking. Providing adequate level of accuracy in case of faults or attacks to the positioning system is equally significant, thus our main interest is on the fault tolerance of positioning methods, rather than the absolute accuracy in the fault-free case. We introduce several fault models to capture the effect of failures in the wireless infrastructure or malicious attacks and discuss how these models can simulate the corruption of signal strength values during positioning. The models are used to investigate the fault tolerance of positioning methods and evaluate them in terms of their performance degradation as the percentage of corrupted signal strength measurements increases. Experimental results using our fault models are also presented.

One of the main applications of Wireless Sensor Networks (WSNs) is area monitoring (e.g., environmental monitoring). In such problems, it is desirable to maximize the area coverage which can be achieved by appropriately positioning the sensors (if possible) and/or by increasing the detection range of the sensors. This paper considers the latter. The emphasis is on pairs of closely spaced sensors that can collaborate in order to increase their collective area coverage. The main contribution of this work is to investigate collaborative detection schemes between a pair of sensor nodes and show that the area coverage achieved by each scheme depends on the distance between the two sensors. For closely spaced sensors, we propose the Enhanced Covariance Detector (ECD) that combines the energy and the covariance information from the two nodes by utilizing two different thresholds (one for the energy test statistic and another for the covariance).

This paper investigates Wireless Sensor Networks (WSNs) for achieving fault tolerant localization of an event using only binary information from the sensor nodes. In this context, faults occur due to various reasons and are manifested when a node outputs a wrong decision. The main contribution of this paper is to propose the Fault Tolerant Maximum Likelihood (FTML) estimator. FTML is compared against the Centroid (CE) and the classical maximum likelihood (ML) estimators and is shown to be significantly more fault tolerant. Moreover, this paper compares FTML against the SNAP (Subtract on Negative Add on Positive) algorithm and shows that in the presence of faults the two can achieve similar performance; FTML is slightly more accurate while SNAP is computationally less demanding and requires fewer parameters.

This paper investigates the use of wireless sensor networks for estimating the location of an event that emits a signal that propagates over a large region. In this context, we assume that the sensors make binary observations and report the event (positive observations) if the measured signal at their location is above a threshold; otherwise, they remain silent (negative observations). Based on the sensor binary beliefs, a likelihood matrix is constructed whose maximum value points to the event location. The main contribution of this work is Subtract on Negative Add on Positive (SNAP), an estimation algorithm that provides an efficient way of constructing the likelihood matrix by simply adding \pm 1 contributions from the sensor nodes depending on their alarm state (positive or negative). This simple estimation procedure provides very accurate results and turns out to be fault tolerant even when a large percentage of the sensor nodes report erroneous observations.

This paper investigates the benefits of spatial correlation for the problem of coverage in Wireless Sensor Networks (WSN). Specifically, it studies two detectors, the Mean Detector (MD) and the Covariance Detector (CD) and compares their coverage performance for various scenarios. The main contribution of this paper is to show that one can exploit the possible correlation between measurements of neighboring sensor nodes in order to achieve significantly better coverage. Moreover, the results of this paper have direct implications on the topology of the network. Our results indicate that for the CD the best placement would be pairs of sensor nodes placed on a grid configuration.