Obstacle detection on the railway, a crucial operational safety concern, is a complex task that encompasses a multitude of challenges. While Machine Learning (ML) algorithms are commonly employed in analogous applications such as autonomous car driving [1] [2], the railway field faces a significant barrier due to the scarcity of available data (particularly images), rendering conventional ML approaches impractical. In response to this challenge, this study proposes and evaluates a framework which uses LiDAR (Light Detection and Ranging) data for obstacle detection on the railways. The framework aims to address the limitations posed by image data scarcity while enhancing operational safety in railway environments. The developed methodology combines the use of a long-range LiDAR capable of detecting obstacles at distances of up to 500 meters, with the train’s GPS (Global Positioning System) coordinates to accurately determine its position relative to detected obstacles. The LiDAR data is processed using a data fusion approach, where pre-existing knowledge regarding the track topography is combined with a clustering algorithm, specifically DBSCAN (Density-based spatial clustering of applications with noise), to identify and classify potential obstacles at a pre-defined distance. Tests of the proposed framework were conducted within the confines of a moving locomotive, specifically the CP 2600-2620 series, along a designated section of the Contumil-Leixões line. These tests served to validate the effectiveness and feasibility of the approach under real-world operating conditions. Overall, the utilization of LiDAR data coupled with advanced algorithms presents a promising avenue for enhancing obstacle detection capabilities in railway operations. By overcoming the challenges associated with data scarcity, this framework holds the potential to significantly improve operational safety and efficiency within railway networks. Further research and testing are warranted to validate the framework’s performance across diverse railway environments and operating conditions.

Rail travel is one of the safest means of transportation, with increasing usage in recent years. One of the major safety concerns in the railway relates to intrusions. Therefore, the timely detection of obstacles is crucial for ensuring operational safety. This is a complex problem with multiple contributing factors, from environmental to psychological. While machine learning (ML) has proven effective in related applications, such as autonomous road-based driving, the railway sector faces unique challenges due to limited image data availability and difficult data acquisition, hindering the applicability of conventional ML methods. To mitigate this, the present study proposes a novel framework leveraging LiDAR technology (Light Detection and Ranging) and previous knowledge to address these data scarcity limitations and enhance obstacle detection capabilities on railways. The proposed framework combines the strengths of long-range LiDAR (capable of detecting obstacles up to 500 m away) and GNSS data, which results in precise coordinates that accurately describe the train’s position relative to any obstacles. Using a data fusion approach, pre-existing knowledge about the track topography is incorporated into the LiDAR data processing pipeline in conjunction with the DBSCAN clustering algorithm to identify and classify potential obstacles based on point cloud density patterns. This step effectively segregates potential obstacles from background noise and track structures. The proposed framework was tested within the operational environment of a CP 2600-2620 series locomotive in a short section of the Contumil-Leixões line. This real-world testing scenario allowed the evaluation of the framework’s effectiveness under realistic operating conditions. The unique advantages of this approach relate to its effectiveness in tackling data scarcity, which is often an issue for other methods, in a way that enhances obstacle detection in railway operations and may lead to significant improvements in safety and operational efficiency within railway networks.

This study focuses on the development of a machine learning (ML) model to elaborate on predictions of structural damage in submerged structures due to ocean states and subsequently compares it to a real-life case of a 6-month experiment with a benthic lander bearing a multitude of sensors. The ML model uses wave parameters such as height, period and direction as input layers, which describe the ocean conditions, and strains in selected points of the lander structure as output layers. To streamline the dataset generation, a simplified approach was adopted, integrating analytical formulations based on Morison equations and numerical simulations through the Finite Element Method (FEM) of the designed lander. Subsequent validation involved Fluid–Structure Interaction (FSI) simulations, using a 2D Computational Fluid Dynamics (CFD)-based numerical wave tank of the entire ocean depth to access velocity profiles, and a restricted 3D CFD model incorporating the lander structure. A case study was conducted to empirically validate the simulated ML model, with the design and deployment of a benthic lander at 30 m depth. The lander was monitored using electrical and optical strain gauges. The strains measured during the testing period will provide empirical validation and may be used for extensive training of a more reliable model.

Wide-area maritime surveillance is an increasingly important focus for countries with large Exclusive Economic Zones (EEZ), such as Portugal, which are responsible for monitoring and protecting these zones and their resources. Passive acoustic autonomous buoy networks equipped with hydrophones are a promising approach for wide-area maritime surveillance. However, achieving a discrete, low-cost system introduces many technical challenges. This work describes a practical, low-power, two-state architecture that separates continuous ship detection from detailed vessel class classification. First, an always-on microcontroller performs continuous binary ship presence detection and triggers the higher-power classifier only when a vessel is detected. The high-accuracy acoustic classifier was tested across embedded controllers to identify the minimum platform capable of sustaining its intended 1 Hz classification rate. A Raspberry Pi 5 achieved the 1 s target with a measured continuous consumption of 4 W; however, adding sensing, storage, and communications is expected to raise the always-on consumption to around 5 W. If this node was used by itself, a week-long autonomy requirement, therefore, would imply 840 Wh of usable energy storage, and recovering this deficit rapidly under limited insolation would require several hundred watts of photovoltaic capacity, driving both buoy volume and cost up. To address this, an always-on edge node based on an ESP32-S3 microcontroller was implemented, running a lightweight binary detection of a vessel presence model trained in Edge Impulse using a subset of Ocean Networks Canada recordings. The edge node consumes $0.69$ W continuously and is intended to trigger a wake-up line to power the higher-performance node only when a ship is detected, reducing average energy demand while maintaining the ability to run a richer classifier on demand. The presented architecture, profiling workflow, and energy calculations provide a path to power-aware passive acoustic monitoring systems suitable for extended maritime deployments.

Hand exoprosthesis are commonly assumed as a promising approach to help people regain independence after upper limb losses. Injury-related data from recent years highlights the need to continue developing solutions to increase end-user acceptance. Within this scope, the present review aims to provide up-to-date information related on advancements and current trends in hand exoprosthesis development. Following a PRISMA methodology, 60 studies were included in this review covering a different range of actuation strategies and design approaches. The main features of the devices developed in the literature are also presented in detail. Concerning actuation strategies, linkage-driven and tendon-pulley mechanisms are the most common approaches presented in the literature, however different strategies such as twisted-string actuators differential mechanisms are also proposed. In turn, pneumatic and hydraulic actuation approaches are also presented as soft alternatives to electric motors. Passive elements such as springs or clutch mechanisms are frequently employed to achieve underactuation in these devices. 3D Printed technologies are also suggested as alternatives to the most conventional manufacturing methods. By covering all these topics, the present review is meant to provide useful insights into future developments in this field. End-user-oriented continuous improvement and the development of highly anthropomorphic solutions are still current challenges, that should be addressed in upcoming developments. This work was developed in the scope of the project ‘NerveRepack—Intelligent neural system for bidirectional connection with exoprostheses and exoskeletons’, which has received funding from the Horizon Europe RIA programme under grant agreement N^∘ 101112347.

Monitoring large critical infrastructures is a highly complex and costly task. The use of a network of sensors to aid in the detection and identification of potential anomalies is therefore an important step towards easing maintenance effort while maintaining operational soundness. To address this challenge, a monitoring system was developed and installed in a seaport drawbridge. The structural parameters monitored during operation can be used to assess the bridge’s structural behavior. This provides the ability to identify potential anomalies that could lead to its failure at an early stage, allowing for the better planning of maintenance interventions, saving time and money. In this paper, the monitoring system will be presented and the employed signal identification and analysis methods will be described.