Abstract
Recently, research concerning the navigation of Autonomous Surface Vehicles (ASVs) has been increasing. However, a big scale implementation of these vessels is still held back by a plethora of challenges such as multi-object tracking. This article presents the development of a tracking model through transfer learning techniques, based on referenced object trackers for urban scenarios. The work consisted in training a neural network through deep learning techniques, including data association and comparison of three different optimisers, Adadelta, Adam and SGD, determining the best hyper-parameters to maximise the training efficiency. The developed model achieved decent performance at tracking large vessels in the ocean, being successful even in harsh lighting conditions and lack of image focus.

Abstract
A major effort is put into the production of green energy as a countermeasure to climatic changes and sustainability. Thus, the energy industry is currently betting on offshore wind energy, using wind turbines with fixed and floating platforms. This technology can benefit greatly from interventive autonomous underwater vehicles (AUVs) to assist in the maintenance and control of underwater structures. A wireless charger system can extend the time the AUV remains underwater, by allowing it to charge its batteries through a docking station. The present work details the development process of a housing component for a wireless charging system to be implemented in an AUV, addressed as wireless charger housing (WCH), from the concept stage to the final physical verification and operation stage. The wireless charger system prepared in this research aims to improve the longevity of the vehicle mission, without having to return to the surface, by enabling battery charging at a docking station. This product was designed following a design for excellence (DfX) and modular design philosophy, implementing visual scorecards to measure the success of certain design aspects. For an adequate choice of materials, the Ashby method was implemented. The structural performance of the prototypes was validated via a linear static finite element analysis (FEA). These prototypes were further physically verified in a hyperbaric chamber. Results showed that the application of FEA, together with well-defined design goals, enable the WCH optimisation while ensuring up to 75% power efficiency. This methodology produced a system capable of transmitting energy for underwater robotic applications.

Abstract
This paper presents an algorithm for mapping monopiles from Offshore Wind Farms (OWF). The ASV (Autonomous Surface Vehicle) surveys the environment, detects and localizes monopiles using situational awareness system based on LiDAR, GPS and IMU (Inertial Measurement Unit) data. The position of the monopile is obtained based on the relative localization between the extrapolated center of the structure that was detected and the ASV. A positive detection of a monopile is referenced to a global positioning frame based on the GPS. Results in a simulator environment demonstrate the ability of this situational awareness system to identify monopiles with a precision of 0.005 m, which is relevant for detecting structural disalignments over time that might be caused by the appearance of scour in the structure's foundation.

Abstract
The worldwide context has fostered the innovation geared to the blue growth. However, the aquatic environment imposes many restrictions to mobile robots, as their perceptual capacity becomes severely limited. DIIUS aims to strengthen the perception of distributed robotic systems to improve the current procedures for inspection of aquatic structures (constructions and/or vessels).The perception of large working areas from multiples robots raises a number of unresolved inference problems and calls for new interaction patterns between multiple disciplines, both at the conceptual and technical level. To address this important challenge, the DIIUS project seeks to reinforce the current state-of-art in several scientific domains that fit into artificial intelligence, computer vision, and robotics. Through case studies focused on 3D mapping of aquatic structures (ex., maritime constructions and adduction tunnels), the project investigates new spatio-temporal data association techniques, including the correlation of sensors from heterogeneous robot formations operating in environments with communications constraints.

Abstract
A growing interest in ocean exploration for scientific and commercial research has been shown, mainly due to the technological developments for maritime and offshore industries. The use of Autonomous surface vehicles (ASV) have a promising role to revolutionize the transportation, monitorization, operation and maintenance areas, allowing to perform distinct task from offshore assets inspection to harbor patrolling. This work presents SENSE, an autonomouS vEssel for multi-domaiN inSpection and maintEnance. It provides an open-source hardware and software architecture that is easy to replicate for both research institutes and industry. This is a multi-purpose vehicle capable of acquiring multi-domain data for inspecting and reconstructing maritime infrastructures. SENSE provides a research platform which can increase the situational awareness capabilities for ASVs. SENSE full configuration provides multimodal sensory data acquired from both domains using a Light Detection And Ranging (LiDAR), a stereoscopic camera, and a multibeam echosounder. In addition, it supplies navigation information obtained from a real-time kinematic satellite navigation system and inertial measurement units. Moreover, the tests performed at the harbor of Marina de Leca, at Porto, Portugal, resulted in a dataset which captures a fully operational harbor. It illustrates several conditions on maritime scenarios, such as undocking and docking examples, crossings with other vehicles and distinct types of moored vessels. The data available represents both domains of the maritime scenario, being the first public dataset acquired for multi-domain observation using a single vehicle. This paper also provides examples of applications for navigation and inspection on multi-domain scenarios, such as odometry estimation, bathymetric surveying and multi-domain mapping.

Abstract
The expansion of autonomous driving operations requires the research and development of accurate and reliable self-localization approaches. These include visual odometry methods, in which accuracy is potentially superior to GNSS-based techniques while also working in signal-denied areas. This paper presents an in-depth review of state-of-the-art visual and point cloud odometry methods, along with a direct performance comparison of some of these techniques in the autonomous driving context. The evaluated methods include camera, LiDAR, and multi-modal approaches, featuring knowledge and learning-based algorithms, which are compared from a common perspective. This set is subject to a series of tests on road driving public datasets, from which the performance of these techniques is benchmarked and quantitatively measured. Furthermore, we closely discuss their effectiveness against challenging conditions such as pronounced lighting variations, open spaces, and the presence of dynamic objects in the scene. The research demonstrates increased accuracy in point cloud-based methods by surpassing visual techniques by roughly 33.14% in trajectory error. This survey also identifies a performance stagnation in state-of-the-art methodologies, especially in complex conditions. We also examine how multi-modal architectures can circumvent individual sensor limitations. This aligns with the benchmarking results, where the multi-modal algorithms exhibit greater consistency across all scenarios, outperforming the best LiDAR method (CT-ICP) by 5.68% in translational drift. Additionally, we address how current AI advances constitute a way to overcome the current development plateau.