This study presents a benchmark for evaluating visual correspondence algorithms in underwater environments using both optical and sonar imaging. It analyzes the transferability of state-of-the-art feature matching methods designed initially for terrestrial data, under the specific challenges of marine sensing. Experiments on real and simulated datasets assess their accuracy, robustness, and downstream impact on visual odometry and image mosaicing. The findings highlight key limitations in generalization and provide insights toward developing more trustworthy perception systems for autonomous underwater robots (This work was supported in part by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) - project number 535678995.).

Forward-looking imaging sonars (FLS) enable perception in turbid water where optical cameras fail, but their measurements exhibit strong noise, viewpoint-dependent backscatter, and elevation-induced ambiguities that challenge correspondence based registration. We use here a training-free frequency-domain scan matching method, Fourier-SOFT in 2D (FS2D), on FLS imagery under an explicit quasi-planar motion model. The approach estimates translation by phase correlation and rotation by correlating Fourier magnitudes after a spherical projection evaluated with the SO(3) Fourier transform (SOFT), requiring no initial guess or navigation prior. We provide a quasi-planar validity analysis by sweeping the elevation integration width and quantifying the resulting degradation in translation and yaw accuracy. Experiments on simulated FLS data with ground truth report relative pose errors, runtime, and resolution ablations, and a field trial with an Unmanned Surface Vehicle (USV) qualitatively demonstrates practical mosaicking consistency.

In autonomous driving development, a perception dataset is crucial, as it provides fundamental data for training, testing, and validating algorithms for an autonomous vehicle's multimodal perception systems. So far, most research has concentrated on providing datasets for well-structured urban environments. This work introduces the A2RL V\textsubscript{max} open-source dataset, specifically designed for perception tasks in high-speed autonomous driving and multi-vehicle interaction. The dataset was captured during the 2024 Abu Dhabi Autonomous Racing League (A2RL), held at the Yas Marina F1 Circuit, with participation from all competing teams. It contains diverse scenarios, including single-vehicle data at varying speeds, multi-vehicle sessions, and the full final four-vehicle race. The dataset contains almost 30,000 professionally annotated LiDAR point clouds, along with RADAR point clouds. In particular, it is the first large-scale dataset in autonomous racing to feature professionally annotated LiDAR point clouds, enabling deep learning-based perception research. The data is provided in a developer-friendly format, enabling easy implementation and evaluation in future research. We provide implementation and evaluation for off-the-shelf 3D detection and tracking methods. Although baseline methods show promising results for both 3D detection and tracking, specialized methods are required to address the unique challenges of high-speed autonomous driving. For a detailed description of the dataset, please visit the \href{https://tum-avs.github.io/A2RL_Dataset_website/}{A2RL V\textsubscript{max} Dataset Website}

We promote in this paper the processing of radar data in the frequency domain to achieve higher robustness against noise and structural errors, especially in comparison to feature-based methods. This holds also for high dynamics in the scene, i.e., ego-motion of the vehicle with the sensor plus the presence of an unknown number of other moving objects. In addition to the high robustness, the processing in the frequency domain has the so far neglected advantage that the underlying correlation based methods used for, e.g., registration, provide information about all moving structures in the scene. A typical automotive application case is overtaking maneuvers, which in the context of autonomous racing are used here as a motivating example. Initial experiments and results with Fourier SOFT in 2D (FS2D) are presented that use the Boreas dataset to demonstrate radar-only-odometry, i.e., radar-odometry without sensor-fusion, to support our arguments.

A study is presented that gives an overview of the state of the art in hands-on robotics education. It is based on an international survey among instructors on their use of different forms of equipment, i.e., robot hardware, simulation, and data-sets, in combination with different access forms, i.e., in-presence, online, and take-home, across different teaching activities with practical components, i.e., labs, projects, and the supervision of BSc/MSc theses. Furthermore, it is checked whether the COVID-19 pandemic, which broke out five years before the survey, had lasting effects. It turns out that simulation is the most popular form of regularly used equipment, i.e., almost all instructors report to use simulation in robotics hands-on activities. It is followed by robot hardware where it is noteworthy that in labs about a quarter, i.e., 24% of the instructors do not regularly use it. Data-sets are particular popular as a occasionally used element. With respect to access forms, in-presence dominates for hardware. For simulation and data-sets, in-presence, online, and take-home are all popular and often used in combination. Among the alternatives to in-presence use of hardware, take-home robotics hardware is clearly more popular than online, i.e., remote access to robots. While the effect is less pronounced for participants with more than 10 years of teaching experience, there is a clear sentiment that the COVID-19 pandemic has had lasting effects on hands-on teaching in robotics. Instructors experimented with new forms of hands-on teaching during the pandemic, and they kept (some of) the changes afterward.

Fish age is an important biological variable required as part of routine stock assessment and analysis of fish population dynamics. Age estimates are traditionally obtained by human experts from the count of ring-like patterns along calcified structures such as otoliths. To automate the process and minimize human bias, modern methods have been designed utilizing the advances in the field of artificial intelligence (AI). While many AI-based methods have been shown to attain satisfactory accuracy, there are concerns regarding the lack of explainability of some early implementations. Consequently, new explainable AI-based approaches based on U-Net and Mask R-CNN have been recently published having direct compatibility with traditional ring counting procedures. Here we further extend this endeavor by creating an interactive website housing these explainable AI methods allowing age readers to be directly involved in the AI training and development. An important aspect of the platform presented in this article is that it allows the additional use of different advanced concepts of Machine Learning (ML) such as transfer learning, ensemble learning and continual learning, which are all shown to be effective in this study.

Mechanical Scanning Sonars (MSS) are popular underwater sensors for Unmanned Underwater Vehicles (UUV) due to their low cost, small size, and low power consumption. But due to their simplicity, there are also many research challenges related to their usage. Unfortunately, there is also a lack of data with ground truth UUV localization. We provide MSS datasets using a UUV with standard navigation sensors, i.e., an Inertial Measurement Unit (IMU) and a Doppler Velocity Log (DVL). The UUV is globally localized with a high precision optical tracking system in a large research pool to provide ground truth. The data is of interest for multiple research areas related to MSS, e.g., extraction of range information, registration of sonar scans, and especially mapping including Simultaneous Localization and Mapping (SLAM). Different parameter settings and environment conditions are covered, e.g., dynamics in the scene. The IMU and DVL data is also of interest for research on navigation independent of the MSS data. Results from navigation and mapping with an Extended Kalman Filter (EKF) are in addition provided as baseline solutions.

Determination of individual age is one essential step in the accurate assessment of fish stocks. In non-tropical environments, the manual count of ring-like growth patterns in fish otoliths (ear stones) is the standard method. It relies on visual means and individual judgment and thus is subject to bias and interpretation errors. The use of automated pattern recognition based on machine learning may help to overcome this problem. Here, we employ two deep learning methods based on Convolutional Neural Networks (CNNs). The first approach utilizes the Mask R-CNN algorithm to perform object detection on the major otolith reading axes. The second approach employs the U-Net architecture to perform semantic segmentation on the otolith image in order to segregate the regions of interest. For both methods, we applied a simple postprocessing to count the rings on the output masks returned, which corresponds to the age prediction. Multiple benchmark tests indicate the promising performance of our implemented approaches, comparable to recently published methods based on classical image processing and traditional CNN implementation. Furthermore, our algorithms showed higher robustness compared to the existing methods, while also having the capacity to extrapolate missing age groups and to adapt to a new domain or data source.

An open-source software framework is presented that allows real-time underwater mapping with popular marine robotics components, namely a BlueRobotics BlueROV2 with its standard Ping360 Mechanical Scanning Sonar (MSS) and a A50 Doppler Velocity Log (DVL), which are low-cost devices for their respective types - if not even the most affordable ones on the market. The software runs with low computational power on a Raspberry Pi4. The framework builds upon Synthetic Scan Formation (SSF) where single MSS beams or scan-lines are embedded into a pose-graph. The rendering of scans is not only based on navigation, but based on the graph itself. Scans formed from scan-lines can be optimized by online Simultaneous Localization and Mapping (SLAM) and result in improved scans, based on the current state of the graph. In subsequent steps this leads to improved registration results. To this end, a combination of two different types of loop-closures is presented. Namely a consecutive loop closure, and a proximity based loop closure, which reduces the overall drift. The framework is validated in three different test-environments, namely a pool, a test-tank with a gantry for ground truth motion, and the flooded basement of a WW-II submarine bunker. Among others, it is shown that there is an increased accuracy compared to conventional SLAM and that the software is usable in real-time during a mission with the low-cost hardware.

CONTEXT: Successful agile teams advance their work practices continuously. The continuous improvement of effective tool-based requirements practices is an important foundation of business agility. However, requirements tool practices are still widely rooted in plan-based approaches. They are not yet suited well for agile teams or agile businesses. OBJECTIVE: Report and make available an approach for continuous improvement of requirements practices so that tool-based requirements management can drive business agility. METHOD: Industry experience report based on a series of cases from different sources, including ones with involvement of the author. RESULTS: Processes and work practices for evolutionarily introducing and adapting requirements tools and tool-based requirements practices, in a way that supports business agility. CONCLUSION: The presented practices can guide organizations towards establishing effective, tool-based requirements practices that support business agility. A foundation is laid for further systematic investigation and development of the approach.

In this paper, we introduce Fourier-SOFT 2D (FS2D) as a new robust registration method. FS2D operates in the frequency domain where it exploits the well-known decoupling of rotation and translation. The challenging part of determining the rotation parameter is solved here based on a projection of the Fourier magnitude on a sphere and the SO(3) Fourier Transform (SOFT). The underlying use case is underwater mapping with sonar, i.e., with very noisy and partially overlapping environment data under non-trivial localization and navigation challenges. Fourier-SOFT 2D is compared with openly available registration methods on two real-world datasets and a simulated dataset. Results show the robustness of FS2D, i.e., its capabilities to handle large amounts of noise and occlusions of consecutive scans. The implementation in C++ is openly available.

Mechanical Scanning Sonars (MSS) are popular devices for Unmanned Underwater Vehicles (UUV), i.e., Remotely Operated Vehicles (ROV) and Autonomous Underwater Vehicles (AUV), as they function under low visibility conditions and over extended ranges. They are comparatively low-cost and easy to integrate. But they require motion-compensation due to the low updates rates caused by the mechanical scanning and the slow speed of sound. We present here a new form of scan formation for MSS where the data from single beams is embedded into a pose-graph. The rendering of scans is not as usual based on only core navigation sensors, but it can improve in the spirit of a synthetic aperture. To this end, online Simultaneous Localization and Mapping (SLAM) is used to form scans from the single beams. These can be optimized and improved scans in turn lead to improved registration results in subsequent steps. This Synthetic Scan Formation (SSF) leads to better mapping results than state-of-the-art SLAM with MSS. The method is validated with several real-world experiments. First, different trajectories with precise ground-truth in a pool with a gantry set-up are used. Second, results from field trials in a WW-II submarine bunker are presented. It is shown that there are clear quantitative and qualitative improvements, and that SSF can be used in real-time for mapping during a mission.

The REFrame workshop “REFraming Elicitation” aims at bringing together a broad spectrum of experts interested in requirements elicitation, concerning practical experiences and scientific findings alike. It provides a space for reflecting on and discussing current research, challenges, and experiences. The ultimate goal of the community organizing the workshop, and to which the workshop shall contribute, is to collaboratively build and publish a compendium of elicitation frameworks (methods, techniques, best practices etc.) that shall help address and solve elicitation challenges. The workshop includes peer-reviewed paper submissions, invited talks, and interactive discussion sessions.

Purpose of Review This review provides an overview of the current state of the art in Underwater Human-Robot Interaction (U-HRI), which is an area that is quite different from standard Human-Robot Interaction (HRI). This is due to several reasons. First of all, there are the particular properties of water as a medium, e.g., the strong attenuation of radio-frequency (RF) signals or the physics of underwater image formation. Second, divers are bound to special equipment, e.g., the breathing apparatus, which makes, for example, speech recognition challenging, if not impossible. Third, typical collaborative marine missions primarily requires a high amount of communication from the diver to the robot, which accordingly receives a lot of attention in U-HRI research. Recent Findings The use of gestures for diver-to-robot communication has turned out to be a quite promising approach for U-HRI as gestures are already a standard form of communication among divers. For the gesture front-ends, i.e., the part dealing with the machine perception of individual signs, Deep Learning (DL) has become to be a very prominent tool. Summary Human divers and marine robots have many complementary skills. There is hence a large potential for U-HRI. But while there is some clear progress in the field, the full potential of U-HRI is far from being exploited, yet.