Abstract
Human oversight can benefit scenarios with complex tasks, such as pallet docking and loading and unloading containers, beyond the current capabilities of autonomous systems without any failures. Furthermore, teleoperation systems allow remote control of mobile ground robots, especially with the surge of 5G technology that promises reliable and low latency communication. Current works research on exploring the latest features from the 5G standard, including ultra-Reliable Low-Latency Communication (uRLLC) and network slicing. However, these features may not be available depending on the Internet Service Provider (ISP) and communication devices. Thus, this work proposes a network architecture for the teleoperation of ground mobile robots in industrial environments using commercially available devices over the 5G Non-Standalone (NSA) standard. Experimental results include an evaluation of the network and End-to-End (E2E) latency of the proposed system. The results show that the proposed architecture enables teleoperation, achieving an average E2E latency of 347.19 ms.

Abstract
Data used by automated decision-making systems, such as Machine Learning models, often reflects discriminatory behavior that occurred in the past. These biases in the training data are sometimes related to label noise, such as in COMPAS, where more African-American offenders are wrongly labeled as having a higher risk of recidivism when compared to their White counterparts. Models trained on such biased data may perpetuate or even aggravate the biases with respect to sensitive information, such as gender, race, or age. However, while multiple label noise correction approaches are available in the literature, these focus on model performance exclusively. In this work, we propose Fair-OBNC, a label noise correction method with fairness considerations, to produce training datasets with measurable demographic parity. The presented method adapts Ordering-Based Noise Correction, with an adjusted criterion of ordering, based both on the margin of error of an ensemble, and the potential increase in the observed demographic parity of the dataset. We evaluate Fair-OBNC against other different pre-processing techniques, under different scenarios of controlled label noise. Our results show that the proposed method is the overall better alternative within the pool of label correction methods, being capable of attaining better reconstructions of the original labels. Models trained in the corrected data have an increase, on average, of 150% in demographic parity, when compared to models trained in data with noisy labels, across the considered levels of label noise.

Abstract
In fields that rely on high-stakes decisions, such as medicine, interpretability plays a key role in promoting trust and facilitating the adoption of deep learning models by the clinical communities. In the medical image analysis domain, gradient-based class activation maps are the most widely used explanation methods and the field lacks a more in depth investigation into inherently interpretable models that focus on integrating knowledge that ensures the model is learning the correct rules. A new approach, B-cos networks, for increasing the interpretability of deep neural networks by inducing weight-input alignment during training showed promising results on natural image classification. In this work, we study the suitability of these B-cos networks to the medical domain by testing them on different use cases (skin lesions, diabetic retinopathy, cervical cytology, and chest X-rays) and conducting a thorough evaluation of several explanation quality assessment metrics. We find that, just like in natural image classification, B-cos explanations yield more localised maps, but it is not clear that they are better than other methods' explanations when considering more explanation properties.

Abstract
Intervention studies have been explored to identify actions to effectively remediate indoor environmental quality (IEQ) problems and to improve people's health, well-being, comfort, and productivity. This study assessed a comprehensive set of IEQ indicators related to ventilation, air pollution, thermal comfort, illuminance, and noise for the first time in Portuguese office buildings. The purpose was to derive evidence-based corrective measures for a further environmental intervention program. The study monitored and surveyed 15 open-space offices from six modern office buildings in Porto (Portugal) during a workday between September and December 2022. Illuminance was of most concern among the assessed IEQ indicators since the measured levels were below the minimum limit required in 27% of the evaluated workplaces. For CO2, although mean concentrations were below 1000 ppm, absolute values exceeding that level were consistently registered in 20% of the offices during the afternoon period. Mean levels of PM2.5, PM10, and ultrafine particles exceeding the WHO guidelines were found in 13%, 7%, and 7% of the offices, respectively. The assessed thermal comfort levels were typically neutral, corresponding to an estimated mean of 6% of dissatisfied people. Based on the findings, an intervention plan was designed to be implemented in the further stages of this work. The priority interventions to test include relocation of printers (PM source removal), optimisation of ventilation rates (using real-time data from CO2 sensors), adjustment of desk positions to improve illuminance, and introduction of indoor plants.

Abstract
Over the years, mobile networks were deployed using monolithic hardware based on proprietary solutions. Recently, the concept of open Radio Access Networks (RANs), including the standards and specifications from O-RAN Alliance, has emerged. It aims at enabling open, interoperable networks based on independent virtualized components connected through open interfaces. This paves the way to collect metrics and to control the RAN components by means of software applications such as the O-RAN-specified xApps. We propose a private standalone network leveraged by a mobile RAN employing the O-RAN architecture. The mobile RAN consists of a radio node (gNB) carried by a Mobile Robotic Platform autonomously positioned to provide on-demand wireless connectivity. The proposed solution employs a novel Mobility Management xApp to collect and process metrics from the RAN, while using an original algorithm to define the placement of the mobile RAN. This allows for the improvement of the connectivity offered to the User Equipments.

Abstract
Human-robot collaboration (HRC) is becoming increasingly important in advanced production systems, such as those used in industries and agriculture. This type of collaboration can contribute to productivity increase by reducing physical strain on humans, which can lead to reduced injuries and improved morale. One crucial aspect of HRC is the ability of the robot to follow a specific human operator safely. To address this challenge, a novel methodology is proposed that employs monocular vision and ultra-wideband (UWB) transceivers to determine the relative position of a human target with respect to the robot. UWB transceivers are capable of tracking humans with UWB transceivers but exhibit a significant angular error. To reduce this error, monocular cameras with Deep Learning object detection are used to detect humans. The reduction in angular error is achieved through sensor fusion, combining the outputs of both sensors using a histogram-based filter. This filter projects and intersects the measurements from both sources onto a 2D grid. By combining UWB and monocular vision, a remarkable 66.67% reduction in angular error compared to UWB localization alone is achieved. This approach demonstrates an average processing time of 0.0183s and an average localization error of 0.14 meters when tracking a person walking at an average speed of 0.21 m/s. This novel algorithm holds promise for enabling efficient and safe human-robot collaboration, providing a valuable contribution to the field of robotics.