Abstract
[No abstract available]

Abstract

Abstract
Purpose: Visual perception enables robots to perceive the environment. Visual data is processed using computer vision algorithms that are usually time-expensive and require powerful devices to process the visual data in real-time, which is unfeasible for open-field robots with limited energy. This work benchmarks the performance of different heterogeneous platforms for object detection in real-time. This research benchmarks three architectures: embedded GPU-Graphical Processing Units (such as NVIDIA Jetson Nano 2 GB and 4 GB, and NVIDIA Jetson TX2), TPU-Tensor Processing Unit (such as Coral Dev Board TPU), and DPU-Deep Learning Processor Unit (such as in AMD-Xilinx ZCU104 Development Board, and AMD-Xilinx Kria KV260 Starter Kit). Methods: The authors used the RetinaNet ResNet-50 fine-tuned using the natural VineSet dataset. After the trained model was converted and compiled for target-specific hardware formats to improve the execution efficiency.Conclusions and Results: The platforms were assessed in terms of performance of the evaluation metrics and efficiency (time of inference). Graphical Processing Units (GPUs) were the slowest devices, running at 3 FPS to 5 FPS, and Field Programmable Gate Arrays (FPGAs) were the fastest devices, running at 14 FPS to 25 FPS. The efficiency of the Tensor Processing Unit (TPU) is irrelevant and similar to NVIDIA Jetson TX2. TPU and GPU are the most power-efficient, consuming about 5 W. The performance differences, in the evaluation metrics, across devices are irrelevant and have an F1 of about 70 % and mean Average Precision (mAP) of about 60 %.

Abstract
Wearable Health Devices (WHDs) are increasingly becoming an integral part of daily life and significantly contributing to self-monitoring in healthcare. WHDs have a wide range of applications, ranging from sports to clinical settings, where the monitoring of cardiovascular health, particularly through ECG, plays a crucial role. This study introduces a unique WHD called VitalSticker, which exhibits distinctive features such as having a comfortable tiny patch form-factor to be attached to the chest, collecting multiple vital signs with medical-grade quality (ECG, respiration, temperature and actigraphy) and seamlessly sending data to a companion app. This paper encompasses a detailed description of the hardware, firmware, and case design of the WHD. A study was conducted to assess the quality of the ECG signal acquired by VitalSticker, comparing it with the signal obtained from a CE medical-grade certified ambulatory device. The results demonstrate that our VitalSticker achieves similar medicalgrade quality when compared to the reference device, surpassing its counterpart in several specifications. Furthermore, this study presents the successful implementation of an ECG baseline wander correction filter that runs on the tiny on-board wearable microcontroller without introducing any artifacts into the ECG signal, reducing the need for further processing for this outside the wearable patch.

Abstract

Abstract
Offshore wind farms are becoming the main alternative to fossil fuels and the future key to mitigating climate change by achieving energy sustainability. With favorable indicators in almost every environmental index, these structures operate under varying and dynamic environmental conditions, leading to efficiency losses and sudden failures. For these reasons, it's fundamental to promote the development of autonomous solutions to monitor the health condition of the construction parts, preventing structural damage and accidents. This paper introduces a new simulation environment for testing and training autonomous inspection techniques under a more realistic offshore wind farm scenario. Combining the Gazebo simulator with ROS, this framework can include multi-robots with different sensors to operate in a customizable simulation environment regarding some external elements (fog, wind, buoyancy...). The paper also presents a use case composed of a 3D LiDAR-based technique for autonomous wind turbine inspection with UAV, including point cloud clustering, model estimation, and the preliminary results under this simulation framework using a mixed environment (offshore simulation with a real UAV platform).