Abstract
This paper presents the development of an automated water sampling system designed to enhance quality control in water treatment facilities. The system is built around a mechanical structure that houses a watertight box containing all electronic components. A display inside the box allows users to program sampling schedules, including parameters such as the day, time, number of samples, sample volume, and intervals between samples. A balance integrated into the structure holds a bottle that collects the water samples, while a reservoir at the bottom accumulates water to ensure an adequate supply. A water pump connected to the structure enables controlled sample collection. The design ensures that all components are compactly integrated, while a non-invasive method is used to measure the volume of the sampled water, thereby avoiding direct contact between the sensors and the sample. A Project-Based Learning (PBL) approach, coupled with direct industry collaboration, has reinforced the effectiveness of active learning methodologies in engineering education.

Abstract
This paper explores an innovative distributed real-time control system for a 3D-printed robotic leg. The system is constructed on a modular multi-board architecture that seamlessly integrates with ROS2 and micro-ROS, demonstrating the use of 3D printing for rapid prototyping and customized solutions. A notable feature of this robotic leg is its 360-degree rotating joint, which extends its range of motion, enabling intricate and versatile movements. Incorporating a shoulder joint further facilitates sideways mobility, augmenting its operational capabilities. A multi-board architecture is designed to ensure efficient communication, ease of component interchangeability, and robust scalability for future development. Additionally, advanced control techniques, including tuning of proportional-integral-derivative (PID) controllers, ensure responsive joint actuation tailored to the unique properties of 3D-printed materials. Experimental validation indicates low latency and stable operation, underscoring the system's effectiveness for real-time robotic applications.

Abstract

Abstract
Inspecting reflective parts is challenging due to strong specular reflections that conceal small porosities and reduce defect visibility. This work presents a framework that combines augmented reality with a deep learning detector. An augmented reality headset is used to capture multi-view images under natural illumination, enabling the operator to adjust the viewpoint and obtain angles that reduce glare. The collected data form a 640 × 480 dataset used to train a yolov8 detection model, integrated into a Robot Operating System 2 architecture for real-time processing. Testing on an independent set of unseen parts yields a precision of 86.70 %, a recall of 87.26 %, and an F1-score of 86.97 %. Additional qualitative examples confirm that the model can identify low-contrast porosities despite reflective surfaces. The results demonstrate the feasibility of AR-assisted acquisition combined with deep learning for real-time inspection of machined aluminum components in a laboratory case study. © 2026 IEEE.

Abstract
Accurate dynamic modeling of ground robots (Unmanned Ground Vehicles - UGVs) is essential for robust control and navigation in agricultural environments, where variations in soil friction and rolling resistance significantly affect system dynamics. This work proposes a Linear Parameter-Varying (LPV) model parameterized by the friction coefficient, identified under different soil conditions using two excitation strategies: Amplitude-Pseudo-Random Binary Sequence (APRBS) and standard maneuvers (SM). A simulated ground robot - the Clearpath Husky - was used under multiple soil friction scenarios within the ROS 2 and Gazebo simulation environment. The results show that the LPV model effectively captures the influence of soil friction, with both LPV APRBS and LPV SM yielding similar RMSE values across scenarios. The results also highlight the feasibility of using SM-based excitation for identifying the robot dynamics. © 2026 IEEE.

Abstract
Building 3D maps in agricultural environments is challenging due to dense vegetation, irregular terrain, lack of landmarks, and unreliable GPS. This paper proposes a Bounding Box-Based 3D Mapping method using collaboration between an Unmanned Ground Vehicle (UGV) and an Unmanned Aerial Vehicle (UAV). The method simplifies crop rows and tree canopies by enclosing their point clouds in 3D bounding boxes, fused with original UAV and UGV data, producing compact maps that preserve essential structures for autonomous navigation and trajectory planning. Evaluation in a simulated Orchard scenario shows that the method could reduce map size by up to 60% while maintaining 83.6% coverage. Multi-robot collaboration proved crucial, with the UGV contributing 74% and the UAV 26% of the merged map. Overall, the proposed method demonstrates potential and deserves further investigation in more complex agricultural scenarios. © 2026 IEEE.