Abstract
Pastoral activities bring several benefits to the ecosystem and rural communities. These activities are already carried out daily with goats, cows and sheep in Portugal. Still, they could be better applied to take advantage of their benefits. Most of these pastoral ecosystem services are not remunerated, indicating a lack of making these activities more attractive to bring returns to shepherds, breeders and landowners. The monitoring of these activities provides data to value these services, besides being able to indicate directly to the shepherds’ routes to drive their flocks and the respective return. There are devices in the market that perform this monitoring, but they are not adaptable to the circumstances and challenges required in the Northeast of Portugal. This work addresses a system to perform animals tracking, and the development of a test platform, through long-range technologies for transmission using LoRaWAN architecture. The results demonstrated the use of LoRaWAN in tracking services, allowing to conclude about the viability of the proposed methodology and the direction for future works.

Abstract
The technological advances in Unmanned Aerial Vehicles (UAV) related to energy power structure inspection are gaining visibility in the past decade, due to the advantages of this technique compared with traditional inspection methods. In the particular case of power pylon structure and components, autonomous UAV inspection architectures are able to increase the efficacy and security of these tasks. This kind of application presents technical challenges that must be faced to build real-world solutions, especially the precise positioning and path following for the UAV during a mission. This paper aims to evaluate a novel architecture applied to a power line pylon inspection process, based on the machine learning techniques to process and identify the signal obtained from a UAV-embedded planar Light Detection and Ranging - LiDAR sensor. A simulated environment built on the GAZEBO software presents a first evaluation of the architecture. The results show an positive detection accuracy level superior to 97% using the vertical scan data and 70% using the horizontal scan data. This accuracy level indicates that the proposed architecture is proper for the development of positioning algorithms based on the LiDAR scan data of a power pylon.

Abstract
Nowadays, educational robotics is part of the learning activities in many K-12 schools. With the increasing interest in Computer Thinking education and acknowledging the importance of using tangible devices, many different educational robots for primary education have become available. With them, new research activities bring about new results concerning the use of robots in classes and how they can improve learning in STEAM areas. In this paper, a prototype of a new robot for primary school is presented. It has similar features to many other robots used in early school years (e.g. easy robot’s interface and one or two sensors, motor actuators), but with the advantage of having a low cost, being a do-it-yourself (DIY) kit and including a participation strategy, clarifying some of the learning targets, addressing the concept of alignment in learning activities.

Abstract
Industrial robot manipulators are widely used for repetitive applications that require high precision, like pick-and-place. In many cases, the movements of industrial robot manipulators are hard-coded or manually defined, and need to be adjusted if the objects being manipulated change position. To increase flexibility, an industrial robot should be able to adjust its configuration in order to grasp objects in variable/unknown positions. This can be achieved by off-the-shelf vision-based solutions, but most require prior knowledge about each object to be manipulated. To address this issue, this work presents a ROS-based deep reinforcement learning solution to robotic grasping for a Collaborative Robot (Cobot) using a depth camera. The solution uses deep Q-learning to process the color and depth images and generate a ? -greedy policy used to define the robot action. The Q-values are estimated using Convolutional Neural Network (CNN) based on pre-trained models for feature extraction. Experiments were carried out in a simulated environment to compare the performance of four different pre-trained CNN models (RexNext, MobileNet, MNASNet and DenseNet). Results show that the best performance in our application was reached by MobileNet, with an average of 84 % accuracy after training in simulated environment.

Abstract
The advancement of technology and techniques applied to robotics contributes to increasing the quality of life and safety of humanity. One of the most widespread applications of mobile robotics is related to monitoring indoor environments. However, due to factors such as the size of the environment impacting the monitoring response, battery autonomy, and autonomous navigation in environments with unknown obstacles, they are still significant challenges in the diffusion of mobile robotics in these areas. Strategy adopting multiple robots can overcome these challenges. This work presents an approach to use multi-robots in hazardous environments with gas leakage to perform spatial mapping of the gas concentration. Obstacles arranged in the environment are unknown to robots, then a fuzzy control approach is used to avoid the collision. As a result of this paper, spatial mapping of an indoor environment was carried out with multi-robots that reactively react to unknown obstacles considering a point gas leak with Gaussian dispersion.

Abstract
Task Scheduling assumes an integral topic in the efficiency of multiple mobile robots systems and is a key part in most modern manufacturing systems. Advances in the field of combinatorial optimisation have allowed the implementation of algorithms capable of solving the different variants of the vehicle routing problem in relation to different objectives. However few of this approaches are capable of taking into account the nuances associated with the coordinated path planning in multi-AGV systems. This paper presents a new study about the implementation of the Simulated Annealing algorithm to minimise the time and distance cost of executing a tasks set while taking into account possible pathing conflicts that may occur during the execution of the referred tasks. This implementation uses an estimation of the planned paths for the robots, provided by the Time Enhanced A* (TEA*) to determine where possible pathing conflicts occur and uses the Simulated Annealing algorithm to optimise the attribution of tasks to each robot, in order to minimise the pathing conflicts. Results are presented that validate the efficiency of this algorithm and compare it to an approach that does not take into account the estimation of the robots paths.