Abstract
Thin-film sensors surfaces are becoming popular to collect data in several specific and complex processes, namely plastic injection or metal stamping, allowing the digitization of such processes through the use of Internet of Things technologies. A particular challenge in such thin-film sensors surfaces is the data acquisition and signal conditioning system, which implementation is complex due to the characteristics of these sensors (e.g., low amplitude and noisy signals), but even more complex when implemented in real industrial processes, which are subject to harsh conditions, namely noise, dirt and aggressive elements. This work describes a modular data acquisition and signals conditioning system for thin-film sensors surfaces, meeting the requirements of scalability, robustness and low-cost, meaning that it can be easily expanded according to the number of sensors required for the application scenario.

Abstract
Autonomous inspection Unmanned Aerial Vehicle systems are an essential research area, including power line distribution inspection. Considerable efforts to solve the demanding presented in the autonomous UAV inspection process are present in technical and scientific research. One of these challenges is the precise positioning and fly control of the UAV around the energy structures, which is vital to assure the security of the operation. The most common techniques to achieve precise positioning in UAV fly are Global Positioning Systems with Real-Time Kinematic. This technique demands a proper satellite signal receiving to work appropriately, sometimes hard to achieve. The present work proposes a complementary position data system based on augmented reality tags (AR Tags) to increase the reliability of the UAV fly positioning system. The system application is proposed for energy power tower inspections as an example of use. The adaptation to other inspection tasks is possible whit some small changes. Experimental results have shown that an increase in the position accuracy is accomplished with the use of this schema.

Abstract
The collaboration between humans and machines, where humans can share the same work environment without safety equipment due to the collision avoidance characteristic is one of the research topics for the Industry 4.0. This work proposes a system that acquires the space of the environment through an RGB-Depth sensor, verifies the free spaces in the created Point Cloud and executes the trajectory of the collaborative manipulator avoiding collisions. It is demonstrated a simulated environment before the system in real situations, in which the movements of pick-and-place tasks are defined, diverting from virtual obstacles with the RGB-Depth sensor. It is possible to apply this system in real situations with obstacles and humans, due to the results obtained in the simulation. The basic structure of the system is supported by the ROS software, in particular, the MoveIt! and Rviz. These tools serve both for simulations and for real applications. The obtained results allow to validate the system using the algorithms PRM and RRT, chosen for being commonly used in the field of robot path planning.

Abstract
New approaches on industrial mobile robots are changing the localization systems from old methods such as magnetic tapes to laser beacons based systems and natural landmarks since they are more adaptable and easier to install on the shop floor. Sensor fusion methods needs to be applied since there is information provided from different sources. Extended Kalman Filters are very used in the pose estimation of mobile robots with sensors that detect beacons and measure its distance and angle in a local referential frame. In certain situations, like for example wheels slippage, the number of impulses read for the encoders is wrong, resulting in a very large displacement or rotation and causing a bad estimation at the end of the prediction step. This bad estimation is used for the linearization of the non-linear equations, causing a bad linear approximation and probably a failure in the Kalman Filter. In this paper it is demonstrated that if we use the last state estimation calculated in the update step at the last cycle, instead of the estimation from the prediction step in the actual cycle, the result is an estimator much more robust to errors in the odometry information. Simulated and real results from several experiments are illustrated to demonstrate this new approach. © 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the scientific committee of the FAIM 2021.

Abstract
The 4th industrial revolution promotes the automatic inspection of all products towards a zero-defect and high-quality manufacturing. In this context, collaborative robotics, where humans and machines share the same space, comprises a suitable approach that allows combining the accuracy of a robot and the ability and flexibility of a human. This paper describes an innovative approach that uses a collaborative robot to support the smart inspection and corrective actions for quality control systems in the manufacturing process, complemented by an intelligent system that learns and adapts its behavior according to the inspected parts. This intelligent system that implements the reinforcement learning algorithm makes the approach more robust once it can learn and be adapted to the trajectory. In the preliminary experiments, it was used a UR3 robot equipped with a Force-Torque sensor that was trained to perform a path regarding a product quality inspection task.

Abstract
Many real-world robotic scenarios require performing task planning to decide courses of actions to be executed by (possibly heterogeneous) robots. A classical centralized planning approach has to find a solution inside a search space that contains every possible combination of robots and goals. This leads to inefficient solutions that do not scale well. Multi-Agent Planning (MAP) provides a new way to solve this kind of tasks efficiently. Previous works on MAP have proposed to factorize the problem to decrease the planning effort i.e. dividing the goals among the agents (robots). However, these techniques do not scale when the number of agents and goals grow. Also, in most real world scenarios with big maps, goals might not be reached by every robot so it has a computational cost associated. In this paper we propose a combination of robotics and planning techniques to alleviate and boost the computation of the goal assignment process. We use Actuation Maps (AMs). Given a map, AMs can determine the regions each agent can actuate on. Thus, specific information can be extracted to know which goals can be tackled by each agent, as well as cheaply estimating the cost of using each agent to achieve every goal. Experiments show that when information extracted from AMs is provided to a multi-agent planning algorithm, the goal assignment is significantly faster, speeding-up the planning process considerably. Experiments also show that this approach greatly outperforms classical centralized planning.