ABSTRACT. In this paper we introduce PolySLAM, a SLAM algorithm that produces PolyMaps. Our PolyMap utilizes polygons built from vectors to model the environment, and as such this is a special case of vector-based SLAM algorithms. The results of our experiments both in simulation and on real world data sets show that this SLAM is promising. First, maps are more compact which reduces memory consumption compared to standard grid-maps making it attractive for map sharing over wireless connection such as in multi-robots systems. Second, the quality of maps produced by PolySLAM is already good for localization despite the lack of global optimization or loop closure.

ABSTRACT. The photogrammetry, 3D reconstruction from images, is an old technique but it's potentials could only be seen after the development of computers and digital photographs. Nowadays it has many applications, as creating scenarios for games, acquiring human expressions, roof inspection, stockpile measurement, high voltage transformer inspection, etc. As new technologies appear, new applications to photogrammetry are created. In this paper the use of available open and closed-source algorithms for 3D reconstruction and texturization is investigated. To achieve this goal, images of a fountain from several points-of-view were used. Next a comparison between several open and closed-source algorithms was performed, evaluating the number of faces, time consumption, RAM memory, GPU memory and the generated textured 3D models. The results obtained demonstrate that with the right setup, current open-source algorithms can achieve results near or better than proprietary software. Regarding the comparison, 3Dflow and MeshRecon presented the most accurate textured 3D models. When comparing quantitative measures, though, MeshRecon presented a slightly better performance in time consumption, but 3Dflow had a better RAM memory usage and a lower quantity of faces with a similar level of details.

ABSTRACT. Mobile robot navigation mostly relies on information gathered by lidars and cameras to detect and map obstacles. However, these sensors become useless both in fires and other environments that have reduced visibility due to smoke, dust and smog. Radars, on the other hand, are not affected by these and therefore provide a viable solution to support robots in low visibility environments. However, radars provide sparse measurements and may detect multi-path reflections, resulting in perception of ghost objects. This work demonstrates the use of a radar to support mobile robot mapping in indoor environments. A range based filtering approach is proposed to reduce the effect of ghost detections and the approach is validated in different scenarios.

ABSTRACT. 2D Laser Range Finders, or 2D-LRFs, are essential sensors in the field of robotics, providing accurate range measurements with high angular resolution. These sensors can be assembled on top of systems which, by granting additional degrees of freedom to the movement of the LRF, enable the 3D reconstruction of a scene. The reconstruction procedure consists of the concatenation of each scan in a single point cloud representation. To do so, the extrinsic transformation between the LRF and the motorized system, in this case, a Pan-tilt unit, must be known with high accuracy, otherwise, the quality of the 3D reconstructed point clouds is insufficient. In this work, a calibration procedure which determines this transformation is proposed. The method does not require a dedicated marker, which is commonly necessary and has numerous disadvantages. Qualitative inspections show that the proposed method is able to significantly reduce artifacts which typically appear on uncalibrated point clouds. Furthermore, quantitative results demonstrate that the calibrated point cloud represents the geometries present in the scene with much higher accuracy when compared with the uncalibrated point cloud.

ABSTRACT. Robotic platforms are being developed for precision agriculture, to execute repetitive and long term task. Autonomous monitoring, pruning, spraying and harvesting are some of these agricultural tasks, which requires an advanced path planning system aware of maximum robot capabilities (mobile platform and arms), terrain slopes and plant/fruits position. The state of the art path planning systems have two limitations: are not optimized for large regions and the path planning is not aware of agricultural tasks requirements. This work presents two solutions to overcome these limitations. It considers the VGR2TO (Vineyard Grid Map to Topological) approach to extract from a 2D grid map a topological map, to reduce the total amount of memory needed by A* and to reduce path search space. Besides, introduces an extension to the A* algorithm, to unsure a safety path and a maximum distance from the vine trees to enable robotic operations on the trees and its fruits.

ABSTRACT. The use of mobile robots is growing every day. Path planning algorithms are needed to move them, allowing the coordination of several robots and make them travel with the least cost and without collisions. With this emerged the interest in studying some path planning algorithms, in order to better understand the operation of each one when applied in this type of robots. The objective of this paper is to present a state of the art survey of some algorithms of path planning for mobile robots. A brief introduction on mobile robots and trajectory planning algorithms is made. Then, the basis of each algorithm is explained, their relative advantages and disadvantages are presented and areas of application for each of them are mentioned.

ABSTRACT. This paper presents a different strategy for a self-driving car short-term path planning among multiple moving obstacles. The main task is to study and implement a motion planning and execution framework in order to make ATLASCAR2 coexist with other moving obstacle vehicles by avoiding collision and overtake them when necessary and possible. The proposed technique, based on the Model Predictive Control paradigm, solves an optimization problem formulated in terms of cost minimization under constraints. Simulation results demonstrate and verify the feasibility and the usefulness of the method considering different scenarios, opening space for real scenario implementation.

ABSTRACT. In this paper we develop a method to coordinate the deployment of a multi-robot team to reach some locations of interest, so-called primary goals, and to transmit the information from these positions to a static Base Station (BS), under connectivity constraints. The relay positions have to be established for some robots to maintain the connectivity at the moment in which the other robots visit the primary goals. Once every robot reaches its assigned goal, they are again available to cover new goals, dynamically re-distributing the robots to the new tasks. The contribution of this work is a two stage method to deploy the team. Firstly, clusters of relay and primary positions are computed, obtaining a tree formed by chains of positions that have to be visited. Secondly, the order for optimally assigning and visiting the goals in the clusters is computed. We analyze different heuristics for sequential and parallel deployment in the clusters, obtaining sub-optimal solutions in fast time for different number of robots and for a large amount of goals.

ABSTRACT. Robotic grasping in domestic environments is nowadays a topic of intensive research, since the big number of variables and constraints in a household scenario makes it a difficult task to accomplish. This paper addresses the problem of grasp planning in these situations. The first problem to deal with is the selection of the grasp pose for the end-effector - the position and orientation in which it should grasp the desired object. To tackle this issue, this work applies a method to detect grasp poses in point clouds with objects that may be unknown to the robot, followed by an approach to select the grasp candidate in terms of its appropriateness for a given scene. The second part of the grasping task is the motion planning that places the end-effector in the grasp pose. This involves getting a Inverse Kinematics (IK) solution for the goal arm configuration and a trajectory for the arm avoiding any obstacle in the scene. This was addressed by using MoveIt! framework together with an additional tolerance method to compute feasible movements. A comparative study regarding the used motion planners and IK solvers was conducted. We show experimental results of the pipeline, featuring the components mentioned above, which can plan and execute grasps with a success rate that may go up to 90%, depending on the scenario.

ABSTRACT. Powered assistive devices have been playing a major role in gait rehabilitation. This work aims to develop a user-oriented assistive strategy with an EMG-based control using a powered knee orthosis (PKO) to provide assistive commands according to the user’s motion intention tracked by electromyography (EMG) signals. To achieve this goal, the work first comprised the development of a wired EMG acquisition system, the study and implementation of a knee joint torque estimation method, and the development of a real-time controller, which uses the estimated torque and the actuator’s torque measured from the PKO to provide user-oriented assistance in walking. We used a proportional gain method to estimate the knee torque, which required a calibration procedure, allowing to determine the relation between the EMG signal and the actuator’s torque. The EMG-based control was validated with two subject walking in a treadmill. The overall performance of the EMG-based control was in accordance with the expectations since it proved to be functional and time- effective when assisting the user’s movements in walking at different walking speeds. Findings show that the developed assistive strategy can effectively follow the user’s motion intention and has potential for gait rehabilitation of patients with residual muscular strength.

ABSTRACT. Scientific competitions are crucial in the field of service robotics. They foster knowledge exchange and allow teams to test their research in unstandardized scenarios and compare result. Such is the case of RoboCup@Home. In this paper, present a summary of the trending solutions and approaches used in RoboCup@Home. Besides, we discuss the attained achievements and challenges to overcome in relation with the progress required to fulfill the long-term goal of the league.

Hence, considering the current capabilities of the robots and their limitations, we propose a set of milestones to address in upcoming competitions. With this work we lay the foundations towards the creation of roadmaps that can help to direct efforts in testing and benchmarking in robotics competitions.

ABSTRACT. This paper presents a methodology for cooperative load transportation with two quadrotors, given the state variables estimates, based on measurements from motion sensors installed on-board. The proposed controllers and estimators resort to optimal control techniques, namely, Linear Quadractic Regulators (LQRs) and Kalman filters, respectively. The proposed control system is validated both in simulation and experimentally, resorting to a commercially available quadrotor equipped with an Inertial Measurement Unit (IMU), an ultrasound height, vertical and frontal cameras, among other sensors.

ABSTRACT. This paper describes the design and development of an autonomous robotic manipulator with four degrees of freedom. The manipulator is named RACHIE - "Robotic Arm for Collaboration with Humans in Industrial Environment". The idea was to create a smaller version of the industrial manipulators available on the market. The mechanical and electronic components are presented as well as the software algorithms implemented on the robot. The manipulator has as its main goal the detection and organization of cans by color and defects. The robot is able to detect a human operator so it can deliver defective cans by collaborating with him/her on an industrial environment. To be able to perform such task, the robot has implemented a machine learning algorithm, a Haar feature-based cascade classifier, on its vision system to detect cans and humans. On the handler motion, direct and inverse kinematics were calculated and implemented, and its equations are described in this paper. This robot presents high reliability and robustness in the task assigned. It is low-cost as it is a small version of commercial ones, making it optimized for smaller tasks.

ABSTRACT. Teaching complex assembly and maintenance skills to human operators usually requires extensive reading and the help of tutors. In order to reduce the training period and avoid the need for human supervision, an immersive teaching system using spatial augmented reality was developed for guiding inexperienced operators. The system provides textual and video instructions for each task while also allowing the operator to navigate between the teaching steps and control the video playback using a bare hands natural interaction interface that is projected into the workspace. Moreover, for helping the operator during the final validation and inspection phase, the system projects the expected 3D outline of the final product. The proposed teaching system was tested with the assembly of a starter motor and proved to be more effective and intuitive than reading the traditional user manuals. This proof of concept use case served to validate the fundamental technologies and approaches that were proposed to achieve an intuitive and accurate augmented reality teaching application. Among the main challenges were the proper modeling and calibration of the sensing and projection hardware along with the 6 DoF pose estimation of objects for achieving precise overlap between the 3D rendered content and the physical world. On the other hand, the conceptualization of the information flow and how it can be conveyed on-demand to the operator was also of critical importance for ensuring a smooth and intuitive experience for the operator.

ABSTRACT. In the past few years, the number of robots being deployed in society has been continuously increasing. Robots are coming to family houses as personal assistants in domestic tasks and entertainers (e.g. toys) as well as in elderly care, handicap assistance, and nursing centers. The new generation of service robots have now to interact with humans in uncertain environments. For this, the robot needs to localize, engage and identify the target subject. The identification of the target could be done in different ways. Image-based face recognition is one example. It is a well-studied problem and state-of-the-art solutions achieve remarkable performance. However, most of the proposed solutions are not adapt to the robot environment. In this work, we explore a new approach to the problem of online person recognition. We present the Recognition-by-Tracking framework that uses pedestrian tracking in order to accumulate evidence about the face identities what leads to more accurate predictions.

ABSTRACT. Robot navigation in human-populated environments is a subject of great interest among the international scientific community. In order to be accepted in these scenarios, it is important for robots to navigate respecting social rules. Avoid getting too close to a person, not interrupting conversations or asking for permission or collaboration when it is required by social conventions, are some of the behaviors that robots must exhibit. This paper presents a social navigation system that integrates different software agents within a cognitive architecture for robots and describes the corpus that allows to establish dialogues between robots and humans in real situations to improve the human-aware navigation system. The corpus has been experimentally evaluated by the simulation of different daily situations, where robots need to plan interactions with real people. The results are analyzed qualitatively, according to the behavior expected by the robot in the interaction performed. The results show how the corpus presented in this paper improves the robot navigation, making it more socially accepted.

ABSTRACT. Mobile robotic applications are increasing everyday in several areas not only in industries but also service robots. The Industry 4.0 promoted even more the digitalization of factories that opened space for smart-factories implementation and for the logistics. Robotic competitions are a key to improve the research and to motivate the learning. This paper addresses a new competition proposal, the Robot@Factory Lite, on the scope of the Portuguese Robotics Open. Beyond the competition, a robot with all its components is proposed and a simulation environment is also provided. To minimize the gap between the simulation and the real implementation, an Hardware-in-the-loop technique is provided that allows to control the simulation with a real arduino board. Results show the same code can control both simulation model and real robot.

ABSTRACT. Robotics competitions are a way to challenge the researchers, roboticists and enthusiastic to address the robot applications. One of the well-known international competition is the micromouse where the fastest mobile robot to solve a maze is the winner. There are several topics addressed in this competition such as robot prototyping, control, electronics, path planning, optimization, among others while keeping the size of the robot as small as possible and as fast as possible . A simulation can be used to speed-up the development and testing algorithms but faces the gap between the reality in the dynamics behaviour. There are some simulation environments that allow to simulate the micromouse competition, but in this paper, an Hardware-in-the-loop simulator tool is presented where the simulated robot is controlled by the same microcontroller used to assembly the robot. By this way, the algorithms are tested and validated with the limitations and constraints presented in the real hardware, such as memory, processing and capabilities. The dynamics of the robot, the slippage of the wheels, the friction and the 3D visualization are presented in the simulator. The presented results show the same code and hardware can control the simulated and the real robot identically.

ABSTRACT. People tracking is a basic capability in almost any robotic application. So it is in robotic competitions, where many robot skills rely on this ability. This problem is still challenging, particularly when are implemented using low definition sensors as LIDARs in crowded environments. This paper describes a solution based on a single laser sensor that uses the gait to keep a continuous identification in time and space of the individual. The system described in this article is based on PeTra (PEople TRAcking) package, which uses convolutional neural networks to identify legs in populated environments. Experimental validation proposes a test in an apartment replicating realistic competition arena.

ABSTRACT. We present Scratchy---a modular, lightweight robot built for low budget competition attendances. Its base is mainly built with standard 4040 aluminium profiles and the robot is driven by four mecanum wheels on brushless DC motors. In combination with a laser range finder we use estimated odometry -- which is calculated by encoders -- for creating maps using a particle filter. A RGB-D camera is utilized for object detection and pose estimation. Additionally, there is the option to use a 6-DOF arm to grip objects from an estimated pose or generally for manipulation tasks. The robot can be assembled in less than one hour and fits into two pieces of hand luggage or one bigger suitcase. Therefore, it provides a huge advantage for student teams that participate in robot competitions like the European Robotics League or RoboCup. Thus, this keeps the funding required for participation, which is often a big hurdle for student teams to overcome, low.

ABSTRACT. The field of automated learning has been steadily growing in robotic tasks. This phenomenon is supported by the evolution of computational resources and new reinforcement learning algorithms. Researchers have drawn their attentions to methods that are easy to implement and tune, while achieving state-of-the-art performance. This trend also affects the world of robotic soccer, where new papers delve systematically into the optimization of basic skills. However, when learning higher-level strategies, there is space for improvement on two fronts. First, the simulation environment should allow the agent to abstract from low-level details. Second, the existing methods to train this kind of behaviors are still scarce. This paper contributes with innovative problem-solving methods, specifically in the rewards field. To test alternative approaches, an extended version of the RoboCup's official Soccer Server simulator was used. The results have confirmed the importance of the proposed reward components and their relationship with the episodes' initial conditions.

ABSTRACT. A solution for an optimal robotic soccer strategy is yet to be found. Multiple agents interacting in an environment with continuous state and action spaces is a recipe for machine learning lethargy. Fortunately, on the one hand, due to the increasing hardware performance and new algorithms, the area of reinforcement learning is growing considerably. On the other hand, simulating the environment for strategy purposes is not following this trend. Several simulators are available, including those used in major soccer competitions (e.g. RoboCup). However, no option combines a good repository of teams with a simple command set that abstracts low-level actions. To clarify this problem, we surveyed the most promising simulators and proposed an extension for the well-established Soccer Server. The objective was to simplify the process of learning strategy-related behaviors through automated optimization algorithms. The results have shown a clear advantage in using the extension to improve the agent's performance. These results were confirmed through an ablation study which emphasized the most important components of the proposed extension. This work contributes to the development of future strategies related with RoboCup or other soccer competitions. Despite the good results, there is space for improvement in computational efficiency and behavior diversity.