Search

Human-machine interfaces (HMIs) have received an increased attention during the past decades, specially due to the use of robots in many fields of the industry. In this paper, we employ a hand gesture recognition (HGR) system and the inertial measurement unit (IMU) integrated in the Myo Armband sensor as a HMI to control the position and orientation of a virtual six-degree-of-freedom (DoF) UR5 robot. As part of the HMI, this work also focuses on solving the trajectory tracking problem for the UR5 robot using two different control strategies: a minimum norm PID and a controller based on linear algebra. Both controllers are designed from the UR5 kinematic model. Finally, we test and compare their performance by computing their IAE performance criterion and the position errors obtained from two experiments: tracking pre-defined input trajectories and commanding the virtual robot by the HGR system plus the Myo Armband's IMU signals.

A practical testbed is crucial to reduce the gap between theory and real-world applications. Path-planning and collision prevention routines are essential for a reliable operation of these multi-robot platforms, so increasingly complex control algorithms can be proposed, designed and tested. This paper describes the recent progress made on the functionality of a mini-sized multi-robot testbed that runs based on ROS (Robotic Operating System). A high-level path planner, a collision prevention routine and motion controllers are proposed as part of the expansion to more complex applications that can run on the testbed. A PID-based orientation controller and an Inverse Kinematics position controller are designed and compared as viable solutions for following the waypoints generated by the high-level path planner. Improvements on software modularity, scalability and the addition of an extra mobile agent also comprise the main improvements for this new iteration of the testbed. The results of the path-planning routine, as well as the performance of the motion controllers are presented and discussed.

Systems used for pattern recognition are usually divided into 4 stages: signal acquisition, preprocessing, feature extraction and classification. However, the use of algorithms in these last 3 stages is not justified and researchers use them without criteria other than the result achieved at the end of the process. In this paper we propose a new methodology and show its particular application to the recognition of five hand gestures based on 8 channels of Electromyography using the Myo armband device placed on the forearm. If n features are extracted, they will form clusters of points in n-dimensional space and now the selection of the best preprocessing algorithms and the best features are based on maximizing the distance among clusters. On the other hand, both intrapersonal and interpersonal variability are treated to facilitate the understanding of the phenomenon. As a demonstration, it was applied to 12 people and the recognition accuracy was 97%.

Unmanned Aerial Vehicles (UAVs) have become one of the most promising technologies in our days due to their outstanding autonomous capabilities, versatility and flexible operation. However, the autonomy offered by these systems limits their use in long-range applications. Also, the efficiency of the UAV propulsion systems remains low even though most of its components have evolved. The two main reason for these issues are: the energy density of the available batteries is quite low, and the propulsion system is not always properly sized according to the needs of the application. Therefore, increasing the flight time of the UAV by enhancing the efficiency of the propulsion systems is still a major challenge. In this sense, this paper aims for the design and implementation of a test-bench for propulsion system essays. Based on a mobile structure, it obtains thrust measures produced by the pair motor-propeller of a UAV. Moreover, by measuring the electric power consumption, the test-bench computes the energy efficiency of a given propulsion system configuration. The system is tested using the propulsion system components of a Skywalker X8 Fixed Wing UAV for two different configurations of the motor-propeller sets. Also, efficiency curves are plotted for each pair along with speed and power measures.

Hand gesture recognition (HGR) based on electromyography (EMG) has been a research topic of great interest in recent years. Designing an HGR to be robust enough to the variation of EMGs is a challenging problem and most of the existing studies have explored supervised learning to design HGRs methods. However, reinforcement learning, which allows an agent to learn online while taking EMG samples, has barely been investigated. In this work, we propose a HGR system composed of the following stages: pre-processing, feature extraction, classification and post-processing. For the classification stage, we use Q-learning to train an agent that learns to classify and recognize EMGs from five gestures of interest. At each step of training, the agent interacts with a defined environment, obtaining thus a reward for the action taken in the current state and observing the next state. We performed experiments using a public EMGs dataset, and the results were evaluated for user-specific HGR models by using a method that is robust to the rotations of the EMG bracelet device. The results showed that the classification accuracy reach up to 90.78% and the recognition up to 87.51% for two different test-sets for 612 users in total. The results obtained in this work show that reinforcement learning methods such as Q-learning can learn a policy from online experiences to solve both the hand gesture classification and the recognition problem based on EMGs.

This work focuses on two main problems in the field of robotic manipulators: modeling and non-linear robust control for trajectory tracking. The kinematic and dynamic model of a three degree of freedom (DoF) commercial manipulator is summarized. Through 3D simulation, the robot is implemented in a virtual environment using Autodesk Inventor and Matlab-Simulink in order to have a robotic computational platform for didactic and research purposes. Using the virtual robot, a non-linear controller is implemented for each joint and the tracking response of the tip to a sinusoidal input is verified. This controller presents a good performance when the manipulator parameters are known with accuracy, so the non-linearities are canceled. However, it is more challenging to consider a real case in which the manipulator holds an object that changes its dynamics. Therefore, an adaptive algorithm is designed and implemented for this case when the cancellation of non-linearities is not exact. This algorithm is based on Lyapunov stability to generate a control law that allows the estimation of the link masses and allowing a better cancellation of non-linearities. This control scheme provides a better performance compared with the case of a non-linear controller but not adaptive.

This paper details the design and implementation of an experimental testbed of a group of three mini-sized mobile robots focused on trajectory tracking formation control. A cascade architecture is proposed which consists of an internal speed control in series with an external controller that helps to maintain the desired formation and posture of the group at all times. The pose of each robot is tracked based on artificial vision and the ArUco reference markers. This monitoring component together with the communication and control systems are implemented on ROS (Robotic Operating System) which facilitates the interconnection between these modules. The results using different predefined trajectories demonstrate the robustness of the controller on maintaining the formation while tracking the desired path.

Cargo-bearing unmanned aerial vehicles (UAVs) have tremendous potential to assist humans by delivering food, medicine, and other supplies. For time-critical cargo delivery tasks, UAVs need to be able to quickly navigate their environments and deliver suspended payloads with bounded load displacement. As a constraint balancing task for joint UAV-suspended load system dynamics, this task poses a challenge. This article presents a reinforcement learning approach for aerial cargo delivery tasks in environments with static obstacles. We first learn a minimal residual oscillations task policy in obstacle-free environments using a specifically designed feature vector for value function approximation that allows generalization beyond the training domain. The method works in continuous state and discrete action spaces. Since planning for aerial cargo requires very large action space (over 106 actions) that is impractical for learning, we define formal conditions for a class of robotics problems where learning can occur in a simplified problem space and successfully transfer to a broader problem space. Exploiting these guarantees and relying on the discrete action space, we learn the swing-free policy in a subspace several orders of magnitude smaller, and later develop a method for swing-free trajectory planning along a path. As an extension to tasks in environments with static obstacles where the load displacement needs to be bounded throughout the trajectory, sampling-based motion planning generates collision-free paths. Next, a reinforcement learning agent transforms these paths into trajectories that maintain the bound on the load displacement while following the collision-free path in a timely manner. We verify the approach both in simulation and in experiments on a quadrotor with suspended load and verify the method's safety and feasibility through a demonstration where a quadrotor delivers an open container of liquid to a human subject. The contributions of this work are two-fold. First, this article presents a solution to a challenging, and vital problem of planning a constraint-balancing task for an inherently unstable non-linear system in the presence of obstacles. Second, AI and robotics researchers can both benefit from the provided theoretical guarantees of system stability on a class of constraint-balancing tasks that occur in very large action spaces.

Hand gesture recognition consists of identifying the class among a set of classes of a hand movement given. Surface electromyography (sEMG) measures the electrical activity generated by voluntary contractions of skeletal muscles. The performance of a recognition system is affected significantly by the orientation of the armband. This orientation could change every time that the user wears the armband. In this paper, a novel technique to improve the robustness in a recognition system with variation in the orientation of the armband is proposed. To test the performance of the proposed model, 4 experiments at recognizing 6 hand gestures are executed. In these experiments the proposed method shows a recognition accuracy of 92.4% versus 59.5%, which corresponds to the accuracy of a traditional recognition model without the correction of orientation.

In the last years, the growing interest in consensus and coordinated control of multi-agent mobile systems has originated the development of a variety of testbeds. These experimental platforms constitute an essential component in the research and education process in the robotics field. This paper describes the design and implementation of a testbed based on the position control and trajectory tracking of mini-sized ground robots. The testbed architecture consists of a monitoring system composed by a web cam and a control algorithm running in a main computer that communicates with the robots wirelessly. A unique feature of the testbed is the use of ArUco markers on top of the mobile platforms for motion tracking purposes using OpenCV and Visual Basic. Additionally, the testbed has a user graphic interface that allows to manage and visualize results of the robot behavior for the tracking of different pre-defined trajectories.

This work summarizes the development of an automatic tracking system for an Unmanned Aerial Vehicle (UAV) realized in cooperation with Latitude Aerospace Solutions (LAS). This system, in joint action with the use of directional antennas, allows increasing the range of communication (telemetry) between the UAV and its ground monitoring station. The tracking system is controlled through a Pixhawk autopilot in which the parameters of a PID controller, designed based on the direct synthesis method, are entered. Additionally, a long range video link is implemented from the drone to the ground monitoring station, which is useful, for example, for a video-surveillance system. Experimental results allow corroboration of the correct performance of the tracking system and the communication links.

International audience; This chapter introduces a series of definitions, metrics, and theory related to heterogeneity and resilience in the context of multi-robotic systems. It presents a case study based on the containment formation control problem for a group of homogeneous and heterogeneous agents. The chapter considers that system failures or adversarial attacks affect a robot such that it communicates incorrect information and on a global level prevents achievement of the team's task. It discusses how heterogeneity has been defined or categorized in the context of multi-agent robotic systems. The variety and disparity capabilities of heterogeneous agents greatly increase their resilience and ability to successfully accomplish complex tasks. The effects of incorporating heterogeneity and/or resilience as design variables of a multi-agent system have been mostly unexplored in the literature. A better understanding of how heterogeneity can be analyzed can help to find the range of missions that an intelligent multi-agent system can or cannot accomplish.

This work presents the design, developing and simulation of a Null-Space based robust controller for a quadcopter formation in a windy environment. The proposed controller is used to perform trajectory tracking of a formation of three quadcopters through wind gusts combining the advantages of the Null-Space based controller with a SMC controller, allowing to merge the capabilities of these two control structures; the flexibility of the Null-Space Controller and the Robustness of the SMC controller. The proposed control system is validated through simulations and with the Lyapunov stability criteria analysis.

This paper details a proof-of-concept development of an adaptive staircase system-type capable of user-specific mechanical reconfigurations actuated by facial-, object-, and voice-recognition. The system is described via two variation-prototypes - developed at Technology Readiness Level 4 - as instances of the same system-type. Accordingly, each prototype is informed by the same use-case considerations and requirements. Nevertheless, by means of their mechanical particulars, advantages and disadvantages specific to each variation are identified and explored. The present adaptive staircase system-type consists of two main components, one computational and the other mechanical. The computational component is built upon an inherited System Architecture previously developed and implemented by the authors. More specifically, the computational component uses Google's TensorFlow for facial-recognition; BerryNet for multi-object detection; and VoiceIt for voice-recognition. These three cloud-compatible, -based, or -dependent recognition mechanisms are used to ascertain the identity three user-types: (1) a person without perceivable physical disabilities; (2) a person reliant on a walking-cane; and (3) a person on a wheelchair. With the exception of the first case, the computational component proceeds to actuate mechanical transformations pertinent to each variety of disabilities depending on which user-type is identified. The objective of this implementations is to present an intuitive and automated vertical mobility solution capable of supporting users with varying degrees of reduced mobility.

Wetlands in the Andean region have been altered and have been lost as a result of the agricultural frontier expansion and human activities. The disturbance of the paramo ecosystem by the destruction or alteration of the wetlands modifies the load and endowment of water to the hydrological systems, which provide water to main cities in the highlands. Therefore, the present work focuses on setting up the framework for wetland monitoring in the Andean paramo region using Unmanned Aerial Vehicles (UAVs). For this aim, the study was based on a mission profile using a fixed wing UAV incorporated with a RGB camera in one of the most documented wetlands in the Ecuadorian paramo region, Pugllohuma wetland. Furthermore, to assess the saturation of the wetland, field testing data has been collected to set the range values of saturation for the monitoring system. In the same way, a review regarding multispectral imagery for the assessment of water and vegetation indices is explored and highlighted for future work. This work is a first stage in the monitoring process and hence it aims to set a baseline study for the implementation of a more detailed methodology.

To meet the academic objectives for courses related with Control Systems, it is important to have access to real educational platforms that help to extend the theory to a hands-on experience. This paper introduces the implementation of an experimental testbed used to control the orientation of a multirotor in an interactive fashion. The testbed is implemented based on the configuration of a quadcopter by restricting its ability to fly, but allowing it to rotate around its three main axes. The orientation model of a quadcopter is also presented focusing on the experimental estimation of the thrust and drag coefficients. A PID-based attitude controller is designed and discretized in order to implement it in an embedded system. Simulation results and initial experimental tests are included showing the response of the multi-rotor experimental testbed and the performance of the controller.

Distributed autonomous robotic systems (DARS) can provide essential support to human task forces in a variety of missions. Maintaining reliable communications on these mobile networks is critically important for cooperative autonomy. Radio-frequency is the common wireless communication technology employed for task coordination. However, it has some limitations that can be mitigated by complementing radio-frequency systems with optical wireless communications. In this chapter, we present our efforts for the development and implementation of an optical wireless communication link in DARS, more specifically between aerial (e.g., quadrotors) and ground robots. We describe a line-of-sight directed optical link between these two platforms. In addition, we detail our strategy to maintain an adequate transmitter-receiver relative position to enhance the optical signal strength. Furthermore, we provide some details about our optical transceiver prototype that can be mounted on small aerial and ground autonomous robotic systems.

Group of robots have become important platforms in applications such as search and rescue, active mapping, and load transportation. Maintaining the group formation while tracking prescribed trajectories is critical in these applications. The performance of the formation can be compromised due to time-delays present in the natural response of the internal actuators. Furthermore, the stability of the system can be compromised. In this work, we proposed a cascade control scheme to tackle an input delay in the kinematics of the mobile robots. An inner velocity control loop is designed based on an approximation for the input delay, while an outer loop takes care of maintaining the shape and posture of the formation. We validate our approach by comparing the performance of our control methodology in simulation with a standard PI and a sliding controller under the assumption of different values input delays.

This paper describes a framework for controlling a heterogeneous wireless robotic network consisting of aerial and ground vehicles. By use of the term heterogeneous, we imply the synergy of multiple robotic platforms characterized by different dynamics and specialized sensing capabilities. Two main scenarios concerning wireless communications are presented: 1) a decentralized connectivity strategy in which a mesh of ground mobile routers swarms in a cluttered environment maintaining communication constraints based on spring-mass virtual physics, potential functions, and routing optimization and 2) an autonomous communications relay in GPS-denied environments via antenna diversity and extremum-seeking SNR optimization. For both scenarios, we validate the proposed methodologies by numerical simulations and experiments. One important feature of our test bed is that it can be used for both indoor and outdoor operations.

In this paper we consider the problem of pairing a ground sensor with an aerial vehicle, both equipped with a hybrid communication system - radio frequency for non-line of sight transmission and optical for line-of-sight transmission. These communication technologies are complementary and by coordinating them it is possible to mitigate their respective weaknesses. A challenging problem is positioning the flying robot within optical communication range, especially when the distance is large and the sensor location is unknown. In this work, we propose a solution to the problem of autonomously localizing the sensor node relative to the aerial vehicle. We take advantage of the hybrid communication scheme by developing a control strategy that uses the radio signal to guide the aerial platform to the sensor node. Once the optical-based signal strength is over a desired threshold the robot hovers within optical range. The control strategy is demonstrated through simulations that incorporate a realistic model for the hybrid communication link.

By taking advantage of complementary communication technologies, distinct sensing functionalities, and different motion dynamics of a heterogeneous multirobotic system, one can accomplish effectively a mission objective. Furthermore, an adequate coordination of the robotic agents is an important factor to fully exploit the unique capabilities that these systems can offer. In this chapter we focus on a multirobotic network made by aerial relays and ground sensors that is deployed to sense areas of interest in an environment populated with obstacles. We use potential field methods to guide the ground mobile sensors to target regions. On the other hand, antenna diversity and reinforcement learning are combined to control the flying relays such that they increase the received signal strength among them and the mobile sensors. We demonstrate the effectiveness of our control strategy through simulation tests contrasting independent and cooperative learning.

Teams of agents with different skills can solve critical missions by efficiently joining their complementary abilities. One critical step to exploit the distinct resources available on a set of agents is to form a coalition, i.e., an alliance that satisfies the requirements imposed by a mission. In this work, we represent the relation between agent capabilities and required resources for executing a given task by a weighted bipartite graph. Using this graph, we find an assignment between agents and resource capabilities such that the total weight of capabilities is maximized. From this assignment, also known as matching in Graph Theory, we compute a coalition of agents whose total resource capabilities can satisfy the task resource requirements. Finally, we measure the heterogeneity of the computed coalition and analyze how it is affected by the task constraints and the amount of resources present in the agents.

Autonomous multi-rotor aerial vehicles, specially quadrotors, have become popular platforms for the transportation of cable-suspended loads. Before transporting the load, the lift maneuver is a crucial step that needs to be planed. In order to perform this essential maneuver, we decompose it into simpler hybrid modes which characterize the dynamics of the quadrotor-load system in specific regimes during the maneuver. In this work, we represent the maneuver as a hybrid system and show that it is differentially flat. This property facilitates the generation of a prescribed trajectory and the design of a trajectory tracking controller based on geometric control. Numerical simulations show promising results on the performance of the proposed control architecture.

In this paper we propose an application of adaptive synchronization of chaos to detect changes in the topology of a mobile robotic network. We assume that the network may evolve in time due to the relative motion of the mobile robots and due to unknown environmental conditions, such as the presence of obstacles in the environment. We consider that each robotic agent is equipped with a chaotic oscillator whose state is propagated to the other robots through wireless communication, with the goal of synchronizing the oscillators. We introduce an adaptive strategy that each agent independently implements to: (i) estimate the net coupling of all the oscillators in its neighborhood and (ii) synchronize the state of the oscillators onto the same time evolution. We show that by using this strategy, synchronization can be attained and changes in the network topology can be detected. We go one step forward and consider the possibility of using this information to control the mobile network. We show the potential applicability of our technique to the problem of maintaining a formation between a set of mobile platforms, which operate in an inhomogeneous and uncertain environment. We discuss the importance of using chaotic oscillators and validate our methodology by numerical simulations., 17 pages, 10 figures, accepted for publication in Physica D

In this paper, we are interested in exploiting the heterogeneity of a robotic network made of ground and aerial agents to sense multiple targets in a cluttered environment. Maintaining wireless communication on this type of networks is fundamentally important specially for cooperative purposes. The proposed heterogeneous network consists of ground sensors, e.g., OctoRoACHes, and aerial routers, e.g., quadrotors. Adaptive potential field methods are used to coordinate the ground mobile sensors. Moreover, a reward function for the aerial mobile wireless routers is formulated to guarantee communication coverage among the ground sensors and a fixed base station. A sub-optimal controller is proposed based on an approximate control policy iteration technique. Simulation results of a case study are presented to illustrate the proposed methodology.

Attaining autonomous flight is an important task in aerial robotics. Often flight trajectories are not only subject to unknown system dynamics, but also to specific task constraints. This paper presents a motion planning method for generating trajectories with minimal residual oscillations (swing-free) for rotorcraft carrying a suspended loads. We rely on a finite-sampling, batch reinforcement learning algorithm to train the system for a particular load. We find criteria that allow the trained agent to be transferred to a variety of models, state and action spaces and produce a number of different trajectories. Through a combination of simulations and experiments, we demonstrate that the inferred policy is robust to noise and the unmodeled dynamics of the system. The contributions of this work are 1) applying reinforcement learning to solve the problem of finding swing-free trajectories for rotorcraft, 2) designing a problem-specific feature vector for value function approximation, 3) giving sufficient conditions for successful learning transfer to different models, state and action spaces, and 4) verification of the resulting trajectories in both simulation and autonomous control of quadrotors with suspended loads.

In this paper, we address the problem of agile swing-free trajectory tracking of a quadrotor with a suspended load. This problem has great practical significance in many UAV applications. However, it has received little attention in the literature so far. Flying with a suspended load can be a very challenging and sometimes hazardous task as the suspended load significantly alters the flight characteristics of the quadrotor. In order to deal with this problem, we propose a technique based on dynamic programming which ensures swing-free trajectory tracking. We start by presenting the mathematical model of a quadrotor with suspended load dynamics and kinematics. A high-level planner is used to provide desired waypoints, and then a dynamic programming approach is used to generate the swing-free trajectory for the quadrotor carrying a suspended load. Effectiveness of this method is demonstrated by numerical simulations and experiments.

In the past few decades, unmanned aerial vehicles (UAVs) have become promising mobile platforms capable of navigating semiautonomously or autonomously in uncertain environments. The level of autonomy and the flexible technology of these flying robots have rapidly evolved, making it possible to coordinate teams of UAVs in a wide spectrum of tasks. These applications include search and rescue missions ; disaster relief operations, such as forest fires [1] ; and environmental monitoring and surveillance. In some of these tasks, UAVs work in coordination with other robots, as in robot-assisted inspection at sea [2]. Recently, radio-controlled UAVs carrying radiation sensors and video cameras were used to monitor, diagnose, and evaluate the situation at Japans Fukushima Daiichi nuclear plant facility [3].