Showing papers in "Journal of Intelligent and Robotic Systems in 2019"
TL;DR: This is the first work that addresses the continuous UAV landing maneuver on a moving platform by means of a state-of-the-art deep reinforcement learning algorithm, trained in simulation and tested in real flights.
Abstract: The use of multi-rotor UAVs in industrial and civil applications has been extensively encouraged by the rapid innovation in all the technologies involved. In particular, deep learning techniques for motion control have recently taken a major qualitative step, since the successful application of Deep Q-Learning to the continuous action domain in Atari-like games. Based on these ideas, Deep Deterministic Policy Gradients (DDPG) algorithm was able to provide outstanding results with continuous state and action domains, which are a requirement in most of the robotics-related tasks. In this context, the research community is lacking the integration of realistic simulation systems with the reinforcement learning paradigm, enabling the application of deep reinforcement learning algorithms to the robotics field. In this paper, a versatile Gazebo-based reinforcement learning framework has been designed and validated with a continuous UAV landing task. The UAV landing maneuver on a moving platform has been solved by means of the novel DDPG algorithm, which has been integrated in our reinforcement learning framework. Several experiments have been performed in a wide variety of conditions for both simulated and real flights, demonstrating the generality of the approach. As an indirect result, a powerful work flow for robotics has been validated, where robots can learn in simulation and perform properly in real operation environments. To the best of the authors knowledge, this is the first work that addresses the continuous UAV landing maneuver on a moving platform by means of a state-of-the-art deep reinforcement learning algorithm, trained in simulation and tested in real flights.
141 citations
TL;DR: A fully-autonomous aerial robotic solution, for executing complex SAR missions in unstructured indoor environments, based on the combination of a complete hardware configuration and a flexible system architecture which allows the execution of high-level missions in a fully unsupervised manner.
Abstract: Search and Rescue (SAR) missions represent an important challenge in the robotics research field as they usually involve exceedingly variable-nature scenarios which require a high-level of autonomy and versatile decision-making capabilities. This challenge becomes even more relevant in the case of aerial robotic platforms owing to their limited payload and computational capabilities. In this paper, we present a fully-autonomous aerial robotic solution, for executing complex SAR missions in unstructured indoor environments. The proposed system is based on the combination of a complete hardware configuration and a flexible system architecture which allows the execution of high-level missions in a fully unsupervised manner (i.e. without human intervention). In order to obtain flexible and versatile behaviors from the proposed aerial robot, several learning-based capabilities have been integrated for target recognition and interaction. The target recognition capability includes a supervised learning classifier based on a computationally-efficient Convolutional Neural Network (CNN) model trained for target/background classification, while the capability to interact with the target for rescue operations introduces a novel Image-Based Visual Servoing (IBVS) algorithm which integrates a recent deep reinforcement learning method named Deep Deterministic Policy Gradients (DDPG). In order to train the aerial robot for performing IBVS tasks, a reinforcement learning framework has been developed, which integrates a deep reinforcement learning agent (e.g. DDPG) with a Gazebo-based simulator for aerial robotics. The proposed system has been validated in a wide range of simulation flights, using Gazebo and PX4 Software-In-The-Loop, and real flights in cluttered indoor environments, demonstrating the versatility of the proposed system in complex SAR missions.
125 citations
TL;DR: Mona is a low-cost, open-source and open-hardware mobile robot, which has been developed to be compatible with a number of standard programming environments and has been successfully used for both education and research at The University of Manchester.
Abstract: Mobile robots are playing a significant role in Higher Education science and engineering teaching, as they offer a flexible platform to explore and teach a wide-range of topics such as mechanics, electronics and software. Unfortunately the widespread adoption is limited by their high cost and the complexity of user interfaces and programming tools. To overcome these issues, a new affordable, adaptable and easy-to-use robotic platform is proposed. Mona is a low-cost, open-source and open-hardware mobile robot, which has been developed to be compatible with a number of standard programming environments. The robot has been successfully used for both education and research at The University of Manchester, UK.
100 citations
TL;DR: A novel algorithm called Borderland is proposed, which uses the check and repair approach to rapidly identify and adjust only the portion of path involved by the inception of relevant dynamical changes in the risk factor, which aims at minimizing the risk.
Abstract: This paper presents a risk-aware path planning strategy for Unmanned Aerial Vehicles in urban environments. The aim is to compute an effective path that minimizes the risk to the population, thus enforcing safety of flight operations over inhabited areas. To quantify the risk, the proposed approach uses a risk-map that associates discretized locations of the space with a suitable risk-cost. Path planning is performed in two phases: first, a tentative path is computed off-line on the basis on the information related to static risk factors; then, using a dynamic risk-map, an on-line path planning adjusts and adapts the off-line path to dynamically arising conditions. Off-line path planning is performed using riskA*, an ad-hoc variant of the A* algorithm, which aims at minimizing the risk. While off-line path planning has no stringent time constraints for its execution, this is not the case for the on-line phase, where a fast response constitutes a critical design parameter. We propose a novel algorithm called Borderland, which uses the check and repair approach to rapidly identify and adjust only the portion of path involved by the inception of relevant dynamical changes in the risk factor. After the path planning, a smoothing process is performed using Dubins curves. Simulation results confirm the suitability of the proposed approach.
85 citations
TL;DR: Highlights of the trade-off between autonomy and communication requirements are provided and followed by an overview of promising communication and networking technologies that could encourage the integration of autonomous systems in maritime scenarios.
Abstract: The rapid development of autonomous systems and Information and Communications Technologies (ICT) create new opportunities for maritime activities. Existing autonomous systems are becoming more powerful and utilise the capabilities of several types of devices such as Autonomous Underwater Vehicles (AUVs), Unmanned Surface Vehicles (USVs) – sometimes referred as Autonomous Surface Vehicles (ASVs) –, Unmanned Aerial Vehicles (UAVs), moored and drifting systems and, recently emerging, autonomous vessels. Their importance in providing new services in maritime environments is undeniable and the opportunity for coordinated and interconnected operations is clear. However, continuous wide integration of various technologies in maritime environments still faces many challenges. Operations may take place in remote locations, so that dependence on third-party infrastructures such as satellite communication or terrestrial communication systems must be expected. The reliability, performance, availability, and cost of such systems are important issues that need to be tackled. This work reviews the major advancements on state-of-the-art autonomous maritime vehicles and systems, which are used in several different scenarios, from scientific research to transportation. Moreover, the paper highlights how available technologies can be composed in order to efficiently and effectively operate in maritime environments. Highlights of the trade-off between autonomy and communication requirements are provided and followed by an overview of promising communication and networking technologies that could encourage the integration of autonomous systems in maritime scenarios.
84 citations
TL;DR: A novel backstepping & fuzzy sliding mode controller (BFSMC) is proposed for trajectory tracking of the DWMR in the presence of model uncertainties and external disturbances and numerical simulation shows that the BFSMC has the better accuracy, rapidity, smoothness and robustness, when compared to the conventional SMC.
Abstract: Differential wheeled mobile robot (DWMR) is a typical nonholonomic complex system with the practical importance and theoretically interesting properties. A novel backstepping & fuzzy sliding mode controller (BFSMC) is proposed for trajectory tracking of the DWMR in the presence of model uncertainties and external disturbances. Backstepping control technique is used to eliminate the pose deviations of the mobile robot based on the kinematic model. Sliding mode control is adopted at the dynamic level for velocity tracking of the driving wheels, in which the gain of switching control is adjusted adaptively by means of fuzzy logic inference, in order to mitigate the chattering problem. The tracking error convergence of the BFSMC is demonstrated by means of the Lyapunov stability criteria. Numerical simulation shows that the BFSMC has the better accuracy, rapidity, smoothness and robustness, when compared to the conventional SMC. A vision-guided mobile robot with an onboard camera is developed for the experiment of path tracking. The experimental results further validate the feasibility and effectiveness of the BFSMC.
69 citations
TL;DR: A real-time 3D path planning solution for multirotor aerial robots to obtain a feasible, optimal and collision-free path in complex dynamic environments is presented.
Abstract: Deliberative capabilities are essential for intelligent aerial robotic applications in modern life such as package delivery and surveillance. This paper presents a real-time 3D path planning solution for multirotor aerial robots to obtain a feasible, optimal and collision-free path in complex dynamic environments. High-level geometric primitives are employed to compactly represent the situation, which includes self-situation of the robot and situation of the obstacles in the environment. A probabilistic graph is utilized to sample the admissible space without taking into account the existing obstacles. Whenever a planning query is received, the generated probabilistic graph is then explored by an A⋆ discrete search algorithm with an artificial field map as cost function in order to obtain a raw optimal collision-free path, which is subsequently shortened. Realistic simulations in V-REP simulator have been created to validate the proposed path planning solution, integrating it into a fully autonomous multirotor aerial robotic system.
66 citations
TL;DR: The efficiency of complete coverage path planning of autonomous underwater vehicle (AUV) is high with short path planning time and low overlapping coverage rate by using the algorithm proposed in this paper.
Abstract: For the shortcomings of biologically inspired neural network algorithm in the path planning of robots, such as high computational complexity, long path planning time etc Glasius Bio-inspired Neural Network (GBNN) algorithm is proposed to improve the algorithm, and applied to the complete coverage path planning of autonomous underwater vehicle (AUV) Firstly, the grid map is constructed by discretizing the two-dimensional underwater environment Secondly, the corresponding dynamic neural network is built on the grid map Finally, complete coverage path of AUV is planned based on the GBNN strategy and the path of AUV at the edge of obstacles is optimized by some typical path templates The simulation results show that the AUV can completely cover the entire workspace and immediately escape from deadlocks without any waiting Meanwhile, the efficiency of complete coverage path planning is high with short path planning time and low overlapping coverage rate by using the algorithm proposed in this paper
65 citations
TL;DR: The Lightweight Rover Unit (LRU) as discussed by the authors is a small and agile rover prototype designed for the challenges of planetary exploration, which has a locomotion system with individually steered wheels for high maneuverability in rough terrain and stereo cameras as its main sensors.
Abstract: Planetary exploration poses many challenges for a robot system: From weight and size constraints to extraterrestrial environment conditions, which constrain the suitable sensors and actuators. As the distance to other planets introduces a significant communication delay, the efficient operation of a robot system requires a high level of autonomy. In this work, we present our Lightweight Rover Unit (LRU), a small and agile rover prototype that we designed for the challenges of planetary exploration. Its locomotion system with individually steered wheels allows for high maneuverability in rough terrain and stereo cameras as its main sensors ensure the applicability to space missions. We implemented software components for self-localization in GPS-denied environments, autonomous exploration and mapping as well as computer vision, planning and control modules for the autonomous localization, pickup and assembly of objects with its manipulator. Additional high-level mission control components facilitate both autonomous behavior and remote monitoring of the system state over a delayed communication link. We successfully demonstrated the autonomous capabilities of our LRU at the SpaceBotCamp challenge, a national robotics contest with focus on autonomous planetary exploration. A robot had to autonomously explore an unknown Moon-like rough terrain, locate and collect two objects and assemble them after transport to a third object – which the LRU did on its first try, in half of the time and fully autonomously. The next milestone for our ongoing LRU development is an upcoming planetary exploration analogue mission to perform scientific experiments at a Moon analogue site located on a volcano.
62 citations
TL;DR: A learning-based fuzzy smoke detection approach intended to achieve an effective and early forest fire detection, while greatly reduce the negative impacts from clouds in the sky, illumination variations, and changes of forest features.
Abstract: Forests are potentially and seriously threatened by fires which have caused huge damages and losses of life and properties every year. In general, it is easier to detect smoke than fire in its early stage. Developing an effective and safe smoke detection method is thereby critical for early forest fire fighting and preventing the fire developing into uncontrollable. This paper presents a learning-based fuzzy smoke detection approach intended to achieve an effective and early forest fire detection, while greatly reduce the negative impacts from clouds in the sky, illumination variations, and changes of forest features. First, a fuzzy-logic based smoke detection rule is designed for detecting and segmenting smoke regions in the visual images captured by the camera onboard an unmanned aerial vehicle (UAV). The differences of each two components of red, green, and blue (RGB) model and intensity in hue, saturation, and intensity (HSI) model of images are chosen as inputs of a fuzzy logic rule, while the smoke likelihood is selected as its output. Then, an extended Kalman filter (EKF) is further employed for reshaping the inputs and output of the fuzzy smoke detection rule on-line. It is expected to provide the smoke detection method with additional regulating flexibility adapting to variations of environmental conditions and reliable automatic detection performance. Next, the morphological operation is also adopted to remove imperfections induced by noises and textures distorted nonconvex/concave segments. Finally, extensive studies on several sets of images containing smoke under distinct environmental conditions are conducted to validate the proposed methodology.
62 citations
TL;DR: This paper proposes to simulate the reflection route by a ray-tracing technique, aided by predicted satellite positions and the widely available 3D building model, and designs a new cost function to consider both the distance to the destination and the positioning error at each grid.
Abstract: The mission of future parcel delivery will be performed by unmanned aerial vehicles (UAVs). However, the localization of global navigation satellite systems (GNSS) in urban areas experiences the notorious multipath effect and non-line-of-sight (NLOS) reception which could potentially generate approximately 50 meters of positioning error. This misleading localization result can be hazardous for UAV applications in GNSS-challenged areas. Due to multipath complexity, there is no general solution to eliminate this effect. A solution to guide UAV operation is to plan an optimal route that smartly avoids the area with a strong multipath effect. To achieve this goal, the impact of the multipath effect in terms of positioning error at different locations must be predicted. This paper proposes to simulate the reflection route by a ray-tracing technique, aided by predicted satellite positions and the widely available 3D building model. Thus, the multipath effect in the pseudorange domain can be simulated using the reflection route and multipath noise envelope according, according to specific correlator designs. By constructing the multipath-biased pseudorange domain, the predicted positioning error can be obtained using a least square positioning method. Finally, the predicted GNSS error distribution of a target area can be further constructed. A new A* path planning algorithm is developed to combine with the GNSS error distribution. This paper designs a new cost function to consider both the distance to the destination and the positioning error at each grid. By comparing the conventional and the proposed path planning algorithms, the planned paths of the proposed methods experienced fewer positioning errors, which can lead to safer routes for UAVs in urban areas.
TL;DR: This paper added to PCL a new algorithm for performing correspondence estimation using lookup tables that was inspired by the PM approach to solve this problem and proved to be 40 to 60 times faster than the existing k-d tree approach in PCL and allowed the Iterative Closest Point algorithm to perform point cloud registration 5 to 9 times faster.
Abstract: The self-localization of mobile robots in the environment is one of the most fundamental problems in the robotics navigation field. It is a complex and challenging problem due to the high requirements of autonomous mobile vehicles, particularly with regard to the algorithms accuracy, robustness and computational efficiency. In this paper, we present a comparison of three of the most used map-matching algorithms applied in localization based on natural landmarks: our implementation of the Perfect Match (PM) and the Point Cloud Library (PCL) implementation of the Iterative Closest Point (ICP) and the Normal Distribution Transform (NDT). For the purpose of this comparison we have considered a set of representative metrics, such as pose estimation accuracy, computational efficiency, convergence speed, maximum admissible initialization error and robustness to the presence of outliers in the robots sensors data. The test results were retrieved using our ROS natural landmark public dataset, containing several tests with simulated and real sensor data. The performance and robustness of the Perfect Match is highlighted throughout this article and is of paramount importance for real-time embedded systems with limited computing power that require accurate pose estimation and fast reaction times for high speed navigation. Moreover, we added to PCL a new algorithm for performing correspondence estimation using lookup tables that was inspired by the PM approach to solve this problem. This new method for computing the closest map point to a given sensor reading proved to be 40 to 60 times faster than the existing k-d tree approach in PCL and allowed the Iterative Closest Point algorithm to perform point cloud registration 5 to 9 times faster.
TL;DR: This paper addresses the problem of cooperative flight path planning where the air vehicles should arrive at the destinations simultaneously or sequentially with specified time delays, while minimizing the total mission time.
Abstract: The engagement of a group of autonomous air vehicles against several targets is a major challenge in mission planning. This paper addresses the problem of cooperative flight path planning where the air vehicles should arrive at the destinations simultaneously or sequentially with specified time delays, while minimizing the total mission time. This involves finding an optimal assignment of air vehicles to targets and generating trajectories in compliance with the kinematic constraints of the vehicles. The trajectories have to avoid nofly-areas, threats and other obstacles, and must prevent the air vehicles from colliding with each other. The presented algorithm for simultaneous arrival first calculates shortest flight paths between all pairs of air vehicles and targets using a network-based routing model. An optimal assignment and a critical path is found by solving a linear bottleneck assignment problem with costs corresponding to the lengths of the shortest paths. The other flight paths are prolongated to the length of the critical path by automatic insertion of waypoints. This is achieved by concatenating subpaths stored in different shortest-path-trees. Due to the special structure of the network, all concatenated flight paths are flyable and feasible. Sequential arrival at a target is realized by sorting the flight paths according to their lengths and prolongating them whenever necessary to accomplish the desired time delays. The capability of the approach is demonstrated by simulation results.
TL;DR: This paper surveys various applications of artificial evolution in the field of modular robots to identify the most promising methods that can lead to developing autonomous adaptive robotic systems that require the minimum task related knowledge on the designer side.
Abstract: This paper surveys various applications of artificial evolution in the field of modular robots. Evolutionary robotics aims to design autonomous adaptive robots automatically that can evolve to accomplish a specific task while adapting to environmental changes. A number of studies have demonstrated the feasibility of evolutionary algorithms for generating robotic control and morphology. However, a huge challenge faced was how to manufacture these robots. Therefore, modular robots were employed to simplify robotic evolution and their implementation in real hardware. Consequently, more research work has emerged on using evolutionary computation to design modular robots rather than using traditional hand design approaches in order to avoid cognition bias. These techniques have the potential of developing adaptive robots that can achieve tasks not fully understood by human designers. Furthermore, evolutionary algorithms were studied to generate global modular robotic behaviors including; self-assembly, self-reconfiguration, self-repair, and self-reproduction. These characteristics allow modular robots to explore unstructured and hazardous environments. In order to accomplish the aforementioned evolutionary modular robotic promises, this paper reviews current research on evolutionary robotics and modular robots. The motivation behind this work is to identify the most promising methods that can lead to developing autonomous adaptive robotic systems that require the minimum task related knowledge on the designer side.
TL;DR: The simulation results show that the proposed mapless motion planning system can guide an underactuated AUV in navigating to its desired targets without colliding with any obstacles.
Abstract: This research is concerned with the motion planning problem encountered by underactuated autonomous underwater vehicles (AUVs) in a mapless environment. A motion planning system based on deep reinforcement learning is proposed. This system, which directly optimizes the policy, is an end-to-end motion planning system. It uses sensor information as input and continuous surge force and yaw moment as output. It can reach multiple target points in a sequence while simultaneously avoiding obstacles. In addition, this study proposes a reward curriculum training method to solve the problem in which the number of samples required for random exploration increases exponentially with the number of steps needed to obtain a reward. At the same time, the negative impact of intermediate rewards can be avoided. The proposed system demonstrates good planning ability for a mapless environment and excellent ability to migrate to other unknown environments. The system also has resistance to current disturbances. The simulation results show that the proposed mapless motion planning system can guide an underactuated AUV in navigating to its desired targets without colliding with any obstacles.
TL;DR: An adaptive dynamic controller to guide an unmanned aerial vehicle (UAV) when accomplishing trajectory tracking tasks can be used as if it were an on-line identification subsystem, since the parameters converge to values that effectively represent the UAV dynamics.
Abstract: This work proposes an adaptive dynamic controller to guide an unmanned aerial vehicle (UAV) when accomplishing trajectory tracking tasks. The controller structure consists of a kinematic controller that generates reference commands to a dynamic compensator in charge of changing the reference commands according to the system dynamics. The final control actions thus generated are then sent to the UAV to make it to track an arbitrary trajectory in the 3D space. The parameters of the dynamic compensator are directly updated during navigation, configuring a directly updated self-tuning regulator with input error, aiming at reducing the tracking errors, thus improving the system performance in task accomplishment. After describing the control system thus designed, its stability is proved using the Lyapunov theory. To validate the proposed system simulations and real experiments were run, some of them are reported here, whose results demonstrate the effectiveness of the proposed control system and its good performance, even when the initial values of the parameters associated to the dynamic model of the UAV are completely unknown. One of the conclusions, regarding the results obtained, is that the proposed system can be used as if it were an on-line identification subsystem, since the parameters converge to values that effectively represent the UAV dynamics.
TL;DR: An admittance controller that generates haptic signals to induce the tracking of a predetermined path to improve the stability in people with lower limb weakness or poor balance and for people that cannot safely use conventional walkers is presented.
Abstract: Smart Walkers are robotic devices that may be used to improve the stability in people with lower limb weakness or poor balance. Such devices may also offer support for cognitive disabilities and for people that cannot safely use conventional walkers. This paper presents an admittance controller that generates haptic signals to induce the tracking of a predetermined path. During use, when deviating from such path, the method proposed here varies the damping parameter of an admittance controller by means of a spatial modulation technique, resulting in a haptic feedback, which is perceived by the user as a difficult locomotion in wrong direction. The UFES’s Smart Walker uses a multimodal cognitive interaction composed by a haptic feedback, and a visual interface with two LEDs to indicate the correct/desired direction when necessary. The controller was validated in two experiments. The first one consisted of following a predetermined path composed of straight segments. The second experiment consisted of finding a predetermined path starting from a position outside of such path. When haptic feedback was used, the kinematic estimation error was around 0.3 (± 0.13) m and the force applied to move the walker was approximately 5 kgf. When the multimodal interaction was performed with the haptic and visual interfaces, the kinematic estimation error decreased to 0.16 (± 0.03) m, and the force applied dropped to around 1 kgf, which can be seen as an important improvement on assisted locomotion.
TL;DR: This work is extending CACOC by a Collision Avoidance (CA) mechanism and testing its efficiency in terms of area coverage by the UAV swarm and evaluating the performance of the proposed CAC OC with CA algorithm in an area exploration scenario with 3 UAVs.
Abstract: The recent development of compact and economic small Unmanned Aerial Vehicles (UAVs) permits the development of new UAV swarm applications. In order to enhance the area coverage of such UAV swarms, a novel mobility model has been presented in previous work, combining an Ant Colony algorithm with chaotic dynamics (CACOC). This work is extending CACOC by a Collision Avoidance (CA) mechanism and testing its efficiency in terms of area coverage by the UAV swarm. For this purpose, CACOC is used to compute UAV target waypoints which are tracked by model predictively controlled UAVs. The UAVs are represented by realistic motion models within the virtual robot experimentation platform (V-Rep). This environment is used to evaluate the performance of the proposed CACOC with CA algorithm in an area exploration scenario with 3 UAVs. Finally, its performance is analyzed using metrics.
TL;DR: A novel kinematic formulation for surgical systems is derived and the movement restriction in incision point is resolved by active control of the system through a so-called RCM-constrained Jacobian.
Abstract: This paper presents kinematic control of surgical robotic systems subject to Remote Center of Motion (RCM) constraint in Minimally Invasive Robotic Surgeries (MIRS) A novel kinematic formulation for surgical systems is derived and the movement restriction in incision point, known as RCM constraint, is resolved by active control of the system through a so-called RCM-constrained Jacobian The proposed minimal Jacobian matrix can realize fixed/moving trocar constraint effectively in comparison with the state-of-the-arts In the following, an analysis related to the dexterity of the constrained system is introduced and an index for manipulability of the constrained system is introduced The proposed approach is validated through several numerical simulations as well as experiments in a 7DoFs and 9DoFs MIRS scenarios The results show the efficiency and the precision of the proposed method
TL;DR: The developed augmented reality system for stroke rehabilitation is able to restore the functions of the upper limb and mainly give patients more motivation to undergo the rehabilitation exercises.
Abstract: Motivating game-based training have the potential to improve therapy for people with neurological impairments. In recent years, the serious games have become extremely useful tools in rehabilitation field. They aim to stimulate the mobility of the body through an immersive experience that puts the user in interactive virtual environment. This paper is concerned about developing a customized augmented reality system for stroke rehabilitation. This will be done through integrating an interactive serious game interface with a hand exoskeleton device. This game-based rehabilitation system allows users to carry out physical rehabilitation therapies using a natural user interface based on Kinect’s skeletal tracking features and the electromyography (EMG) sensor. During game playing, the interactive user interface provides useful real-time feedback information such as the time required to grasp a desired dynamic virtual object, and the assigned score and thus the ability of the proposed system to provide a compensatory action regarding the dynamic behavior of the virtual target. The main goal of the developed virtual environment is to create positive influences on the rehabilitation process. Patient movement information and signals obtained from the developed exoskeleton device are used together to monitor the rehabilitation progress. The developed exoskeleton hand is a 3D printed low cost device suitable for grasping tasks that can be used even for domestic stroke patients. The developed exoskeleton device is not only a mechanical system able to perform the rehabilitation act but also it presents an effective tracking and traceability software solution. The EMG signals measured during hand motion are used to detect the intention of hand opening or closing which in turn will actuate the mechanical structure to accomplish the desired task. Parameters and results of patients’ exercises are stored and analyzed when needed to evaluate patients’ progress. The developed system is tested experimentally and it is able to restore the functions of the upper limb and mainly give patients more motivation to undergo the rehabilitation exercises.
TL;DR: A novel robotic platform designed and constructed to facilitate teaching Persian Sign Language (PSL) to children with hearing disabilities, which has a relatively low development cost for a robot in its category.
Abstract: This paper introduces a novel robotic platform, called RASA (Robot Assistant for Social Aims). This educational social robot is designed and constructed to facilitate teaching Persian Sign Language (PSL) to children with hearing disabilities. There are three predominant characteristics from which design guidelines of the robot are generated. First, the robot is designed as a fully functional interactive social robot with children as its social service recipients. Second, it comes with the ability to perform PSL, which demands a dexterous upper-body of 29 actuated degrees of freedom. Third, it has a relatively low development cost for a robot in its category. This funded project, addresses the challenges resulting from the at times divergent requirements of these characteristics. Accordingly, the hardware design of the robot is discussed, and an evaluation of its sign language realization performance has been carried out. The inspected recognition rates of certain signs of PSL, performed by RASA, have also been reported.
TL;DR: Inferences based on the results from simulations and experiments show that the proposed planner is more effective in providing an optimal feasible path as compared to existing methodologies, demonstrating clear advantages for rough, unstructured terrain planning.
Abstract: This paper describes a novel physics-based path planning architecture for autonomous navigation of tracked vehicles in rough terrain conditions. Unlike conventional path planning applications for smooth and structured environments, factors such as slip, slope of the terrain, robot actuator limitations, and dynamics of robot terrain interactions must be considered for rough terrain applications. The proposed path planning method consists of a hybrid planner/simulator, which takes into account all of the above factors by simulating the closed loop motion of the robot with a low-level controller on a realistic terrain model inside a physics engine. Once a feasible path to the goal is obtained, the same low-level closed loop controller is then used to execute the proposed path on the actual robot. The proposed architecture uses the D* Lite algorithm working on a 2D grid representation of the terrain as the high-level planner, Bullet as the physics engine and a hybrid automaton as the low-level closed loop controller. The proposed method is validated both in simulation and through experiments. Inferences based on the results from simulations and experiments show that the proposed planner is more effective in providing an optimal feasible path as compared to existing methodologies, demonstrating clear advantages for rough, unstructured terrain planning. Based on the results, possible improvements to the method are proposed for future work.
TL;DR: This paper presents a high-throughput method for Above Ground Estimation of Biomass (AGBE) in rice using multispectral near-infrared (NIR) imagery captured at different scales of the crop using an Unmanned Aerial Vehicle (UAV).
Abstract: This paper presents a high-throughput method for Above Ground Estimation of Biomass (AGBE) in rice using multispectral near-infrared (NIR) imagery captured at different scales of the crop. By developing an integrated aerial crop monitoring solution using an Unmanned Aerial Vehicle (UAV), our approach calculates 7 vegetation indices that are combined in the form of multivariable regressions depending on the stage of rice growth: vegetative, reproductive or ripening. We model the relationship of these vegetation indices to estimate the biomass of a certain crop area. The methods are calibrated by using a minimum sampling area of 1 linear meter of the crop. Comprehensive experimental tests have been carried out over two different rice varieties under upland and lowland rice production systems. Results show that the proposed approach is able to estimate the biomass of large areas of the crop with an average correlation of 0.76 compared with the traditional manual destructive method. To our knowledge, this is the first work that uses a small sampling area of 1 linear meter to calibrate and validate NIR image-based estimations of biomass in rice crops.
TL;DR: This work presents a method for quantifying the probability of fatalities resulting from an uncontrolled descent of an unmanned aircraft conducting a BVLOS flight, based on a standard stochastic model, and employs a parameterized high fidelity ground impact distribution model.
Abstract: One of the major challenges of conducting operations of unmanned aircraft, especially operations beyond visual line-of-sight (BVLOS), is to make a realistic and sufficiently detailed risk assessment. An important part of such an assessment is to identify the risk of fatalities, preferably in a quantitative way since this allows for comparison with manned aviation to determine whether an equivalent level of safety is achievable. This work presents a method for quantifying the probability of fatalities resulting from an uncontrolled descent of an unmanned aircraft conducting a BVLOS flight. The method is based on a standard stochastic model, and employs a parameterized high fidelity ground impact distribution model that accounts for both aircraft specifications, parameter uncertainties, and wind. The method also samples the flight path to create an almost continuous quantification of the risk as a function of mission flight time. The methodology is exemplified with a 180 km flight in Danish airspace with a Penguin C aircraft.
TL;DR: COACH (Crective Advice Communicated by Humans), a new learning framework that allows non-expert humans to advise an agent while it interacts with the environment in continuous action problems, outperforms the other methods in terms of learning speed and final performance.
Abstract: The main goal of this article is to present COACH (COrrective Advice Communicated by Humans), a new learning framework that allows non-expert humans to advise an agent while it interacts with the environment in continuous action problems. The human feedback is given in the action domain as binary corrective signals (increase/decrease the current action magnitude), and COACH is able to adjust the amount of correction that a given action receives adaptively, taking state-dependent past feedback into consideration. COACH also manages the credit assignment problem that normally arises when actions in continuous time receive delayed corrections. The proposed framework is characterized and validated extensively using four well-known learning problems. The experimental analysis includes comparisons with other interactive learning frameworks, with classical reinforcement learning approaches, and with human teleoperators trying to solve the same learning problems by themselves. In all the reported experiments COACH outperforms the other methods in terms of learning speed and final performance. It is of interest to add that COACH has been applied successfully for addressing a complex real-world learning problem: the dribbling of the ball by humanoid soccer players.
TL;DR: A new wave variable compensation (WVC) structure is proposed to improve the tracking performances with less conservative condition and comprehensive analysis to keep stable and improved tracking performance is also provided.
Abstract: The master-slave teleoperated robotic systems have advanced the surgeries in the past decades. Time delay is usually caused due to the data transmission between communication channel connecting the master and slave in bilateral teleoperation, which is crucial because even small time delay could destabilize the whole teleoperation system. Motivated to solve the instability caused by time delay in bilateral teleoperation, wave variable transformation (WVT) structure has been proposed to passivate the delayed communication channel. However, conventional WVT structure provides poor velocity, position and force tracking performances which are not sufficient for surgical applications. In this paper, a new wave variable compensation (WVC) structure is proposed to improve the tracking performances with less conservative condition and comprehensive analysis to keep stable and improved tracking performance is also provided. In order to better facilitate certain surgical procedures with special requirements, e.g. robotic-assisted neurosurgery, velocity/position and force scalings are designed in the proposed structure with guaranteed system passivity, and transparency of the scaled WVC structure is also analyzed. Simulation and experimental studies were carried out to verify the performance of the proposed structure with time delay. System performance comparisons with several existing wave based bilateral teleoperation structures are also provided through simulation studies to show the improvements brought by the proposed teleoperation structure.
TL;DR: The approach enables to fully assist the UAV during its take-off and landing on the target, as it is able to detect anomalous situations, such as the loss of the target from the image field of view, and the precise evaluation of the drone attitude when only a part of thetarget remains visible in the image plane.
Abstract: In this paper we present an on-board Computer Vision System for the pose estimation of an Unmanned Aerial Vehicle (UAV) with respect to a human-made landing target. The proposed methodology is based on a coarse-to-fine approach to search the target marks starting from the recognition of the characteristics visible from long distances, up to the inner details when short distances require high precisions for the final landing phase. A sequence of steps, based on a Point-to-Line Distance method, analyzes the contour information and allows the recognition of the target also in cluttered scenarios. The proposed approach enables to fully assist the UAV during its take-off and landing on the target, as it is able to detect anomalous situations, such as the loss of the target from the image field of view, and the precise evaluation of the drone attitude when only a part of the target remains visible in the image plane. Several indoor and outdoor experiments have been carried out to demonstrate the effectiveness, robustness and accuracy of developed algorithm. The outcomes have proven that our methodology outperforms the current state of art, providing high accuracies in estimating the position and the orientation of landing target with respect to the UAV.
TL;DR: This work proposes a graph-based visual place recognition method that is able to obtain significantly better performance than that of FAB-MAP, a commonly used method for place recognition based on handcrafted features, especially on some challenging datasets.
Abstract: Visual place recognition is a critical and challenging problem in both robotics and computer vision communities. In this paper, we focus on place recognition for visual Simultaneous Localization and Mapping (vSLAM) systems. These systems have been limited to handcrafted feature based paradigms for a long time, which normally use local visual information of images and are not sufficiently robust against variations applied to images. In this work, we address place recognition with the features automatically learned from data. First, we propose a graph-based visual place recognition method. The graph is constructed by combining the visual features extracted from convolutional neural networks (CNNs) and the temporal information of the images in a sequence. Second, we propose to employ diffusion process to enhance the data association in the graph to achieve more accurate recognition results. Finally, to evaluate the proposed method, we experiment on four commonly used datasets. Experimental results indicate that the proposed method is able to obtain significantly better performance (e.g. 95.37% recall at 100% of precision) than that of FAB-MAP (47.16% recall at 100% of precision), a commonly used method for place recognition based on handcrafted features, especially on some challenging datasets.
TL;DR: Experimental results show the ability of the new approach to avoid areas that are considered risky for legged locomotion, and the A∗ algorithm to guide the RRT-Connect method, which yields detailed motion plans for the multi-d.o.f. legged robot.
Abstract: This paper considers motion planning for a six-legged walking robot in rough terrain, considering both the geometry of the terrain and its semantic labeling. The semantic labels allow the robot to distinguish between different types of surfaces it can walk on, and identify areas that cannot be negotiated due to their physical nature. The proposed environment map provides to the planner information about the shape of the terrain, and the terrain class labels. Such labels as “wall” and “plant” denote areas that have to be avoided, whereas other labels, “grass”, “sand”, “concrete”, etc. represent negotiable areas of different properties. We test popular classification algorithms: Support Vector Machine and Random Trees in the task of producing proper terrain labeling from RGB-D data acquired by the robot. The motion planner uses the A∗ algorithm to guide the RRT-Connect method, which yields detailed motion plans for the multi-d.o.f. legged robot. As the A∗ planner takes into account the terrain semantic labels, the robot avoids areas which are potentially risky and chooses paths crossing mostly the preferred terrain types. We report experimental results that show the ability of the new approach to avoid areas that are considered risky for legged locomotion.
TL;DR: The Lyapunov stability theorem is utilized to demonstrate that all the position errors, orientation errors, velocity tracking errors, observer estimation errors, and NN weight estimation errors are semi-globally uniformly ultimately bounded (SGUUB) in the presence of bounded disturbances.
Abstract: In this paper, a novel nonlinear output feedback neural network (NN)-based consensus controller is developed for a group of quadrotor unmanned aerial vehicles (UAVs). One UAV in the group tracks a desired trajectory while the rest of the group uses consensus-based formation controllers without knowledge of the desired trajectory. Each UAV estimates its own and its neighbor’s velocities through a novel nonlinear NN-based observer by using position and orientation information. Neighboring UAV positions and orientation information is assumed to be available via wireless communication or obtained through local sensors. Since quadrotor UAVs have six degree of freedom with only four control inputs, the UAV’s pitch and roll angles are utilized as virtual control inputs to bring all UAVs to consensus points along x and y directions. The Lyapunov stability theorem is utilized to demonstrate that all the position errors, orientation errors, velocity tracking errors, observer estimation errors, and NN weight estimation errors are semi-globally uniformly ultimately bounded (SGUUB) in the presence of bounded disturbances. The effectiveness of our consensus-based output feedback formation control of quadrotor UAVs is demonstrated in simulation validating our theoretical claims.