The paper addresses the problem of the generation of collision-free trajectories for a robotic manipulator, operating in a scenario in which obstacles may be moving at non-negligible velocities. In particular, the paper aims to present a trajectory generation solution that is fully executable in real-time and that can reactively adapt to both dynamic changes of the environment and fast reconfiguration of the robotic task. The proposed motion planner extends the method based on a dynamical system to cope with the peculiar kinematics of surgical robots for laparoscopic operations, the mechanical constraint being enforced by the fixed point of insertion into the abdomen of the patient the most challenging aspect. The paper includes a validation of the trajectory generator in both simulated and experimental scenarios.
Beyond robot hardware and control, one major element for an efficient, constructive and safe mission of teleoperated robots in disaster scenarios such as Fukushima is the quality of the connection between operator and robot. In this contribution, we present the concept of using an exoskeleton and utilizing 3D simulation as a central interface component for the operator to intuitively collaborate with mobile teleoperated robots. keywords: 3D simulation, exoskeleton, force feedback, operator interface
This paper describes a walk-through programming technique, based on admittance control and tool dynamics compensation, to ease and simplify the process of trajectory learning in common industrial setups. In the walk-through programming, the human operator grabs the tool attached at the robot end-effector and "walks" the robot through the desired positions. During the teaching phase, the robot records the positions and then it will be able to interpolate them to reproduce the trajectory back. In the proposed control architecture, the admittance control allows to provide a compliant behavior during the interaction between the human operator and the robot end-effector, while the algorithm of compensation of the tool dynamics allows to directly use the real tool in the teaching phase. In this way, the setup used for the teaching can directly be the one used for performing the reproduction task. Experiments have been performed to validate the proposed control architecture and a pick and place example has been implemented to show a possible application in the industrial field.
We present a human-robot interaction framework that integrates multimodal collaborative execution and kinesthetic teaching to accomplish complex and new collaborative tasks in an industrial scenario. We consider the case of a hospital scenario where a human user interacts with a robotic arm in order to collect and to arrange tools.
Known as the ’art of combining strokes to form complex letters’, the Chinese and Korean Calligraphy can be learned by assimilating the skill of drawing these strokes. Once this dexterity is mastered, it can be used to compose a calligraphic letter afterwards. The very notion of calligraphy is endeavored to be implemented on robots in this research. Korean calligraphy is the nucleus of this research and the image of the calligraphic character is used as an input. Unlike humans or calligraphers, the robots are unacquainted with combination of various strokes used to draw a calligraphic letter. Hence it is quite cardinal and arduous to fragment the calligraphic letter into diverse strokes used to draw it. Therefore this has been the area of interest for many researchers which led to couple of unalike approaches like in [1], [2] and [3] but with some drawbacks. Authors in [1] used geometric properties of contour of the character to extract the stroke which fails in even very simple case like shown in Figure 13 in [1]. Hence a novel approach ensuing Gaussian Mixture Model (GMM) is proposed in this research to segregate the input image into assorted strokes. Once the strokes are extracted, they are combined to reproduce the same character using Gaussian Mixture Regression (GMR). The resulted learned strokes are then implemented on KUKA Light Weight Robot and further improved using RL.
segmental angular momenta analysis in human walking subjected to lateral pelvic perturbations. The results show that angular momenta of specific segments, right upper leg and right foot in the study, were mostly affected by the applied perturbations. This finding will be further utilized to develop a perturbation detection method for monitoring stability while lower limb exoskeleton-supported walking.
The main objective of this work is obtaining a framework able to provide bio-inspired torque assistance to disabled humans during walking and stair ascending/descending. The method is based on bio-inspired models copying natural dynamics of the leg muscles, and containing top-down primitives, short-loop reflexes, and a torso stabilization mechanism.
Robots are still outperforming humans regarding their efficiency and dexterity. On the other hand, however, muscles are more efficient and have a higher power to weight ratio than electric motors. As a consequence, among others, manipulators for human robot interaction in industry or service applications have a low payload to weight ratio and a low energy efficiency. Therefore, we developed a novel actuation concept which we name Series-Parallel Elastic Actuation (SPEA). The concept enables variable recruitment and locking of multiple springs in parallel. In this paper the latest MACCEPA-based SPEA is presented and first results are shown. Servomotors are generally used for industrial robots as they make the joint’s mechanical impedance very high, often considered as infinite, so they are ideal for precise tracking with a high bandwidth. As a consequence, however, industrial robots are unsafe for human-robot interaction and are positioned in cages or secured by safety light curtains and sensors. The next generation of robots will strongly collaborate with humans, implying new requirements such as safety and energy-efficiency. For example manipulators for assistance technologies in industrial settings, also known as co-workers, are currently a vastly emerging technology. Multiple manipulators, of which some are listed in Table I, are being developed in research projects or are already commercially available. Mekabot’s A2 compliant robot arm and Rethink Robotics’ Baxter are rather exeptional in Table I since these are the first commercialized manipulators which are driven by complaint actuators, also known as Series Elastic Actuators [1]. The limited torque to weight ratio of current actuators, stiff or compliant, however limits the payload to weight ratio of these manipulators. This can be seen from Table I. On the contrary, the average specific power density of a mammalian skeletal muscle (0.05 W/g) is an order or magnitude lower compared to electric motors (0.5 W/g) [2]. The maximum energy efficiency of an electric motor (>80%) is also higher compared to a muscle (<40%). Despite this higher power density and maximum efficiency, the electric motors are not yet able to better All authors are with the Robotics & Multibody Mechanics Research Group, Faculty of Mechanical Engineering, Vrije Universiteit Brussel, 1050 Elsene, Belgium. http://mech.vub.ac.be/robotics. Glenn Mathijssen is with the Robotics & Multibody Mechanics Research Group at the Vrije Universiteit Brussel and with Centro E. Piaggio of the University of Pisa. This work has been funded by the European Commission as a part of the ERC-grant SPEAR (no.337596). 1 Corresponding author: Glenn.Mathijssen@vub.ac.be Stiff actuators ABB’s dual arm concept FRIDA 2x0.5 kg/20 kg Universal Robots UR5 5 kg/18.4 kg Yaskawa’s SDA robots 5 kg/110 kg Kawada NextAge 2x1.5 kg/28 kg Kuka lightweight LWR4+ 7 kg/16 kg
