Conference Paper

Kinematically Optimal Catching a Flying Ball with a Hand-Arm-System

DLR Inst. of Robot. & Mechatron., Wessling, Germany
DOI: 10.1109/IROS.2010.5651175 Conference: Proc. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
Source: DLR


A robotic ball-catching system built from a multi- purpose 7-DOF lightweight arm (DLR-LWR-III) and a 12 DOF four-fingered hand (DLR-Hand-II) is presented. Other than in previous work a mechatronically complex dexterous hand is used for grasping the ball and the decision of where, when and how to catch the ball, while obeying joint, speed and work cell limits, is formulated as an unified nonlinear optimization problem with nonlinear constraints. Three different objective functions are implemented, leading to significantly different robot movements. The high computational demands of an online realtime optimization are met by parallel computation on distributed computing resources (a cluster with 32 CPU cores). The system achieves a catch rate of > 80% and is regularly shown as a live demo at our institute.

Download full-text


Available from: Thomas Wimböck
  • Source
    • "[12] the task of thrown ball catching is pursued by a multi-purpose 7-DOF lightweight arm and a sophisticated 12-DOF four-fingered hand. In this work the ball is again detected by a stereo camera system. "
    [Show abstract] [Hide abstract]
    ABSTRACT: This study proposes a method for performing velocity estimation by motion blur in a single image frame and intercepting a moving object with a robotic arm. It is shown that velocity estimation in a single image frame improves the system's performance up to five times. The majority of previous studies in this area require at least two image frames to measure the target's velocity. The speed estimation is practiced by a Kinect camera and the implemented control system converts the position and velocity information of the moving object to the suitable joint angles and torques to intercept within a certain time interval. To test the system, two different sets of experiments are designed in which the target swings on a string or is thrown within the working volume of the robotic arm. According to the recorded velocity data and the success of the experiments, the limitations of the method and setups will be discussed.
    Full-text · Conference Paper · Aug 2015
  • Source
    • "Nishiwaki et al. realized the ball-catching behavior by a humanoid robot (Nishiwaki et al., 1997). These works have been subsequently followed by recent studies (Riley and Atkeson, 2002) (Bäuml et al., 2010) (Birbach et al., 2011). However, this motion has not yet been realized in commercial robot systems. "
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a controller design of a robotic manipulator for soft catching of a falling object. If a robotic system is able to catch a falling object softly, there will be many applications expected in human activities such as industry, welfare, nursing, housework and office work, because this ability allows a human operator or another robot system to move an object to the catching robot without any transportation systems such as an conveyor or a mobile structure. First, this paper considers a nonlinear decoupling control of a robotic manipulator. Next, a controller design is presented for catching a falling object with a small impact force. This controller consists of two parts: a position tracking controller that tracks a desired trajectory before contact between the object and the robot end-effector, and a force controller that is triggered after the contact. We employ a position-based impedance controller so that the entire control system can be constructed as a position-based controller. In order to achieve the soft catching, precise motion control is required to achieve the same velocity of the robot end-effector with a falling object when they are in contact. Hence, we employ an adaptive controller that consists of a feedback controller to compensate for disturbance such as friction and a feed forward controller to improve the tracking performance to the desired trajectory by adjusting controller parameters in real time. Experimental results with a falling raw egg demonstrate the effectiveness of the proposed approach.
    Preview · Article · Jan 2015
  • Source
    • "We now consider the problem of intercepting a ball as presented in [2]. Assume the situation of Fig. 6, where two robots R 1 , R 2 are chasing a ball B. The movement of the robot R 1 with respect to the robot R 2 is represented by "
    [Show abstract] [Hide abstract]
    ABSTRACT: We present a logic approach to reason with moving objects under fuzzy qualitative representation. This way, we can deal both with qualitative and quantitative information, and consequently, to obtain more accurate results. The proposed logic system is introduced as an extension of Propositional Dynamic Logic: this choice, on the one hand, simplifies the theoretical study concerning soundness, completeness and decidability; on the other hand, provides the possibility of constructing complex relations from simpler ones and the use of a language very close to programming languages.
    Full-text · Article · Jul 2013 · Fuzzy Sets and Systems
Show more