Syungkwon Ra

Human-like catching motion of humanoid using Evolutionary Algorithm(EA)-based imitation learning

A framework to generate a human-like arm motion of a humanoid robot using an Evolutionary Algorithm(EA)-based imitation learning is proposed. The framework consists of two processes, imitation learning of human arm motions and real-time generating of a human-like arm motion using the motion database evolved in the learning process. The imitation learning builds the database for the humanoid robot that is initially converted from human motion capture data and then evolved using a genetic operator based on a Principal Component Analysis (PCA) in an evolutionary algorithm. This evolution process also considers the minimizing of joint torques in the robot. The database is then used to generate humanoid robots arm motions in real-time, which look like humans and require minimal torques. The framework is examined for humanoid robot to reach its arms for catching a ball. Additionally, the inverse kinematics problem to determine the final posture of 6-DOF robot arm with a waist for the task of catching a ball, is proposed.

Self-stabilizing bipedal locomotion employing neural oscillators

For attaining a stable and robust dynamic bipedal locomotion, we address an efficient and powerful alternative based on biologically inspired control framework employing neural oscillators. Neural oscillators can be used to generate sustained rhythmic signals, and show superior features for stabilizing bipedal locomotion particularly when coupled with virtual impedance components. By building a network of neural oscillators, we can enable humanoid robots to walk stably and exhibit robustness against unexpected disturbances. Specifically, in order to maintain stability, the neural oscillator plays an important role by controlling the trajectory of the COM in phase with the ZMP input. The effectiveness of the proposed control scheme is verified through simulations.

Imitation Learning of Robot Movement Using Evolutionary Algorithm

Abstract This paper presents a new framework to generate human-like movement of a humanoid robot in real time using the movement primitive database of a human. The framework consists of two processes: (1) the offline motion imitation learning based on Evolutionary Algorithm and (2) the movement generation of a humanoid robot using the database updated by the motion imitation learning. For the offline process, the initial database contains the kinetic characteristics of a human, since it is full of humans captured motions. The database then develops through the proposed framework of motion learning based on Evolutionary Algorithm, having characteristics of a humanoid in aspect of minimal torque. The humanoid generates a human-like movement corresponding to a given task in real-time by linearly interpolating the primitive movements in the developed database. The proposed framework is a systematic methodology for a humanoid robot to learn human motions, considering the dynamics of the robot. The experiment of catching a ball thrown by a man is performed to show the feasibility of the proposed framework.

Walking pattern generation method with feedforward and feedback control for humanoid robots

This paper proposes a new walking pattern generation method for humanoid robots. This paper uses the linear inverted pendulum model (LIPM) which is composed of zero moment point (ZMP) and center of mass (CoM). Based on LIPM, the proposed method consists of feedforward control and feedback control for walking pattern generation of humanoid robots. The linear quadratic regulator(LQR) as a feedback controller tracks the desired ZMP according to footprints of humanoid robots and makes poles of LIPM stable. The feedback controller, pole-zero cancelation by series approximation (PZCSA) plays a role of reducing the inherent property of LIPM and approximating the transfer function of the overall system including LIPM and controllers to be unity. The usefulness of the proposed method is verified by simulations such as arbitrary time intervals of support phases, arbitrary desired ZMP position and sudden changed desired ZMP position. And the validity of the proposed method is confirmed by the experiment of a humanoid robot using a joystick.

PCA-based genetic operator for evolving movements of humanoid robot

This paper proposes a new genetic operator in order to evolve the humanoid movements, which is composed of principal component analysis (PCA) and descent-based local optimization with respect to robot dynamics. The aim of the evolution is to let humanoid robots generate human-like and energy-efficient motions in real-time. We first capture human motions and build a set of movement primitives. The set is then evolved to the optimal movement primitives for the specific robot, which contain its dynamic characteristics, by using an evolutionary algorithm with the proposed genetic operator. Finally, the humanoid robot can generate arbitrary motions in real-time through the mathematical interpolation of the movement primitives in the evolved set. The evolved set of movement primitives endows the humanoid robot with natural motions which require minimal torque. This technique gives a systematic methodology for a humanoid robot to learn natural motions from human considering dynamics of the robot. The feasibility of our genetic operator is investigated by simulation experiments in regard to catching a ball that a man throws of the humanoid robot.

Regenerating human-like arm motions of humanoid robots for a movable object

A method to regenerate a new human-like arm motion by modifying and scaling a human arm motion is presented. The humanoid robot may not obtain all the necessary human-like motions from motion capture data. When the robot communicates with a person, the robot has to keep attention to the person by aligning the direction of the motion. The robot needs to modify human arm motions and regenerate new human-like arm motions without loosing the original meanings. The developed method modifies and scales the wrist trajectory of a human arm motion and reproduces the human-like arm. The motion of drawing multiple circles with a various radius and direction is examined.

Self-stabilizing human-like motion control framework for humanoids using neural oscillators

We propose an efficient and powerful alternative for adaptation of human motions to humanoid robots keeping the bipedal stability. For achieving a stable and robust whole body motion of humanoid robots, we design a biologically inspired control framework based on neural oscillators. Entrainments of neural oscillators play a key role to adapt the nervous system to the natural frequency of the interacted environments, which show superior features when coupled with virtual components. The coupled system allows an unstable system to stably move according to environmental changes. Hence the feature of the coupled system can be exploited for sustaining the bipedal stability of humanoid robots. Also based on this, a marionette-type motion conversion method to adapt captured motions to a humanoid robot is developed owing that there are the differences in capabilities of dynamics and kinematics between a robot and a human. Then this paper discuss on how to stably show human motions with a humanoid robot. We verify that a real humanoid robot can successfully sustain the bipedal stability exhibiting captured whole body motions from various simulations and experiments.