We present a new method to translate videos to commands for robotic manipulation using Deep Recurrent Neural Networks (RNN). Our framework first extracts deep features from the input video frames with a deep Convolutional Neural Networks (CNN). Two RNN layers with an encoderdecoder architecture are then used to encode the visual features and sequentially generate the output words as the command...

Biped walking by using all joint movements and DOFs in both directions (sagittal plane and coronal plane) is one of the most complicated research topics in robotics. In this paper, angular trajectories of a stable biped walking for a humanoid robot are generated by a Truncated Fourier Series (TFS) approach. The movements of legs and arms in sagittal plane are implemented by an optimized gait ge...

An autonomous mobile humanoid robot able to assist in a workshop environment and interact with human beings, must have the basic capability of learning manipulation tasks by demonstration. In this paper two aspects to reach this goal are considered. First we describe how the mechanics must be constructed to manipulate objects like humans do. Then we propose an approach to transfer human arm mov...

In this paper, the inclinometer is used to control the balance of a humanoid robot on a miniature surfboard. GHR (Griffith University Humanoid Robot) is the humanoid robot developed in the Robotics and Games laboratory, at Griffith University. In order to measure the angle of the inclining surfboard, the inclinometer is attached to the back of the right foot of the GHR. To remove the high frequ...

have achieved a world-first success in their joint development of a tele-operated humanoid robot that wears protective clothing and drives an industrial vehicle (backhoe) outdoors in lieu of a human operator as shown in Fig. 1. It shows us that a humanoid robot can expand its capability by using machines designed for the human even in the rain. The use of a tele-operated humanoid robot has a po...

This paper deals with on-line motion imitation of a human being by a humanoid robot using inverse kinematics (IK). First, the human observed trajectories are scaled in order to match the robot geometric and kinematic description. Second, a task prioritization process is defined using both equality and minimized constraints in the robot IK model, with four tasks: balance management, end-effector...

In this paper we present a humanoid robot system that is able to mimic the motion of a person. The system performs mimicking, using real-time stereo vision. Our discussion will focus on the detection and motion inferencing of the action of a person, and the simultaneous reproduction of the same motion on a humanoid robot. Our attempt, is to remove from the conventional view, that the use of a f...

In this paper, we introduced a concept of realworld oriented humanoid robot that can support humans’ daily life behavior. Such realworld oriented robots have to behold humans, understand behaviors of humans and support daily life tasks. Especially, realworld behavior such as handling of table wares, delivering daily commodities by hand and so on, are required. We developed a humanoid robot HRP-...

The goal of this work is to build a cognitive system for the humanoid robot, Cog, that exploits human caregivers as catalysts to perceive and learn about actions, objects, scenes, people, and the robot itself. This thesis addresses a broad spectrum of machine learning problems across several categorization levels. Actions by embodied agents are used to automatically generate training data for t...

Theoretically the birth of the humanoid robot arises from the study of human characteristics of movement. However, following the human model does not mean copying it. It is still a big challenge to develop an advanced humanoid robot. Most researchers have shown an interest in developing an ideal theory to keep the balance in various grounds while walking since it is so important for the surviva...