Humanoid robot

About: Humanoid robot is a research topic. Over the lifetime, 14387 publications have been published within this topic receiving 243674 citations. The topic is also known as: 🤖.

Real-time robot learning with locally weighted statistical learning

Abstract: Locally weighted learning (LWL) is a class of statistical learning techniques that provides useful representations and training algorithms for learning about complex phenomena during autonomous adaptive control of robotic systems. This paper introduces several LWL algorithms that have been tested successfully in real-time learning of complex robot tasks. We discuss two major classes of LWL, memory-based LWL and purely incremental LWL that does not need to remember any data explicitly. In contrast to the traditional beliefs that LWL methods cannot work well in high-dimensional spaces, we provide new algorithms that have been tested in up to 50 dimensional learning problems. The applicability of our LWL algorithms is demonstrated in various robot learning examples, including the learning of devil-sticking, pole-balancing of a humanoid robot arm, and inverse-dynamics learning for a seven degree of-freedom robot.

Omni-directional closed-loop walk for NAO

Abstract: —This paper describes the new walk algorithmimplemented on the NAO robot.NAO is a small fully actuated biped robot provided by theFrench company Aldebaran Robotics. Since the beginning of thecompany in July 2005, a major goal has been the developmentof robust walk for the robot.After 5 years of mecatronic design and improvements inrobustness of the robot (7 prototypes) and multiple prototypesof humanoid dynamic walk algorithm, an omni-directional walkrobust against small obstacles is now available for all the NAOunits (more than 700) in the world. I. I NTRODUCTION Humanoid robot research is an active eld. Over thepast decade, several anthropomorphic robots have been con-structed. Some of them became well known even for nonspecialists : Honda Robot Asimo [1], Sony's biped QRIO[2] and Kawada's humanoid HRP-2 [3]. For each of theserobots, there is much scientic literature describing theirwalk engines [4][5][6], even on unknown and rough ground[7]. All these scientic contributions give a good descriptionof the algorithm used to create different walk gaits. Butunfortunately it's hard to nd explicit information about howthe stabilization of the walk is implemented and how tocontrol a real robot with problems like calibration, sensorsdelay, sensors errors etc.Recently a new biped robot has arrived on the worldstage: NAO (cf. Fig.1). This is a small (56cm) and light(4.8Kg) biped robot with 25 degrees of freedom (5 in eachleg, 1 in the pelvis, 2 in the head, 5 in each arm and 2actuated hands). It is equipped with two cameras, an inertialmeasuring unit, sonar sensors in its chest, and force-sensitiveresistors under its feet [8]. NAO is different to all other bipedplatforms because it's an affordable robot (12K e ) which isfully programmable at high level (ex: cognition) or low level(ex: motion primitives) with Aldebaran Robotics SoftwareDevelopment Kit. Since 2008, NAO is the robot used in theRoboCup Standard Platform League [9]. In this competitionthe walk engine is critical for the game result : some teamshave developed their own walk process [10]. The best walkwas reached in the 2009 RoboCup edition with the universityof Bremen [11]. They created a very stable walk with a speedof 12 cm/s. The algorithm exposed in this paper is quitedifferent and based on more general concept.The main contribution of this paper is to describe, withoutomitting any assumptions or implementation details, the newFig. 1: The NAO robotrobust and omni-directional walk of NAO. The rst sectionpresents the walk engine based on an inverted Pendulummodel from the foot planner to trajectory generation. Theresult of this engine is a stable walk on at ground. Then,the stabilization process of this walk engine is described withconsideration of sensors errors. This results in a walk of10cm/s on NAO. In the last section, the robustness of thewalk is shown through walks across different ground surfaces(carpet, wood...) and walk over small books or cables.II. T

Humanoid motion planning for dynamic tasks

Abstract: This paper addresses an integrated humanoid motion planning scheme including both advanced algorithmic motion planning technique and dynamic pattern generator so that the humanoid robot achieve tasks including dynamic motions A two-stage approach is proposed for this goal First, geometric and kinematic motion planner first computes collision-free paths for the humanoid robot Then the dynamic pattern generator provides dynamically feasible humanoid motion including both locomotion and task execution such as object transportation or manipulation If the generated dynamic motion causes collision due to dynamic movements, the planner go back to the planning stage to remove the collision by path reshaping This iterative planning scheme enables robust planning against variation of task dynamics Simulation results are provided to validate the proposed planning method

3D walking based on online optimization

Abstract: We present an optimization based real-time walking controller for a full size humanoid robot. The controller consists of two levels of optimization, a high level trajectory optimizer that reasons about center of mass and swing foot trajectories, and a low level controller that tracks those trajectories by solving a floating base full body inverse dynamics problem using Quadratic Programming. Our controller is capable of walking on rough terrain, and also achieves longer foot steps, faster walking speed, heel-strike and toe push-off. Results are demonstrated with Boston Dynamics' Atlas robot in simulation.

A friendly gesture: Investigating the effect of multimodal robot behavior in human-robot interaction

Abstract: Gesture is an important feature of social interaction, frequently used by human speakers to illustrate what speech alone cannot provide, e.g. to convey referential, spatial or iconic information. Accordingly, humanoid robots that are intended to engage in natural human-robot interaction should produce speech-accompanying gestures for comprehensible and believable behavior. But how does a robot's non-verbal behavior influence human evaluation of communication quality and the robot itself? To address this research question we conducted two experimental studies. Using the Honda humanoid robot we investigated how humans perceive various gestural patterns performed by the robot as they interact in a situational context. Our findings suggest that the robot is evaluated more positively when non-verbal behaviors such as hand and arm gestures are displayed along with speech. These findings were found to be enhanced when the participants were explicitly requested to direct their attention towards the robot during the interaction.