Rafael Cisneros

Humanoid robot HRP-2Kai — Improvement of HRP-2 towards disaster response tasks

This paper presents a humanoid robot HRP-2Kai, which is the improvement version of HRP-2 towards disaster response tasks. HRP-2 stands for a humanoid robotics platform-2, which was developed in phase two of the Japanese national project HRP (Humanoid Robotics Project, from 1998FY to 2002FY), while Kai means improvement in Japanese. In a year of the ninth year from releasing HRP-2, the Great East Japan Earthquake shook Japan on March 11, 2011. Since we reflected on that we were not able to deploy our robots for emergency response at that time, we started a study of the disaster response humanoid robots by improving HRP-2. Improvements are presented with its basic specification in this paper.

Task sequencer integrated into a teleoperation interface for biped humanoid robots

This study describes our teleoperation interface, through which an operator can dictate various tasks to a biped humanoid robot. In this interface, the robot model and the environmental data are presented in three-dimensional computer graphics, and the robot actions are specified by several operational markers. The interface also provides a task sequencer function, enabling the operator to perform a task merely by proceeding a task sequence, which structures the procedure of a particular task. The task sequencer provides a much more efficient operation than the direct use of operational markers. We verified the interface by participation in the DARPA Robotics Challenge (DRC) Finals. Our humanoid robot HRP-2kai operated through our teleoperation interface completed most of the DRC tasks.

Task-level teleoperated manipulation for the HRP-2Kai humanoid robot

This paper presents the strategy used by the team AIST-NEDO at the DARPA Robotics Challenge to deal with the designated manipulation tasks by means of a task-level teleoperation of the HRP-2Kai humanoid robot, considering a disaster-hit scenario that is inherently non-structured and a limited communication between the user and the robot. The strategy, based on the information provided by a laser rangefinder and a set of cameras installed at the head and at both hands, consisted in the alignment of 3D models representing the desired manipulation targets with a measured point cloud, in order to provide a reference frame to describe the manipulation motion required for each task. Each motion was carefully planned in advance by assuming minimum information of the object representing the manipulation target. In order to exemplify the before mentioned approach, two representative tasks of the DARPA Robotics Challenge are described, as well as the corresponding results obtained during the competition.

Impact acceleration of falling humanoid robot with an airbag

Maximum absolute acceleration at impact is widely used for evaluating damage at robot fall. In this paper, we measured the impact acceleration of a life-size humanoid robot when it fell down. The experiment showed that a robot suffered maximum acceleration of over 100G at the impact. To mitigate this we tested an airbag system. Our experiment showed the airbag could reduce the peak acceleration to between 20G and 30G, which is an acceptable level for the most of robot hardware.

Stable simulation of flexible cable-like objects by using serial kinematic chains with high number of passive degrees-of-freedom

This paper presents a technique to achieve a stable simulation of flexible cable-like objects approximated as serial kinematic chains with a high number of passive degrees of freedom. This is achieved by implementing a dry-friction model based on implicit Euler integration of Coulomblike discontinuous friction, which does not exhibit chattering nor unbounded drift. The applicability of this technique is demonstrated through its implementation to simulate a hose for the Japan Virtual Robotics Challenge.

Effective teleoperated manipulation for humanoid robots in partially unknown real environments: team AIST-NEDO’s approach for performing the Plug Task during the DRC Finals

Graphical Abstract Abstract This paper introduces the approach used by AIST-NEDO team during the DARPA Robotics Challenge (DRC) Finals to perform the Plug Task, which is taken as a case study to show how we dealt, in general, with the manipulation tasks proposed by this challenge, using similar strategies. For that purpose, this paper also describes the system used to teleoperate our humanoid robot, HRP-2Kai, in a disaster-hit scenario that is partially unknown, inherently non-structured and where communications are not reliable. This description includes an overview of our robotic platform, as well as the overall structure of the teleoperation interface and the algorithms behind. Also, the way the Plug Task was performed using these algorithms as operation units is explained in detail, by means of a high-level description. Finally, this paper presents the results for this task obtained during the actual competition, as well as a comparison with the performance achieved by other teams that performed the Plug Task in the DRC Finals.

Developing semi-autonomous humanoid robots that perform various composite tasks via a task sequencer and dynamics simulator

This paper proposes a system to enable a humanoid robot to perform various tasks semi-autonomously according to higher-layer procedures called task sequences. A task sequence is a sequentially structured procedure that describes how to perform a particular task based on sensing data, motion planners, and a robot controller. These task sequences are processed by the task sequencer to move the robot. The goal of our system is to support comprehensive tasks, such as disaster response, with a number of task sequences defined for various subtasks. To efficiently develop required task sequences, the task sequencer is designed such that task sequences can be executed interactively; further, the task sequencer is integrated with a dynamics simulator that can be used to test the sequences without actually moving the robot. We verified our system by using it to make our HRP-2Kai robot perform the tasks defined in the DARPA Robotics Challenge, a competition involving disaster response robots.

Enabling a teleoperated humanoid robot to pass through debris-filled terrain using manipulation

This paper describes the methodology followed to enable a semi-autonomous humanoid robot to pass through a narrow lane filled with debris, a task that was originally inspired by the DARPA Robotics Challenge Trials. During this competition, each team had 30 minutes to complete the task, a time limit that was not suitable if 8 tasks had to be completed within one hour, as in the Finals. It was then our objective to speed up the performance of this task, and try to complete it as fast as we could. The resulting strategy is described here, as well as an analysis of the outcome and a comparison of performance with the teams that participated in the Trials.