Human friendly robot that works not replacing human being but working cooperatively with them, are required in the 21st century. The authors developed a human friendly robot to assist human being in physical tasks, and named such kind robot as Human Assist Robot (H.A.R.) . In this report, the authors describe the development of a prototype of H.A.R. for nursing use and physical assistance applications. Since some of the hardest tasks for nurses are to carry the patient to up/down from nursing bed or wheelchairs, the prototype robot is supposed to assist physically in such kinds of tasks. Experiments were carried out to certify the ability of the prototype H.A.R. in such tasks.
This paper proposes a multi-layered reinforcement learning system that integrates lower learning modules and generates one of higher purposive behaviors based on which an autonomous robot learns from lower level behaviors to higher level ones through its life time. We decompose a large state space at the bottom level into several subspaces and merge those subspaces at the higher level. This allows the system to reuse the policies already learned and to learn the policy against the new features. As a result, curse of dimension is avoided. To show its validity, we apply the proposed method to a simple soccer situation in the context of RoboCup, and show the experimental results.
This paper addresses the development and the experimental results of an aerial ski robot, which can perform a somersault coupled with twist motion in the air. The robot is composed of three rigid links and two rotational joints driven by DC motors and contains control devices so that complicated motion can be performed in the air. To plan the motion of the robot, Fourier Basis Algorithm (FBA) is applied, which approximates the optimal joint trajectories by the finite terms of the Fourier basis. The joint trajectories to perform a twisting somersault from FBA were inputted to the robot, and the flight experiment was carried out. As the result, the robot achieved one backward somersault with half twist. From this, the validity of the analytical model and the motion planning method was shown.
Precise measurements of geometry should accompany robotic equipments in operating rooms if their advantages are further pursued. For deforming organs including a liver, intraoperative geometric measurements play an essential role in computer surgery in addition to pre-operative geometric information from CT, MRI and so on. Previously developed laser-pointing endoscope acquires and visualizes the shape of the area of interest in a flash of time. Using this intraoperative geometry, in this paper we develope a surgical navigator for laparoscopic procedure. This system has the function of intraoperative monitoring and safety management. The authors believe that the proposed function enhances robotically assisted operations not only in ordinary operational environments but also in tele-operative environments. Results of in-vivo experiments on a liver of pig verify the effectiveness of the proposed system. The intraoperative 3D model of liver and 3D geometric calculation in virtual abdominal space were useful to avoid unexpected collisions with the forceps of surgical robot.
When a human user operates multiple mobile robots, it is difficult for him/her to obtain much information simultaneously from robots. In this paper, we propose a method of displaying visual, force, and sound information effectively in order that a human user could operate the robots with less load. The displaying information consists of operation guidance for the arrangement of the robots around an object, transportation guidance for transporting the object, and work guidance for precise movement. With these guidances, the user deals with excess information and controls the multiple robots effectively. We built a real system and verified the effectiveness of the proposed method and the system in experiments.
This paper presents a virtual teaching method for multi-fingered robots based on hand manipulability, in which a position and orientation of the robot hand is determined so as to maximize a manipulability of the robot hand on the condition that the robot grasps the object at taught contact points of the object. In order to get the finger joint angles in movable range, a penalty function is added to the manipulability measure. This approach does not require the accurate human hand model. The simulation results of a pick-and-place task are shown.
This paper presents an advanced leg module developed for HRP-2. HRP-2 is a new humanoid robotics platform, which we have been developing in phase two of HRP. HRP is a humanoid robotics project, which has been lunched by Ministry of Economy, Trade and Industry (METI) of Japan from 1998FY to 2002FY for five years. The ability of the biped locomotion of HRP-2 is improved so that HRP-2 can cope with rough terrain in the open air and can prevent the possible damages to a humanoid robot's own self in the event of tipping over. In this paper, the mechanisms and specifications of leg module, electrical system, simulation results utilized for deciding specifications, and experimental results are also introduced.
Autonomous map construction is one of the most fundamental and significant issues for intelligent mobile robots. While a variety of construction methods have been proposed, most of them are dependent on the quantitative sensor information and accurate physical models for estimating the positions and directions of the external objects and robot itself. This paper proposes a new map construction method based on rough information of “how often two objects are observed simultaneously” . This method is founded on the combination of a simple and general heuristics-“closely located objects are likely to be seen simultaneously more often than distant objects” and a well-known technique in the multivariate data analysis-Multi-Dimensional Scaling (MDS) . A significant feature of this method is that it requires little quantitative models nor precise sensor information, unlike conventional map construction methods. Simulation and experiment results suggest that this method is sufficiently practical for grasping a topological configuration of identifiable landmarks quickly.