Force and torque induced by traveling motion of a mobile robot effect dynamically to the objects being carried on it. If the induced force and torque should be bigger than the static friction force and torque exerting between the carrying objects and the mobile robot, the carrying objects start to slip on it. Since this slipping motion causes increasing the acceleration of the mobile robot, then the slipping of one object leads to dangerous collapse of all carrying objects. Furthermore it interferes with guidance control. On the other hand, mobile robots are desired to transfer the carrying objects as fast as it can. On this view point of this contradicted requirement, this paper purposes a controller to guide a mobile robot along a given course as fast as possible with acceleration restriction not to slip the carrying objects during traveling. The performances have been examined by simulations.
The human communication exists in various situations of our daily life. For human - robot communication or robot - robot communication, it is useful to design a communication model based on communication principal that is an entrainment phenomenon of nonlinear dynamics. In this paper, we focus on the robot motion synchronization for robot - robot communication. The robots are controlled to be entrained to an orbit attractor that corresponds to a robot motion. By exchanging the state variables of each robot, the robots are controlled to entrain one attractor that is possible for both robots and synchronize each other. The results of this paper represent the communication principal by an entrainment phenomenon of nonlinear dynamics.
The ultimate goal of this research is to develop a robot that is able to perform physical tasks such as nursing care skillfully and safely in complex environment of our home and/or hospitals. In order to realize such a robot, the design criteria and approach should be different from that of traditional industrial robots. Specifically, considering the physical interactions between the robot and human subjects, the safety, ease and affinity are more important in addition to the robot's high power, speed and accuracy. Towards this objective, “RI-MAN” is developed in RIKEN BMC which demonstrated the high ability to carry up and hold a doll softly and safely. This paper first discusses the problems on the design of a human-interactive robot, and proposes original key technologies and system integration to solve the problems including the soft and whole body interaction. The details and basic performance of RI-MAN are then introduced together with some experimental results. The further research subjects are also pointed.
Realization of an energy-efficient and high-speed dynamic walking has come to be one of the main subjects in the research area of robotic biped locomotion, and passive dynamic walking has been widely attracted as a clue to solve this problem. It has been empirically known that the effect of convex curve shape of feet, which characterizes passive-dynamic walkers, is important to increase walking speed. This research consists of two parts and aims to investigate the rolling effect of semicircular feet on dynamic bipedal walking. In this paper (I), driving mechanisms of compass-like biped robots and the rolling effect of semicircular feet are mainly investigated. We first analyze the mechanisms of a planar fully-actuated compass-like biped model to clarify the importance of ankle-joint torque introducing generalized virtual gravity concept. In the second, a planar underactuated biped model with semicircular feet is introduced and we show that virtual passive dynamic walking by hip-joint torque only can be realized based on the rolling effect. We then compare it with the flat-feet model through linear approximation, and show that the rolling effect is equivalent to its virtual ankle-joint torque. Throughout this paper, we give a novel approach to dynamic bipedal walking as ZMP-free robots.
This paper investigates what effect semicircular feet had on dynamic bipedal walking. It has been clarified in paper (I) that underactuated virtual passive dynamic walking can be realized by using the rolling effect, which acts as the ankle-joint torque virtually. This paper (II) then studies the rolling effect on the gait performances in detail through theoretical investigations and numerical analyses. We first perform parameter studies, and confirm that a suitable choice of the foot radius increases walking speed effectively. It is then theoretically clarified that increasing walking speed is achieved not by the rolling effect during the single support phase but by reducing the energy-dissipation due to heel-strike collisions. It is also confirmed that mass of the leg is necessary to drive the robot's center of mass in the walking direction. Furthermore, extension to redundant models with knee-joints is discussed.
This paper proposes a method for a humanoid robot to generate 3D model of the environment using a stereo vision, find a movable space using it and plan feasible locomotion online. The model is generated by an accumulation of 3D grid maps which are made from the range data of the field of view obtained by a correlation based stereo vision. The locomotion is planned by an online whole body pattern generator which can modify robot's waist height, an upper body posture and so on according to the size of the movable space.
In the case of control for nonlinear mechanical systems, which have complex actuator dynamics, the total system becomes high-order and nonlinear. Generally speaking, it is not easy to tune feedback gains for state feedback. In this paper, we propose a feedback control scheme for motion control of nonlinear high-order systems. We prove that the proposed scheme can improve the trajectory tracking performance of a robot manipulator system with actuator dynamics. Moreover, some simulation results demonstrate the effectiveness of the proposed control scheme.
This paper addresses the inverse kinematic problem for 7 DOF anthropomorphic manipulators with joint limits. Specifically, the paper analyzes how the joint limits restrict the redundancy to develop practical inverse kinematic resolutions under the joint limits. First, the manipulator model considered in this paper is provided. Second, how to solve the inverse kinematics of the redundant manipulator is described. Third, how joint limits affect the inverse kinematic solutions is investigated. Forth, several suggestions for resolving redundancy under the joint limits are presented. Lastly, the effectiveness of the redundancy resolutions is illustrated by simulation results.
The dynamics after that of the Newton-Euler method uses the symmetric property between a momentum vector and a velocity vector, or between a force vector and a velocity vector. The usage of the symmetric property makes the expression of the kinematic and dynamic analysis independent from the choice of a coordinates system. In this paper, it is clarified that this concept can be applied to the analysis of multi-rigid-body systems and mechanisms, using the concept of tangent and cotangent vectors based on the screw coordinates of joint axes. And the procedures of analyses of the closed loop mechanisms and parallel mechanisms are derived. An outline of the algorithm of the simulator for closed loop mechanism and parallel mechanisms is also described and the simulation results of parallel mechanisms are illustrated.
In this paper, we revised and applied the parallel solution scheme of inverse dynamics to flexible link systems where elastic deformation and vibration normally occur in constituting link members. The calculation process of the scheme is based upon the Finite Element Method (FEM), which evaluates the analyzed model in absolute Cartesian coordinates with the equation of motion expressed in dimension of force. The calculated nodal forces are converted into joint torques using a matrix form equation divided into terms of force, transformation between coordinates, and length. Therefore, information from the entire system can be handled in parallel, which makes the calculation seamless in application to any type of link system regardless of its boundary conditions or stiffness values. The scheme is revised and the calculation time is shortened by applying Bernoulli-Euler beam elements, and the scheme is then combined with a kinematics solution scheme that calculates target trajectories for flexible models. The calculation flow of inverse dynamics is shown for a 5-link system, and some feed-forward control experiments are carried out on a 2-link system with different stiffness members. The accuracies of trajectories and torque curves are verified by applying the system to a sensorless, model-based vibration control.
This paper describes an obstacle map generation by bayesian theorem. A sensor for the map generation can observe the obstacle which sensor can look, and can't observe the obstacle which is occluded from other obstacles. Previous method can't generate the actual map with the observation which has substantial margin of error because generate the map by distant observation model without considering occlusion. In our research, we use the observation model considering occlusion. In order to considering occlusion correctly, we estimate a visibility: Which obstacle is observed? can observe. Additionally in order to estimate the visibility, we have to consider spatial dependencies that the obstacle and the free area has size of space. In order to consider spatial dependencies, we estimate the visibility by a assumption that each line of view is markov chain. Experimental result shows that we obtained more precise generation of the obstacle map by our method than by previous method.
In this paper, we discuss the manipulation of an object under hybrid active/passive closure. The passive part of the system corresponds to the uncontrollable part. Therefore, to assure the stability of the controller, how to deal with the uncontrollable part is a key issue. We show the orthogonality between the active and passive parts in the previous paper. Based on the orthogonality, we decompose the dynamics of grasping system into the active part and the passive part. Then, using the decomposition, we derive a controller only including the dynamics of the active part. The controller assures to converge to the desired trajectories asymptotically. We also discuss how to determine the desired internal forces in order to satisfy frictional constraints during the manipulation. In order to verify the validity of our approach, some simulation results are shown.