This paper presents an on-line pose measurement method of a 3D object. The proposed method utilizes an on-line evolutionary search technique of the genetic algorithm (GA) in pattern recognition and a fitness evaluation based on matching input dynamic images with stereo models whose poses are expressed by unit quaternion. To improve the dynamics of recognition, a motion-feedforward compensation method is proposed for the hand-eye system by predicting target object's motion in camera frame induced by the hand-eye motion through the kinematic calculation of robot's motion. The effectiveness of the proposed method is confirmed by simulation experiments.
We are conducting the research project named Robot Town Project. The aim of this research is to develop a distributed sensor system such as cameras, laser range finders, and IC tags, and its management system so that autonomous robots can work with humans in an ordinary environment for daily human life. This paper presents a sensor network system consisting of distributed cameras and laser range finders for multiple objects tracking. Sensory information from cameras is processed by the Level Set Method in real time and integrated with range data obtained by laser range finders in a probabilistic manner using novel SIR/MCMC combined particle filters. Though the conventional SIR particle filter is a popular technique for object tracking, it has been pointed out that there are some drawbacks in practical applications such as its low tracking performance for multiple targets due to the degeneracy problem. In this paper, the new combined particle filters consisting of a low-resolution MCMC particle filter and a high-resolution SIR particle filter is proposed. Simultaneous tracking experiments for multiple moving targets are successfully carried out and it is verified that the combined particle filters has higher performance than the conventional particle filters in terms of the number of particles, the processing speed, and the tracking performance for multiple targets.
We propose a new weight compensation mechanism with a non-circular pulley and a spring. We show the basic principle and numerical design method to derive the shape of the non-circular pulley. After demonstration of the weight compensation for an inverted/ordinary pendulum system, we extend the same mechanism to a parallel four-bar linkage system, analyzing the required torques with transposed Jacobian matrices. Finally, we develop a 3 D.O.F manipulator with relatively small output actuators and verify that the weight compensation mechanism greatly contributes to decrease static torques to keep the same posture within manipulator's work space.
In order for humanoids to be able to achieve task-oriented behaviors with high responsiveness against instantaneous environmental or self-body changes, the realization of software adaptability is one of the key issues. For example, humanoids which interact with human at a very close distance have to respond to many contact events with outer environments, human bodies, or a self body during the interaction behaviors at the same time. In order to deal with such complex circumstances, behavior integration system is adopted in this paper. By introducing such behavior integration system, adaptability to many local events can be added incrementally. Also the analytically designed and slightly limited behavior modules can be accumulated for generating next richer behaviors. As such behavior integration system, ‘Parallel Evaluating Monitors’ is proposed. ‘Parallel Evaluating Monitors’ always run in the background of the normal task-oriented behavior, and it monitors the sensor changes and modifies the original motion trajectories if it's evaluated as a necessary one from monitoring results. In this paper, the implementation of the ‘Parallel Evaluating Monitors’ is described, and the implemented behavior integration system is applied to the different types of humanoids' interaction behavior for confirming the feasibility of the proposed system.
This paper describes an epoch-making pneumatic driving unit which enables the generation of high-speed motion aiming a higher throw and jump realized by a robot. The proposed unit, named MB cylinder, is basically composed of a pneumatic cylinder, a permanent magnet, a portable tank, and small valves. Since the magnet plays a role to enhance the rapid release function of pneumatic energy instead of using a big and heavy valve, the pressure inside the cylinder can be kept in high condition and can generate high speed with light structure. The height control method of MB cylinder and its design method as well as the analysis of the performance of MB cylinder are also described in this paper. After the developed unit is installed on both the throwing device and the jumping robot, the validity of the proposed methods is experimentally verified in addition to discussion on its application to the rescue operation.
In this research, a Universal map, which can be converted to individual maps for heterogeneous mobile robots, is proposed. A Universal map can be generated using our developed measurement robot, and it is composed of a textured 3D environment model. Therefore, every robot can use a Universal map as a common map, and it is utilized for various localization technologies such as view-based and LRF-based methods. In LRF-based localization, accurate localization is achieved using a specific map, which is generated from a Universal map. In a view-based approach, localization and navigation are achieved using rendered images. The use of a Universal map enables generation of these maps automatically. The effectiveness of this approach is confirmed through experiments.
It is important for autonomous control of a biped robot to obtain the 3D information of the ground on which the robot walks. When a biped robot traverses a floor including slopes, sensors should be able to detect a slope and precisely measure the angle of the slope, along with the positions of its beginning and end. In this paper, we propose a realtime floor sensing method using stereo cameras mounted on a biped robot. We detect a slope in the environment and estimate the inclination angle and the boundary by fitting a set of plane parameter vectors to a two-plane model. For obtaining the parameter vector set, we first determine multiple regions of interest (ROI) in a reference image by using footstep positions up to several steps, scheduled by a current footstep plan. Then the parameter vector of the floor with respect to each ROI are efficiently and accurately estimated by a fast direct method with motion compensation. The fitting result is feedback for stable walking by updating the footstep plan . The validity of the proposed method is demonstrated through online experiments using stereo cameras mounted on the body of a biped robot, Honda ASIMO, traversing a real slope.
This paper describes a method of 3D localization of partially buried objects for automated excavation system in unstructured environments. The objects are limited to surfaces of revolution. Range image obtained by a correlation-based stereo vision is employed for 3D sensing. Candidate regions are extracted as convex surface regions from the range image. For each region, multiple hypotheses for the position and the direction of the rotational axis are generated for each object model. Each hypothesis is verified and improved by an iterative method, and then the most reliable hypothesis is adopted as a recognition result. From the object recognition result, an operator selects an object which a robot grasps. Based on a command from the operator, the robot picks the object and transfers it to a storage box. The experimental results show the effectiveness of the proposed method for various environments including objects which are partially buried, inclined, touching each other and mixed with other objects.