Pyramid- and tree-type structures manipulate hierarchical and multi-resolution representations of data and are used in many image processing applications. However, it has been thought that it is difficult to deal with a spherical image by use of a pyramidal data structure. Because, connectedness of such structure is far weaker than a normal square tessellation. In this paper, we propose a new data structure for spherical image processing on an omni-directional view that is a spherical hexagonal image pyramid. It is composed of hexagonal pixels on a regular polyhedron. Using this data structure, we can represent an omni-directional view in a complete spherical image. And we performed an experiment to extract great circles in a spherical image and to estimate coordinate systems using vanishing points as one of omni-directional image processing applications.
In this paper, we propose two prototype models of micro active catheter (MAC) that has two bending degrees of freedom. One consists of 3 active units with SMA wires in lumina as the servo actuator. The other is active guide wire catheter with ICPF (Ionic Conducting Polymer Film) on its front end as the servo actuator. The bending characteristics of the MAC have been measured by application of electricity in physiological saline solution. We also modeled these MAC for characteristic evaluation (Bending characteristic, Electric characteristic) . Experimental results show that the models of the active catheter are reasonable. By using simulators (whose conditions are similar to those of a blood vessel of human brain), we also carried out simulation experiments “in vitro” and “in vivo”. The experimental results indicate that the proposed micro catheters work properly, and they can effectively improve the operability of traditional procedures for intravascular neurosurgery.
In this paper, we propose two design methods for robust control of master-slave systems, where one method gives a controller to guarantee the robust stability and the robust performance for all passive environments and operators, and the other guarantees the robustness for all environments and operators specified by the gain condition even if the environment is not passive. In both proposed methods, the dynamics of operator and environment is formulated as some kind of structured uncertainties included in the system, and the control problem of master-slave systems is reduced to a problem in terms of μ synthesis. In addition, we confirm the validity of the two proposed methods by some simulation results and discuss the applicable classes of each design method.
In this paper, we propose a method of planning of pushing operation, which is to move an object on a floor to a desired position and orientation by one arm robot. We suppose that the distribution of magnitude of friction between the object and the floor is known. This method is as follows. During one pushing motion, the position of the contact point between the object and the pusher, and the direction of its velocity are kept to be constant with respect to the object, so that the path of it will be an arc or a straight line. Some paths to the desired position and orientation are generated by connecting these arc and straight paths. The optimal path is decided to minimize a performance index as a function of the number of pushing, the total traveling distance, and the total change of angle. We show an example of planning for a polygonal object under the case where a finite number of contact points, push directions, and intermediate orientations of the object are given. Numerical examples show the effectiveness of the proposed method.
A multifingered hand can reorient an object in a hand. This paper discusses one of primitive ways to reorient a prism object. Two fingers (corresponding to the human thumb and middle fingers) grasp the top and bottom faces of the prism object. Their fingertips form a rotational axis. The third finger (corresponding to the human index finger) exerts a force to the side faces to rotate the object about the axis. Since the workspace of the third finger is limited, the side faces reachable by it are also limited depending on the position of the axis formed by the first two fingers. This paper shows an algorithm for computing the possible region for the axis so that the third finger can rotate the object from start to goal angles. The algorithm also determines the side faces to be pushed by the third finger. It is applicable to any shaped prism objects.
This paper provides a new method for obtaining a collision free path for a manipulator. The proposal in this paper consists of two steps. First, the authors define Path-Restricted-Curved-Surface (PRCS) as a free form surface in the configuration space that includes both start and goal points and exclude collision regions. The PRCS is described by Bezier surface in the configuration space. Moving along the PRCS, the manipulator doesn't occur to collide, because the PRCS is a collision-free area. Second step is to generate the optimized path on the PRCS. The path on the PRCS is selected to the geodesic connects from start to goal points. The geodesic in the configuration space is the most suitable path in the point to minimize the total value of manipulator's joint angle changes. The effectiveness of this method is clarified by computer simulation with 3D graphics. This method can be said to apply the differential geometry to configuration space method.
In the exploitation of a magnetic robot hand, it is extremely important to develop a gap sensor of high and stable performance for the relative motion in three dimensional space between the held body and the magnetic robot hand. In this study, first, a prototype gap sensor system for the instrument of the gap between a magnetic robot hand and a spherical magnetic body held and transferred without contact point is developed. The prototype gap sensor system consists of a magnetic flux sensor, an exciting current sensor and a data base calculation system. The data base calculation system is constructed with the instrument data obtained by the beforehand experiments among the gap, the magnetic flux and the exciting current. The magnetic robot hand is composed with the prototype gap sensor system and an electromagnet. Second, the experiments of the non-contact hold and transfer control of the spherical magnetic body have been carried out, and it is confirmed that the spherical magnetic body can be held and transferred successfully without contact point by the magnetic robot hand with the prototype gap sensor system.
The virtual reality (VR) technology which is applicable not only in an image world but also in a audio and a tactile world is extremely useful. This paper deals with a newly developed telephone system applied the VR technology. For the sake of operating a receiver and a push button of virtual telephone (VT), five degrees-of-freedom serial link manipulator and a magnetic sensor have been designed and manufactured. Also, in order to obtain a higher performance of pushing the button of VT, the difference between the depth in the VR world and the depth recognized by human has been examined by the experiment. The proposed VT system would be applied to the other type of information terminal, such as a facsimile, a word processor and a personal computer.
This paper discusses an algorithm to make quadruped machine walk dynamically and omnidirectionally following a real-time command. To produce a smooth body motion in a high speed walk, we introduce a new gait named “Intermittent trot gait, ” in which four-leg-supporting phase and two-diagonal-leg-supporting phase appear in orders. In this gait, pitching and rolling motion of the body is able to be suppressed to smaller level than that of the other gaits such as pace, bounce, and expanded trot gait, which is an intermediate gait of the crawl and the trot. Because, dynamic effects of the swinging legs are almost canceled each other. To realize the “Intermittent trot gait, ” new algorithms to decide a landing point and to plan a motion of the body are discussed. A landing point is decided by considering an operator command and also a conversion to the standard leg formation. On the other hand, a motion of the body is planned to produce a dynamic stability, and to follow a speed and direction command. By using these algorithms, a mechanical vehicle TITAN VI could walk omnidirectionally, smoothly following a real-time operator command.
We designed a new omnidirectional image sensor COPIS (COnic Projection Image Sensor) to guide the navigation of a mobile robot. The feature of COPIS is passive sensing of the omnidirectional image of the environment, in real-time (at the frame rate of a TV camera), using a conic mirror. COPIS is a suitable sensor for visual navigation in a real environment. Under the assumption of the known motion of the robot (COPIS), an environmental map of an indoor scene is generated by monitoring azimuth change in the image. We did several experiments in the simple indoor environment. The precision of obtained environmental maps was sufficient for robot navigation in such environment. In this previous experiment, to examine the potential of COPIS against effects of observational errors in real-time navigation, we simplified the image processing method and the experimental environment. In this paper, we improve the image processing method for extracting and tracking vertical edges, taking care of reliability, and also evaluate the effectivity of omnidirectional image sensor COPIS in a real indoor and outdoor environment.
It is important to set landmarks when a robot needs to navigate in a certain environment and to measure its own location; however it has not been discussed well how to arrange landmarks in the optimal way until now. We, therefore, demonstrate the possibility of design of the optimal arrangement of artificial landmarks. First, we make it clear the optimization problem with regard to the arrangement of landmarks. Second, as one example of the artificial landmarks, we introduce a signboard system. Third, we propose a methodology to evaluate arrangement of landmarks assuming an obstacled rectangular environment. Fourth, we apply the algorithm of spatial competition to the optimal arrangement of landmarks. Finally, simulation results indicate that even though our method is not global optimization approach we can obtain closely optimal and practical arrangement with the minimum number of landmarks which satisfies the necessary condition.
A parallel processing scheme is described for robot control computation on MIMD shared memory multi-processor model. Since dynamic control law of robot arm usually contains a large amount of operations compared with the number of available processors, it is important to consider not only the effect of parallel processing but also the efficiency of serial processing in each of the processors. In order to obtain such a desirable solution for the complex scheduling problem, optimization and quasi-optimization algorithms are proposed. The excellent optimization algorithm is based on a branch-and-bound method. On the other hand, the practical quasi-optimization algorithm is based on a partial enumeration method which effectively combines the optimization algorithm and the approximation algorithm (GCF/LPT) ; that was recently developed by authors. The proposed scheduling algorithms are applicable to parallel processing of any kinds of control laws represented by sum of products. The test results on the dynamic control computation of robot arms demonstrate the usefulness of the algorithms.
A robust vision system against the change of lighting condition for industrial parts positioning has been developed. A wide dynamic range vision sensor, we had developed previously, was employed to avoid the saturation of images of the objects. Performance of the developed system with the wide dynamic range vision sensor was investigated experimentally in comparison with that of the system with a conventional video camera. The experimental result has confirmed the effectiveness of the developed system, revealing that the dynamic range expansion of the video cameras is very effective for realizing robust robot vision systems.