Recently, view-based or appearance-based approaches have been attracting the interests of computer vision research. We have already proposed a visual view-based navigation method using a model of the route called the “View Sequence, ” which contains a sequence of front views along a route memorized in the recording run. In this paper, we apply an omnidirectional vision sensor to our view-based navigation and propose an extended model of a route called “Omni-View Sequence.” Then we propose a map representation named the “View-Sequenced Map”which can represent a whole corridor environment on a floor in a building. Then we also describe the method for acquiring a View-Sequenced Map automatically based on the exploration of a robot in a corridor using both stereo and omnidirectional vision. Finally experimental results of the autonomous navigation and map acquisition are presented to show the feasibility of the proposed map representation.
Recently, robots have been expected to be used in various fields such as not only usual manufacturing industry but also medical welfare, agriculture, security system and so on. Especially, in the coming advanced aged society, the demand of robot will be more increased in medical welfare and human assist technology. In such a field, since robots must work together with human or contact with a human body directly, they are required safety and flexibility. In this study, in order to realize a flexible mechanisms satisfying the above requirements, a soft actuator has been developed. It is composed of a body made of silicone rubber and driven with a pneumatic power. Both the elasticity of rubber material and the air compressibility are effective to realize such a soft actuator. The developed planar type soft actuator has two layer structure built up with silicone rubber hemispheres and balls, which may be available to carry fragile and shapeless objects, furthermore, to execute flexible contacting tasks for human body. This paper describes a principle of operation, a control method and a fundamental control performance of the proposed actuator.
This paper proposes a new control scheme for a coordinated motion and force control of multi-fingered robot hand. Proposed control scheme can achieve simultaneous control of the position of a target object and a so-called ‘internal force’ which does not affect the motion of the object. Based on ‘knetostatic filtering method’ which eliminates a element of internal force from joint angle feedback signal, proposed controller can achieve internal force regulation for the fingers which have kinematical redundancy. Computer simulation results and experimental results illustrates the effectiveness of proposed scheme.
A laser micromanipulator is suitable for non-contact manipulation of a small object, such as a bacteria and microbe, in the closed space. However, laser trapping has two serious problems to control a living microbe. One is the damage caused by direct irradiation of high power laser to the target, and the other is the weakness of the trapping force, which causes swerving of the target from the desired transportation trajectory by disturbances. For safe and secure transportation of the microbe, we proposed to transport a microbe with micro tools trapped by the laser. Here we developed a measurement system of the laser trapping force using a quadrant photo detector (QPD) . We analyzed the measurement area of the QPD and calibrated the trap stiffness. Then we applied the bilateral control to the laser micromanipulation of the micro tool. The operator can feel the magnified force of the laser micromanip-ulation through the master arm. Finally, we demonstrated the indirect micromanipulation of the Yeast using laser manipulated micro tool.
A major problem in mobile robot navigation using a map is that an accurate map needs a lot of building cost. A framework of navigation using a roughly measured map would be a solution of this problem. The paper proposes a Topological-Geometrical map, or TG map for short, which permits inaccurate description, and a method of mobile robot localization on a TG map. A TG map is built by defining the relative poses between geometrical entities in the environment. The models of the geometrical entities are supposed to be predefined, and the relative poses between them can be as inaccurate as those measured by eye. These features allow the map-building cost to be small. The robot pose is represented in a local frame attached on each entity since the robot pose based on a global reference frame might be inconsistent because of the inaccuracy of the map. Errors in relative poses between entities are represented by probability density functions, and the robot pose is estimated using a variant of Markov localization which is augmented so as to incorporate map errors into data fusion process. Simulation and experiments show that the robot pose is estimated correctly on a TG map by the proposed method.
This paper proposes a medical care support system for computational behavioral data accumulation and interactive display by behavior understanding. The objective of this system is to save doctors' labor of recording their clinical behavior for diagnosis and treatment, and to help doctors to use the medical information more efficiently. The support system has three functions as follows: (1) understanding of a doctor's behavior by visual and audio processing, (2) accumulation of the results of understanding as the doctor's behavioral data, (3) displaying of the clinical information for the decision making by judging the content and timing from the results of understanding. In this paper, the overview of the proposed support system is introduced as well as development of a prototype system.
In this paper, we propose a method to estimate state variables based on time-series sensor information of flexible manipulator motion. It is also possible to identify dynamic model parameters of flexible manipulator based on the time-series sensor information. We have proposed two schemes for this identification, namely, driving torque error minimizationandstate variables error minimization. Characteristics of these two schemes are studied. Effectiveness of the proposed methods is verified by some experiments.
As many biomechanists indicate, tendon plays important role for running or jumping motion. It stores the kinetic energy as a potential energy during stance and also absorbs the impulse at touch down. Inspired by such biomechanical studies, we propose a new simple mechanical model for hindlimb of a dog to realize a robot which imitates dog running, and the hardware design of one-legged running robot, “Kenken” The robot has an articulated leg and uses two hydraulic actuators as muscles and a tensile spring as a tendon. Using an empirical controller based on the characteristic dynamics of the model, the robot has succeeded in planar one-legged running. Although the problem related to the stability at the higher running speed remains, the experimental results demonstrate that the proposed leg mechanism is effective for robotic running.