This paper presents a method of mobile robot's autonomous navigation in indoor environments by a teaching and playing-back scheme using memories of an ounri-directional image sequence and robot's motion. The robot memorizes a sequence of environmental images and robot's poses during a teaching stage. And in the course of a playing-back stage. itt calculates the robot's position difference from the memorized and currently taken images, and then decides a trajectory to track the taught route, so that it can navigate autonomously to follow the given path to the destination. An omni-directional image sensor is used, and vertical edges arc extracted as feature lines in the environment for calculation of robot's position difference. The experimental results show the effectiveness of this method in corridors of bulidings.
A method for acquiring an internal representation of a vision-based agent by associating visual input with a tactile sensor is presented. Agents that act in the physical world always suffer from the constraints of embodiment. However, embodiment plays a very important role in the formation of visual functions. Cognitive psychological findings assert that vision does not work without tactile motion experiences. Thus, our method features a tactile sensor for acquiring an internal representation. An internal representation corresponding to the physical relationship between the agent's body and environment in the visual input is acquired as a utility function. The utility value is calculated by a dynamic programming method with rewards given in a tactile state and transition probability model obtained beforehand. We conducted two computer simulation experiments, one being obstacle avoidance in which the agent needs to determine whether it can pass through a gap between objects or not. The other is object recognition in which the agent can distinguish differences in object shape by tactile motion.
We propose a method of running pattern generation for a humanoid robot using the dynamics of a simple inverted pendulum. A dynamic simulation using a model of an actual humanoid robot shows that the robot can perform running by applying a generated pattern with slight modifications. From the simulation we evaluate the required performance of actuators for a real running robot.
This paper proposes a two wheels caster type odometer to estimate the current position and orientation of omnidirectional vehicles. The proposed odometer is composed of two passive wheels and one passive rotational axis. The two passive wheels rotate independently, and the passive rotational axis changes the odometer's orientation. Angular velocities of the two wheels and rotational axis are measured by three rotary encoders. In order to solve an encoder's wiring problem by increasing rotation of the rotational axis, a mechanism of a gear train is proposed. Using measured values of the three encoders, the current velocity vector of an omni-directional vehicle is calculated by kinematics of the proposed odometer. Using the velocity vector, the position and orientation of the omni-directional vehicle are estimated by dead-reckoning. An experimental two wheels caster type odometer is developed. Experimental results of localization by the developed odometer are shown in this paper.
This paper presents analysis and design of a flapping robot with membrane wings. We develop an analysis method of aerodynamic forces for flapping movements for the robot. The optimal aspect ratio for a flapping robot with the membrane wings is designed through the numerical simulation that considers not only the analysis method of aerodynamic forces but also the load of a motor and the mechanical structure between the motor and the membrane wings. The simulation results show that the flapping efficiency for membrane wings is better than that for rigid wings. The experimental results agree well with the simulation results based on the proposed analysis method of aerodynamic forces.
This paper proposes a method of detecting waving hands from images as a key technology for man-machine interface. FFT is applied to time series of intensity images. The images are converted to low-resolution ones, and FFT is applied to each pixel of the low-resolution images. The proposed method is robust to lighting condition and individual difference of skin color, because it doesn't use color information at all. Experiments show the stability and robustness of the proposed method. Additionally, the method is applied to an omnidirectional vision images and an infrared camera, so as to show the wide applicability of the method.
In this paper, we consider a stand-up motion from a face-up lying posture by means of swing-down movement of a leg. Because whether the motion is feasible or not depends on the angular velocity of the leg swing, we propose a method to calculate it approximately. This calculation is based on the phase plane analysis and on conservation of angular momentum. The main problem there is how to evaluate the motion during the rotation around hip, where angular momentum has an effect of the gravity. Here, we calculate it using the initial and final position of COG in the hip rotation phase. This method enables us to estimate a necessary angular velocity without solving dynamic equation. Finally, we examine the effectiveness of the method by simulation and experiments with our constructed robot.
This paper proposes a tactile sensor capable of automatically adjusting its sensitivity, depending upon the contact information. Suppose a strain gauge based sensor, such as multi-axis force sensor and tactile sensor. One big issue for those sensors is the signal saturation due to large input for a particular direction. The proposed sensor system can automatically avoid such a situation with combination of Automatic Gain Control (AGC) . The sensor is composed of a controller and an analyzer. The analyzer provides with the contact force and the controller tunes up the amplitude of each frequency. We discuss the basic design relue where two theorems are introduced for guaranteeing the working principle of the sensor with AGC. Experimental results are also shown for validating the basic idea of the proposed sensor.