Ultrasonic range finders have already been used successfully for indoor mobile robots. However, in outdoor environment, the sensor should be robust against noise from vehicle engines and other sound sources. We already proposed the a simple, low-cost, and robust ultrasonic range finder that is based on the use of correlation technique. The aim of this paper is to describe the fast calculating method to get the correlation value for the this robust sonar. In this method, a calculation quantity is drastically a little compared with conventional method. This method is to calculate difference of correlation using the characteristic of the shape of the transmitted signal, then to calculate the correlation using the difference. The experimental results using the ultrasonic range finder that equips 8 sets sensors and one controller without DSP show the ultrasonic range finder with the conventional method need plenty of time for the mobile robot until it shows measurement results, but that with the fast calculating method needs little time.
This paper presents a new type of holonomic and omnidirectional mobile robot. A steered driving wheel for holonomic and omnidirectional robots is proposed which has an offset distance between the steering axle and the wheel axle along a direction of wheel traveling like a caster which can use a conventional tires (a rubber tire or a pneumatic tire) . A kinematics of the wheel mechanism is analyzed and a new control strategy is introduced to avoid a nonholonomic constraints on the wheels. The driving wheel realizes an active ”caster motion”which enables 2DOF motions in the two dimensional plane. The concept of a holonomic robot with the proposed driving wheels is implemented and tested on a two-wheeled vehicle prototype. Experimental results show that the prototype robot realizes smooth holonomic and omnidirectional motions and precise mobility.
A multiple-layered ac-drive electrostatic actuator as a lightweight mechanical power source for physical aid machines is investigated. The actuator comprises a number of plastic sheets with three phase electrodes embedded. The sheets are stacked in two bundles and interleaved. Ac voltages applied on the electrodes induce traversing potential waves on the surfaces of sheets and interactions between the potential waves generate driving forces to slide the sheets. A 40-layer actuator is fabricated and tested. The actuator produces 80 [N], 4.8 [W] output at 1.6 [kV] . Output is in proportion to the number of sheets stacked. Efficiency is described independently to applied voltage as a function of power ratio, which is defined as output power under a voltage and load condition divided by the maximum output power at the voltage, and the peak efficiency of 55% is obtained at the maximum output condition. Analysis of power loss shows that the most portion of the loss derives from friction caused by attractive force between sheets. Output and loss constants that represent actuator characteristics are introduced and a model to predict the actuator performance is developed. Using this model, an actuator for physical aid machine is designed.
Integrated motion information management for use in highly accurate sensor-based robot systems that increases the precision and quality of robotic tasks is described. In this method, the sources of the recognition errors are easily identified by introducing the error-model using the information obtained by the proximity sensor as well as the information about the robot motion acquired by commanded simple testing motions, and these identified error-sources are used to reduce the recogniton errors. Motion reliability is determined based on the robot-motion information, and this reliability is used to reduce the influence of the recognition errors to the whole sensor-based robot system. As a result of this tuning the sensor-based robot system is set up to be close to optimal as possible. Experimental results from a typical visual path-tracking task showed that the tracking error is significantly reduced by using this motion information management system.
This paper describes an adhesion status detection sensor as well as crabbing motion for an in-pipe magnetic-wheeled robot. The robot targeted in this study can travel in any position within a pipe via magnetic-wheel adhesion, and can easily negotiate bends in pipe elbows and T-pieces. Moreover, the dual structure of the wheels ensures superior travel performance even over gaps like those commonly found in gas and other pipelines. However with no sensor in the robot for accurately judging its own adhesion status, we had a significant problem because the robot had to be attached to the pipe in order to travel. Steps also had to be taken because it looked like the adhesion status of the robot might well deteriorate under certain conditions, such as when traveling over weld lines or when reacting to avoid obstacles. To that end, we used a hall element in the present study to accurately detect adhesion status, and then set about confirming the efficacy of the element using an actual robot. We found with hall element output that inner magnetic-wheel rotation generated periodic vibration, so we eliminated this vibration as well. Here we are also proposing a new means of travel related to crabbing, which is the magnetic car-type robot's unique motion for mobility, and we set about confirming the efficacy of this means of travel using an actual robot. In more specific terms, the method maintains stability and provides mobility with a crabbing motion (much like Bogen moves in skiing) using a system that steers the front and rear wheels in opposite directions and offsetting their speed for forward and backward direction.
A human-demonstration based approach to the object motion design and the recognition of process state transitions in the insertion of a deformable tube is presented. First, human demonstration during the insertion of a deformable hose into a rigid plug is measured by use of a position sensor and a force sensor. Secondly, the measurements are analyzed with regard to the process states during the insertion. Human motion and recognition laws to detect process state transitions from force sensation are then extracted through the analysis of human demonstration. Finally, human motion during the insertion is transplanted to a mechanical manipulator to examine whether the human recognition laws are useful or not for the insertion performed by a mechanical manipulator.
For mobile robots' behavior such as navigation and map building, the estimation of own position is very important. In outdoor environment, the estimated position by odometry has the unpredictable error caused by traveling over an unexpected small obstacle or a bump under the wheels. In such case, the accuracy of the estimated heading direction is suddenly getting worse. In this paper, we propose a robust dead reckoning system fusing odometry with gyroscope in which the measurement mode is switched based on the difference of the angular rate measurements obtained from the odometry and the gyro. The proposed dead reckoning algorithm can reduce the ill-effect of road surface, estimate the bias drift of gyroscope and warn of a failure of sensor. As a result, we could realize the dead reckoning system which is reliable in outdoor as well as indoor environment.
It is desirable to have a mobile robot which can move and perform various tasks on irregular terrain, but current systems are not reliable or adaptable enough for practical use. A traditional method for climbing stairs is to use knobbed tires or tracks which can grip the edge of the stairs. But with this method usually only one knob on each side of the vehicle grips the stairs because the spacing between the knobs does not necessarily coincide with the distance between steps. In this thesis, a new idea crawler is developed for the stair-climbing. Blocks are attached to the face of the crawler-belt and can change shape to match the edge of the steps. Therefore, a large friction force is obtained by the soft deformation of the blocks at every point along the belt. After experimenting with various types of substances, blocks filled with powder were chosen for their desirable deformation characteristics. The results of this stair-climbing experiment prove that this crawler has better reliability than other crawlers when ascending and descending stairs. Also, this crawler can change its posture when climbing stairs or obstacles and this is also used to increase the turning efficiency of the crawler.
In this paper a bipedal robot with flywheel has been proposed. When a bipedal robot has no flywheel, it has been shown that the robot would have only one stable walking pattern to each initial translation speed and condition. This proposed robot with flywheel, however, can perform various stable stationary dynamical walking modes. Therefore, it may be easier to adapt for various walking conditions, for example, with slippery ground surface, a wide ditch, and frequent starts and halts. Several theorems have been derived from the equation of motion for the proposed robot that are on the fundamental characteristics of the walking modes. While the changes of walking conditions would influence on the stable mode of robot locomotion, stationary dynamical walking modes also influence on the interaction between the robot and the ground. To illustrate it clearly, a stationary dynamical walking mode map and a reaction patterns map have been proposed. The former would be a practical guide to the stable walking mode design; the latter would be helpful for designing a robust walking mode against slippery ground surface.