This paper describes the development of an in-pipe mobile robot driven by pneumatic pressure. In recent decades, natural gas has come to be supplied to most residences through polyethylene pipes laid under the roadways. As a result, it is necessary to inspect these pipes regularly to check for gas leaks caused by earth loads and oscillations resulting from vehicles traveling on the roads. Herein, we discuss a prototype pipeline inspection robot fabricated based on the model of a green caterpillar. In nature, caterpillars move forward on stems and leaves by the undulating motion of somites, and cling to leaves using abdominal legs and an anal proleg. Our robot is structured in the form of three somites and four suction brakes. Each somite, which contains four pneumatic bellows actuators, plays the role of a caterpillar somite, and suction brakes play the role of abdominal and an anal proleg. We confirmed that the robot is capable of moving at a speed of 9.7 mm/s inside an acrylic pipe more 34 mm in inner diameter, and can produce a maximum traction force of 4.2 N. In addition, the moving speed mechanism is clarified by using a model of the robot.
A wire suspended manipulator is proposed for extending its working space. A manipulator is fixed on a base plate suspended by six wires. The base plate is completely constrained by the wires, but their tensions change according to the motion of the manipulator or force disturbances acting on the suspended part, which may cause undesirable vibrations or turnover of the base plate. To avoid such situations, four counter weights driven by DC motors are mounted on the base plate. These counter weights are controlled by the feedback signals from tension sensors of six wires to keep the wire tensions positive. After the relationship between wire tensions and the counter weight motion are analyzed, a control law to keep the desirable tensions is proposed and it is installed into the prototype of the wire suspended manipulator system. The results of some fundamental experiments prove the analyzed static performance of the system and also verify the effectiveness of the proposed control system.
This paper proposes a method for recording and reproducing of human motions with a different target object. A bilateral control is known as one of the effective methods to achieve recording and reproducing human motions based on both position and force information. However, it is known that robustness for a different environment is decreased by adding the force information to the recorded motion information. In this paper, we focus an application of “peg-in-hole”, and our goal is to reproduce the recorded motion even if material of a target block is different. First, information of the actuators are transformed into motion-modal space to treat human motion as one system. Then, we propose a reproduction method which omits the effect of force information from a target object. In the focused application, the applied force during the target object and external environment is important. By the proposal, the contacting force is modified and force difference on the basis of material different is compensated. Validity of the proposal is verified by experiments. By the proposed method, the recorded motion is reproduced even if weight and friction of the target object is different.
This study aims to develop an automatic process planning and NC program generation system based on machining feature recognition for a turning-milling machine tool with multi-turrets. Machining feature recognition is carried out based on delta volume decomposition. A delta volume which is a portion removed by machining is separated into element shapes which are recognized as machining features. The developed system uses a CAD system (SolidWorks API) to execute delta volume extraction, delta volume decomposition, machining feature recognition, turret assignment of machining features, toolpath length calculation, machining time prediction, process plan determination, and NC program generation. The system generates various possible candidates for the machining process plan, and evaluates them based on a criterion determined by users. In this study, the shortest machining process time is provided as a user's criterion. After determining the machining plan that has the shortest machining time, the system automatically generates an NC program according to the plan. The system is applied to a turning-milling machine tool with three turrets, and an experiment using the generated NC program is conducted to confirm that the proposed system works well.
Error characteristics of multi-axis machine tools such as geometric errors cause machined errors on machined surfaces of machined parts. To estimate the machined errors on machined parts, shapes of machined parts with machined errors should be derived in computer simulation. By representing shapes of machined parts as difference set between a workpiece and tool swept volumes, the machined errors on machined surfaces can be estimated. A tool swept volume is composed of tool swept areas on tool surfaces that form surfaces of the tool swept volume. Tool swept areas on tool surfaces change its shape depending on tool motion. In order to derive tool swept volumes without any restriction on tool motion, all shape types of tool swept areas should be classified as tool swept types for general tools in multi-axis machining. Hence, in our previous study, tool swept types of cylindrical surface (side surface of a flat endmill), tapered surface (side surface of a taper endmill) and planar surface (bottom surface of an endmill) were classified. In this paper, tool swept types of torus that represents a corner radius of a radius endmill are classified into 8 tool swept types.
This paper deals with the automation of grinding operation. In this study, the developed equipment extracts the bending vibration sound of a grinding wheel using a microphone. And, the discrimination of grinding state is discussed based on signals obtained from the bending vibration sound of a grinding wheel. As a result, the sound pressure of the bending vibration sound becomes high at the specific frequency of a grinding wheel. The bending vibration sound can measure the normal grinding force. In short, the bending sound pressure level at specific frequency can replace the evaluation for grinding state. And, Contact sensing of 1mm unit is possible by observing the bending vibration sound. At last, the automatic cycle of a grinding processing by observing the bending vibration sound was proposed.
Several methods for evaluating the motion accuracy of the rotary axes of five-axis controlled machining centers have been proposed up to now. Since it is known that particular motion errors exist around motion direction changing points, it is important to evaluate the behavior of the rotary axes around these points. In this study, the characteristic of axial displacement around the motion direction changing point of the rotary axis in a five-axis controlled machining center and the influence of this characteristic onto the machined surface are investigated. The dynamic behavior around the motion direction changing point of the rotary axis was measured by the eddy current type displacement sensor. And the axial displacement of the rotary axis was investigated from the measured results. A machining test to evaluate the influence of the axial displacement of the rotary axis on machining surface is also proposed. It is confirmed from the results that the axial displacement of the rotary axis exists caused by the rotational direction change and it has an influence to the machining surface. In addition, it is confirmed that the influence of the axial displacement of the rotary axis can be compensated by translational displacement of linear axis.