Posture change detection device and method of object clamped by mechanical arm

Abstract

The invention discloses a posture change detection device and method of an object clamped by a mechanical arm. The detection device comprises a mechanical arm clamping system. The mechanical arm clamping system comprises a mechanical arm and two mechanical fingers. The mechanical arm is a cross rod. The two mechanical fingers are vertical rods perpendicular to the cross rod and are mounted at the two ends of the cross rod correspondingly. The two mechanical fingers are used for clamping the object. A first sliding feeling sensor 1 and a second sliding feeling sensor 2 are embedded in one mechanical finger, and a third sliding feeling sensor 3 is embedded in the other mechanical finger. Each sliding feeling sensor is used for detecting the sliding displacement quantity and the sliding direction. Each sliding feeling sensor is connected with a storer through a corresponding A / D converter. The storer is connected with a DSP. The DSP is connected with a control computer through a USB interface. Therefore, quantitative detection of posture changes of the object is achieved. The detection method can be further used for continuous trace monitoring and tracking of the clamped object.

Description

technical field [0001] The invention relates to a posture change detection device and method of an object picked up by a manipulator, and relates to the field of robot control. Background technique [0002] The stable and flexible gripping of objects by robots in unstructured environments is one of the most challenging research areas in the development of robotics. In the actual gripping operation, the gripped object will inevitably have attitude changes such as tilting, slipping, and rotation, such as translation along the x, y, and z axes and changes in rotation angles around the three axes. If the posture change of the object to be gripped can be quantitatively detected in real time and fed back to the manipulator, the manipulator controller will make corresponding operations based on these real-time data, which can greatly improve the stability of gripping and the flexibility of operation. [0003] At present, there are mainly two types of attitude change detection meth...

AI Technical Summary

Problems solved by technology

[0003] At present, there are mainly two types of attitude change detection methods for the gripped object: one is to use the visual method, that is, to monitor the attitude change of the gripped object by arranging a camera outside the manipulator. Or the occlusion of the manipulator, which affects the scope of application. In addition, the detection effect is also sensitive to ambient lighting; the second is to use the tactile sensor method, that is, to judge the posture change of the grasped object through the tactile sensing information of the mechanical finger and the grasped object. Although this method can overcome the shortcomings of the visual method, its disadvantage is that it can only detect qualitatively, and can only detect whether the attitude change of the clamped object is translation, rotation or rollover, and cannot quantitatively measure the attitude change.

Images