Research On Slabstone Installation And Positioning Based On Force Feedback And Vision

Slabstone installation inside the large building is a kind of heavily physical labor, so develop and promote the automated slabstone installation tools and methods have great significance. In view of the low installation success rate caused by the incomplete sensing information, the sticky inaccuracy of the connector and other problems existing in the present slabstone installation robot, the study is carried out based on previous work, including automatic dry-hanging process optimization, installation strategy based on multisensor, and a derived issue—the contact state recognition for multiple square peg-in-hole. The main contents are as follow:On the basis of inquiring upon the traditional dry-hanging process of the indoor stone wall, the automatic slabstone dry-hanging process is established, and the connecter sticky method is confirmed. In order to provide foundation to large SE-type slabstone's standardization, the sticky inaccuracy constraints which insure the slabstones installed successfully are analyzed.In view of the drawback of the vision based positioning method which cannot acquire the position relationship between the parts, a force-based approach is introduced. On the basis of this, the thesis propose two methods for searching the hole—plane search and spatial search, which further improve the approach and process for automatic slabstone installation.The physical system for slabstone installation is established and the vision, force sensors are selected and fixed. The robot's kinematics model, vision measurement model and the external force detection model based on multiple one-dimensional force sensors are also established. These three models are unified to one model, which is the theoretical foundation for the implement of installation.The recognition of contact state is the key problem in the searching stage for"multiple square peg-in-hole"issue. The categories of the contact state during the two searching strategy are analyzed; two approaches to recognize the contact state is proposed, one is based on the position of the contact force and the other is based on force pattern recognition. Those approaches caused the searching hole strategy available and more importantly, it provide a reference for multiple contact point recognition which is a relatively universal issue.Finally, experiment based on the slabstone installing robot is implemented and the result verify the correctness of the contact state recognition algorithm and searching strategy.