In our project, to develop robotic methods for deburring process, the main subjects of “Modeling of Deburring/Grinding Process”, “Development of Control Systems for Deburring/Grinding Process”, “Trajectory planning for Deburring Process” and “Testing the developed methods on robotic experimental setup” were discussed. In addition, the "motion primitives" method has also been studied to utilize the arm and hand movements of human experts in deburring processes. Hence, a human arm and hand analogy were performed to a specific configuration consisting of a parallel robot that is mounted to end-effector of an industrial robot.
Furthermore, deflection is one of the problems in grinding / deburring with serial robots’ systems. For this reason, the modeling of grinding forces has been concentrated on modeling of deburring/grinding process. In this regard, “Penetration Tests” have been studied. While this study, design of experiments techniques was utilized. After obtaining the results of the penetration tests, artificial neural networks and then energy methods were concentrated. For estimating the deflection of the cutting tool and making the necessary refinements during grinding process, the classical strength approach and the compensation of the control systems were combined.
Unlike trajectory planning for the edges of the part, the changes considered while processing in the regional trajectory planning should be made with the data obtained from the receivers. The main reason is the problem of flexing. "Movement primitives" which are mathematical substructure for the recording and use of flexible paths those human beings create through their experiences and receptors being examined and different studies have been carried out. In these studies, by inspiring from 1 and 6 DOF haptic devices, human hand and wrist movements and skills were modeled and developed.
A grinding process of maintaining the desired shape on products is the most similar prototype problem to the deburring process. In this respect, PID and Fuzzy logic controllers was developed. The sample surfaces were scanned with an assisted measurement setup, and the surface profiles before and after experiments were compared and a statistical study of success was investigated. Also, the compensation of tool tip deflection was considered and the angular surface profile resulting from deflection was eliminated.
