| Summary: | 碩士 === 國立勤益科技大學 === 工業工程與管理系 === 102 === Robotic arm already has been widely applied in industry, such as transportation, automatic assembly, combination and other highly complex and precise works or dangerous works which are not suitable for manpower. Many large medical centers utilize robotic arms which are operated by manual control and combine the surgical type of robot in the microscopic image display system. However, the motion trajectory has dynamic characteristics of nonlinear coupling during the multiple axes motion of robotic arm which makes uneasy to achieve the requirement of precise positioning. Therefore, the adjustment of controller parameters becomes the very important factor on the precision positioning.
In the manufacturing process, the precision is the essential factor need to be considered during product processing as particularly most critical important in the precision manufacturing of semi-conductors industry. However, accumulated or repetitive errors are existed under the long-term multiple-axes operation of robotic arm which may proceed to affect the precision of robotic arm to have a few errors that can lead to processing failure and result in the poor quality of product. Therefore, it becomes a widely concerning and sustaining explorative issue for how to improve the precision. However, the current way of precision improvement is mostly using the controller and adopts the visual sensor which is added on robotic arm to improve precision while the system cost is increased.
This study used the applying robotic arm to combine with XX-Y positioning platform. Firstly, we used Taguchi-Method to find out the significant affecting variables which can affect the positioning precision of robotic arm; and secondly, conducted the experiments to those key factors by the utilization of response surface experimental design to find out the most optimized parameters of key factors; and then, used the Backpropagation Neural Network (BPN) to execute the verification. Finally, we compared the results between the 3 methods of Taguchi-Method, Response Surface Methodology (RSM), and Backpropagation Neural Network (BPN) and compared the results mutually.
The results of this study indeed has enhanced the improvement of positioning precision and has shown the possible quality enhancement can be achieved by using Taguchi-Method to combine with Response Surface methodology (RSM).
|
|---|
