Increasing the Automation Level of Serial Robotic Manipulators with Optimal Design and Collision-free Path Control
|Kustantaja||Tampere University of Technology|
|Tila||Julkaistu - 1 joulukuuta 2017|
|Nimi||Tampere University of Technology. Publication|
To increase the autonomous level of robotic manipulator, the optimal structure is not enough, but it is a part of the solution toward a fully autonomous manipulator. The control system of the manipulator is the main part of computer-controlled manipulators. A collision avoidance system plays an important role in the ﬁeld of autonomous robotics. Without collision avoidance functionality, it is quite obvious that only very simple movements and tasks can be carried out automatically. With more complicated movement and manipulators, some kind of collision avoidance system is required. An unknown or changing environment increases the need for an intelligent collision avoidance system that can ﬁnd a collision-free path in a dynamic environment.
This thesis deals with these fundamental challenges by optimizing the serial manipulator structure for the desired task and proposing a collision avoidance control system. The basic requirement in the design of such a robotic manipulator is to make sure that all the desired task points can be achieved without singularities. These properties are difficult to achieve with the general shape and type of robotic manipulators. In this research work, a task-based kinematic synthesis approach with the proper optimization method ensures that the desired requirements can be fulﬁlled.
To enable autonomous task execution for robotic manipulators, the control systems must have a collision avoidance system that can prevent different kinds of collisions. These collisions include self-collisions, collisions with other manipulators, collisions with obstacles, and collisions with the environment. Furthermore, there can be multiple simultaneous possible collisions that need to prevented, and the collision system must be able to handle all these collisions in real-time. In this research work, a real-time collision avoidance control approach is proposed to handle these issues. Overall, both topics, covered in this thesis, are believed to be key elements for increasing the automation of serial robotic manipulators.