(1989)
Publication Type: Thesis
Publication year: 1989
Publisher: Carl Hanser
Edited Volumes: Fertigungstechnik - Erlangen 5
The present work should contribute to the development of models for the position control of industrial robots in three areas: kinematics, dynamics and position control through feedback decoupling.
At the beginning, the aid for the kinematics and dynamics of industrial robots, namely the body-fixed coordinate system, was explained. The theory of the input axis coordinate system introduced in Chapter 2 shows that the normal input axis coordinate system has many advantages.
It is advisable to use the normal input axis coordinate system for the kinematic and dynamic tasks of industrial robots.
The methods of solving the kinematic problems were first presented in detail using the example of the Stanford arm.
Chapter 3 also dealt with the kinematic singularity and the method for the treatment of the singularity. While the position singularity can be avoided in the design of the work space, the orientation singularity can occur at any point in the work space. Since the orientation accuracy plays a subordinate role in all manufacturing tasks, a certain component of the angular velocity of the end effector is neglected in the inversion of the Jacobian matrix in order to ensure the position accuracy.
In the following chapter, the recursive Newton-Euler formulation for the dynamic inverse system was derived again on the basis of the normal input axis coordinate system. For this purpose, an algorithm was developed that is less computationally complex than that based on the original formulation [42]. The implementation of this algorithm was implemented on a microprocessor of the multi-microcomputer system MMC 216. The measurement results show that the microprocessor software built on the basis of this algorithm, which was developed for the dynamic inverse system of today's industrial robots, satisfies the scanning time criterion of the industrial robot.
Compared to the conventional joint position control, the Cartesian position control has the advantage that the path deviation of the end effector is regulated directly in the Cartesian world coordinate system. This means that the accuracy of the movement of robots in the entire work area remains exclusive
near kinematic singular points, the same. If a device for measuring the position and orientation of the end effector could be used in the future, annoying non-linearities and elasticities in the drive axles and gears would be avoided by the Cartesian position control. However, the difficulty lies in defining the orientation deviation. This problem was solved by introducing Euler's parameters as an orientation deviation and in Chapter 5 a model of the Cartesian position control was built. The theory could be confirmed by a simulation.
The last chapter dealt with closed chain robots. First, a dynamic relationship was established between a robot with a closed chain and its corresponding cut open chain. Based on this approach, all models developed for robots with a simple chain can easily be extended to robots with closed chains.
All models developed in the present work are of great importance if a very powerful position control is required when using industrial robots for manufacturing tasks.