The SIG-Pack packaging robot box is for packaging small products into bigger units (cardboard boxes for example). The PGx groups the products in bigger units which are then grabbed and putted into cardboard boxes by the robot grip arm. The grouping unit is made of two band conveyors. The system is controlled by means of a servo controlling unit for the velocity loop and a SPS-axis unit for the position loop. Since the positioning of the band conveyors has to be very fast and accurate, the band conveyors begin to overshoot when operating at a high throughput rate. This could damage the product holders which are fixed onto the band conveyors.

As a consequene of the problems mentioned above, we decided to make all investigations about a controlling strategy at the Contraves model of the ZHW. This model might have a similar character to the PGx. To prove that, tests with the real model had to be done. The Contraves model is an electro-mechanical unit with an elastic shaft between engine and load. With the Contraves model, we tested different kinds of controlling strategies and finally compared them.

The fastest transition response is obtained with the Fuzzy controller and the classical cascaded controller with a Notch filter. In contrast to the classical controller, the Fuzzy controller is nonlinear. In addition, it is very complicated to optimize the Fuzzy controller. As a conclusion of this dissertation, we can state that a replacement of the position sensor to the load site and the use of a Notch filter might minimize the overshooting for the least expenditure on the mechanical unit.

The objective of this diploma work was to consider how to control a two-crane container transport system that operates without mechanical coupling. The study was done as part of a project for the firms Neuweiler and Tuchschmid. All control studies were done on a laboratory model. Emphasis was given to the study of the transportation operational states. During the two projects (Wil_PA1_01/2 and Wil_PA2_01/2) operating states for the two-crane system were determined. The transport system consists of two cranes that move a hoisted container. Two different hoisting configurations were studied. The operating state of each configuration is different. In the first configuration the transported container is supported between the cranes in a fixed manor. A single speed controller connected to a separate torque controller on each crane controls the operating state. In the second configuration the transported container is freely hung between the cranes in a manor that allows the container to swing. For proper operation of the system the controller must suppress the oscillations. Controller versions consisted of a classic speed controller, a fuzzy-logic angle controller, a difference master-master controller and a setpoint processing. Several different state controllers were developed. One speed state controller combined the overlaid classic difference controller following the master-master principle. A laboratory model was available for testing each controller versions. It was necessary to formulate the model mathematically and to build it in Simulink (simulation tool) to design the different controllers.

Studies to evaluate slip control were made using the physical crane model from project Wil_PA2_01/2. The two-level controller was determined to be the best for fulfilling this task. Control methods were tested using the laboratory model that can be used for the real two-crane system.

Who doesn't remember the wooden seesaw in the kindergarten? As we were children, there was always the problem to find a vis-a-vis, that had the same physical properties to enjoy the rocking. Today, as big children, this fascinating toy, still attracts our attention. This degree dissertation is based on a similar principle to the seesaw. The two children, which put the dynamic to the seesaw, have now been replaced by a driven cart. This cart can move on a straight track on the hole length of the seesaw and is forced to hold the System in balance.

To control the system, it is necessary to design a mathematical model. The better this model, the better it corresponds with the real system. Using this copy of the reality it is possible to design different kinds of control systems. We decided to make three different kinds of control: on hand classic control on the other hand two modern types of control, state control and fuzzy control. With state control we were able to find in a short time a good solution. Whereas classic and fuzzy control took a lot of effort take. Especially the fuzzy control was very time consuming and did not result in the same good quality of his opponents. Nevertheless, all methods have their advantages, say transient response, response on a distortion etc..

In a further step we enhanced the seesaw to a MIMO system by adding another cart on a parallel track to the existing one. Using the additional cart, we balanced an inverted pendulum which is attached to the cart.

Our first attempt to decouple the MIMO system did not work, so we designed a state feedback controller for the coupled system, that was able to stabilize the seesaw as well as the inverted pendulum.