So far, the most difficult technical challenge in the Miniatur Wunderland ist the ship control. Already before this project was started in 2004, most experts we asked were of the opinion that such a task was at least barely feasable. The aim was, and still is, to develop a system that routes a couple of ships simultaneously and fully automated through the Wunderland oceans. The team around chief-developer Gerrit Braun more than once thought that they made it but then new problems stood in their way during the last years.
The difficulties are not found at running the ships, meaning the electro-mechanical part and it’s precise control. This works perfectly well and even very complicated docking manouvres can be accomplished. Influencing factors like waves, currents, and reciprocal effects between passing ships are well under control.
Basis for an automated and fully computer controlled ship service is a permanent and absolutely precise detection of all ships under way in the North East Sea. Since so far there are no commercially available systems with the required precision on the market, we started the development of our own system in January, 2004. One has to differentiate the precision between navigating (running a ship on the water surface, including meets and other influences) and manouvering (mooring to the pier). The navigation doesn’t call for extremely high precision and doesn’t expose any problems.
Manouvering, however, requires a precision of 1 mm, which relates to the nature of water. Therefore, all values vary a bit and only by many precise measurements, the real position and actual speed of a ship can be detected with the required precision.
First, we tried an infra-red camera surveillance like it is used in systems that record movements of athlets. The different height of the ships and water levels, the distortion of the lenses, refraction of the light, and more factors required a voluminous software. Although six(!) fast PCs were allocated, this process proved to be too complex and was trashed after 15.000 lines of source code were written.
Somebody from the circle of those participating in this project suggested to use ultrasonic devices. 48 piezo loudspeakers would have had to be installed, each fitted out with it’s own processor and the detection would have worked using the cross bearing method. All problems that popped up during testing were solved by August, 2004, for instance: longer ways due to diffraction at reflections. During a test run in a 252-room a precision of 1 mm was achieved - a fantastic result. However, this result was completely destroyed by a simple little mishap: Gerrit dropped his walkie-talkie. The resulting ultrasonic waves disrupted the system.
In order to avoid errors resulting from external influences, tests with tonal series followed. It was proven that only every hundred signals an error occurred. Using 8 measurements per second, the required precision can be achieved and wrong measurements can be avoided using a plausibility test. This system, which has the advantage to work without mechanical parts, is currently prepared for a test run above the North East Sea. Parallel to this, an infrared-lasersystem with an accuracy of about 0.6 mm is developed.
Driven by a motor, a circle segment of laser light is emitted over the area to be controlled – the ships have sensors that report the presence of IR-light. Critical for this function to work properly is to maintain a constant revolution of the motor. A deviation of 1/100 degree at a distance of 10 meters equals 0.6 mm. 20 to 30 of these system will be required. This system with the disadvantage of mechanical parts will be put to a test in the near future.
The nearly finished software now has 75000 lines of source code. Currently the final tests of the system are running. Manual control of the ships using the computer system were already successful. It seems that after long years of development we finally have a working system.