Final Report: Arachnae II

The basic design used for Arachnae II was copied from its predecessor Arachnae I, created by Edmund Schierer. Arachnae II shall have somewhat the basic behavior of a simple insect, that is, move around and react in a certain way to external stimuli like obstacles, temperature and light. At the time of writing, the robot has the ability to walk around on flat surfaces with basic obstacle avoidance capabilities. Obstacle detection is done by the help of four whiskers-like feelers mounted at the head of the robot. If it senses an obstacle it stops, steps back, turns to the opposite direction the obstacle was detected and continues walking. There are a lot of other behaviors proposed in the functional specification which were not implemented yet. Some of these are responding to voice commands, speech synthesizer for spoken text output, sound direction detection to support 'follow-me mode', radio link to a host computer and some kind of radio control.

A considerable amount of work (around 3 months or so) was spent on the mechanical design and construction of the robot itself. Tough this wasn't the intention, since the fundamental design was being copied (theoretically) from Arachnae I, the lack of many years of experience in mechanical engineering culminated in some frustrating fiascos. The very first design of the robot's body was too weighty due to the use of aluminium and hard plastic. Some experiments with other materials resulted finally in the use of 4 mm thick ply-wood. This is both light-weight and stable enough to build-up even large frames (with the help of braces).
The next problem was the design of the legs. Insects normally have different front, middle and back legs [4, Michael B.Binnard]. They are each optimized for their specific role they play in locomotion. Thinking on insects, we know, most of them have powerful rear legs to provide most of the propulsion. On the other hand, making all legs the same simplifies both the design and manufacturing process. Since we hadn't time enough to experiment with different leg designs we decided to choose the latter alternative. To make the legs as light as possible (in other words, to increase Arachnae's power to weight ratio) we first tried to build the legs of aluminium profiles. This failed by some reasons. The next material we tried out was carbon fiber composites to save weight. It has a much higher strength to weight ratio than aluminium and can be made to nearly any shape you can imagine without expensive machining (says the theory). It seems that this manufacturing process needs also a lot of experience (which to get we hadn't time and patience). At the end we agreed on using the same material as we did for the robot's frame, namely repeatedley glued laminated wood cut and milled into a proper shape.
The main objective in the design of the control system was to make it reliable, easy to expand and fail safe. These design constraints was accomplished by using a many of several inexpensive controllers and microprocessors connected together by serial lines in a distributed network. This includes microcomputers for servo control, sensory control, topological map generation and behavioral algorithms.

Last updated: 26.6.1999