Simplest Self-Serving Robot

by Karl Castleton

Introduction

This website documents the simplest robot that can serve it own purpose. And like so many other biological entities this robot emulated, this robots purpose is to stay alive (a.k.a moving). In this case the robot uses the solar power it collects to move. When it does move it attempts to move into the brightest place it can find. In sometimes while actually running though the robot is moving so fast that it moves too far into the dark before it can change directions. See some of the video clips below. The construction has borrowed many lessons from BEAM robotics. See Beam Robotics in the WikiPedia. The main design goal was to insure that all components serve at least two purposes. For example the wire that makes up the "frame" is also the circuit "brain". The motors are the "legs" and the help angle the solar cells. The solar cells are a source of power and sense the appropriate direction to go.

The Simplest Self-Serving Robot is solar powered and has no off-switch (like biological organisms). It can be built from components that are readily available on the internet. The solar cells are from Silicon Solar. The first robot was made of solar cells from Radio Shack and two pager motors from parted out pagers. The second used two motor from "zip-zap" upgrade kits and solar cells from Silicon Solar. Specifically any .5V "high" current 600ma cell will do. Your own design really only needs to create enough voltage and current to get the motors moving. I have used even cheap DC hobby motors with gear boxes. I just had to buy higher current cells. The cells are usually constant voltage (.51V) but will produce higher and higher currents dependent on the surface area (and sunlight).

Construction


The construction is to simply solder the circuit in the diagram. The first thing that might jump out at you is that there is no active components (transistors, CPU, operational amplifiers) in the circuit. Well it is meant to be simple. Basically you have two circuits in the diagram. The two circuits are 1) a loop between R-Motor, Cell-M, and Cell-L (yes Cell-L) and the second is a loop between L-Motor, Cell-M, and R-Motor. Interestingly the amount of light of the left side of the circuit causes the right side to speed up. A similar process is used for the other side.

Cell-M is shared between the right and left circuit. This serves two purposes 1) it insures that only the differences in manufacturing between Cell-R and Cell-L is minimized and 2) Cell-M boosts the voltage from each solar cell to help get the motors moving initially. The picture
shows the placement of these components on the actual robot.

Results

The robots do try to stay in the sun-light. This is a good example of the robot avoiding a dark location. Here is another example. But many times this robot gets going too fast and runs into the dark before it turn out of it. Here is another example of it going too fast.

Conclusions

The robot does work, and is about as simple as can be reasonably be expected. Other experimenters are investigating whether you can build a robot that "lives" in a 2-D environment can have less that two mechanims for controlling direction to find "food". In a 3-D world we would expect that three mechanisms and detectors will be needed.

Another interesting fact about this simple robot is that, like a human brain the left side sensor drives the right side locomotion. So if the left side of the "brain" is damaged the right side stops working. This indicates to me an interesting connection to biological systems.

Future Work

I am currently experimenting with a robot with a gear box on each motor to see if the "too fast" behavior is reduced. It will be a converted wall hugging mouse robot. I am also considering adding a operation amplifier circuit (as a comparator) to sense the difference in light and then have the solar cells store the power a capacitor or battery to avoid the "too fast" behavior and be able to operate in the dark.