My current position does not require much in the way of creative solutions so I have to think all the way back to a school group project for this. The goal of the project was to create a crude robot that could be programmed traverse a predefined path that was several hundred feet in length including numerous turns. Being typical students (i.e. poor) my group's resources were of course quite limited, as was the time we had to complete the project.
We had quickly decided that a tracked chassis such as one would find on a tank or bulldozer would be easier and more economical to implement than a something more akin to an automobile with steerable front wheels. Using assorted junk from our collective garages my colleagues in the group fashioned the tracked chassis out of scrap metal, cordless drills, lawnmower wheels, and tow straps.
My task was to create the guidance system. It would have to be adaptable to several different courses and easily programmed on the fly without a computer. It would have to be capable of being storing and implementing a program that would guide the robot along the numerous twists and turns of each several hundred foot long course.
The system I created would require the user to place the robot at the starting line of the course. The user would then set the robot to record the course and guide the robot along the course via a wired remote control. This remote had only two buttons. One button controlled the left track and the other controlled the right track. Pressing both would cause the robot to go straight while pressing just one would cause it to turn. Once the robot completed the course the user would end the recording. The user would then return the robot to the original starting position of the course, and orient it in the same direction as before. The user would then remove the wired remote and set the robot to execute the program for navigating the course.
I suppose the traditional solution might have involved some sort of microcontroller but I deemed it impractical to implement in the time and financial constraints that we had. Instead of a microcontroller, I used an MP3 player as the brains of the robot. The MP3 player used for the system had the capability to record a signal input to its 3.5mm input jack as a MP3 file. The signal it would record came from two independent tone generators created with 555 timers. Each of the buttons on the remote that controlled the robot's tracks also controlled a corresponding tone generator. As such, when either track was in motion the tone generator created a tone for the MP3 player to record. A lack of motion by the track in question would be recorded as silence. During playback of the recorded stereo MP3 file, the output signal from the MP3 player's headphone jack would trigger each track as appropriate when ever there was a recorded tone. When there was a tone on both channels the robot would go straight. When there was a tone just one channel the robot would turn.
Although the system was quite crude, it worked as intended and my group received a grade of A for the project as well as kudos for original thinking.
