Development of Two Line-Following and Maze-Solving Cars: Analog vs. Arduino
Two robotic cars solving the same maze — one built from analog circuitry, one from code
Same maze, two completely different machines
The task was to build a car that follows a line and solves a maze. The interesting decision was to build it twice, two different ways, and let the comparison be the point.
One car does it in hardware: reflectance sensors feeding comparators and op-amps that drive the motors directly, so the control law is literally the circuit. The other does it in software: the same sensors read by an Arduino, with the following and maze-solving logic written in code.
The analog car reacts; the digital car remembers
The analog car is immediate — there is no loop rate, no sampling delay, the motors respond as fast as the sensors change. But it only ever reacts to what it sees right now, and tuning it means physically changing components.
The Arduino car is slower to respond and needed its timing thought about, but it can hold state: remember which junctions it has taken, store a route, and improve on a second run. Maze-solving needs memory, and that is exactly where the analog approach hits its ceiling.
It was an early, very concrete lesson in choosing the right level of complexity for a problem — and my first real exposure to writing embedded code, which turned into the Arduino work in the verification device four years later.