Adaptive Line-Following via Self-Tuning PID Control
Problem
Fixed PID gains fail when track curvature, surface friction, or lighting changes — robots need re-tuning by hand.
Methodology
Empirical gain-scheduling: profile the error signal in real time and switch between tuned parameter sets.
Hardware
Arduino-class MCU, 8-sensor IR array, N20 gear motors, custom driver board.
Software
C++ firmware, interrupt-driven control loop, serial telemetry logger.
Future work
Port the controller to a learned policy and compare against classical tuning on benchmark tracks.