Choosing sensors for a Rescue Line robot: what actually matters

A Rescue Line robot is only as good as what it can perceive. Here's how each sensor category earns its place on the chassis, and the mistakes that cost teams points.

Line sensors

The workhorse is the reflectance array: a row of IR emitter/phototransistor pairs facing the floor. More elements give smoother position estimates for your PID. Key detail: raw analog readings beat digital thresholds, because they let you interpolate the line position between sensors and adapt thresholds per field via calibration.

Mounting height matters more than datasheet numbers. Most reflectance sensors have a sweet spot a few millimetres from the surface; mount them too high and contrast collapses, too low and they scrape on ramps and speed bumps.

Green markers and color

Intersection markers demand real color discrimination, a pure reflectance reading can confuse dark green with black. Small RGB color sensors placed left and right of the main array are the standard solution. Position them so they see the green square before the robot's turning point, not after: by the time you detect it, you need room to react.

Distance sensors

Obstacles and evacuation-zone walls call for range sensing. Common options: ultrasonic sensors (cheap, wide beam, struggle with angled surfaces) and IR time-of-flight modules (narrow beam, fast, precise at short range). Many teams combine both: ToF forward for obstacle detection, side-facing sensors for wall following in the evacuation zone.

Cameras

For victim detection in the evacuation zone, a camera changes the game: silver and black balls in an open arena are exactly what simple blob detection is good at. Embedded vision modules that run the detection on-board and output coordinates over serial are the pragmatic choice, you get vision without soldering a GPU to your robot.

The boring advice that wins competitions

Whatever you choose: wire it so it can't come loose (competition tables get bumped), log every sensor so you can replay a failed run, and calibrate on site. A mid-range sensor calibrated on the actual field beats a premium one calibrated in your lab, every single time.