Rescue and navigation robotics

Before committing to any single mechanism, we scored multiple concepts against weighted criteria for each robot — ease of manufacturing, ease of use, accuracy, maneuverability, and how many people each design could carry or how fast it could deploy. The arm-based rescue bot and the mega bridge design came out ahead and became our starting points.

Every mechanism went through multiple failed builds before we landed on something reliable. Here's the path for both robots, roughly in order.

Bridge concept 1

Bridge drop bot

Early concept using a lever/piston to push bridge material up and over into place. Simple in theory, but hard to control precisely with a single actuator.

Bridge concept 2

Rotating-arm bridge bot

Replaced the piston with servo-driven rotating arms to lower the bridge more predictably, sketched out with a top-view servo layout.

Bridge build, early

First bridge robot

Struggled to clear the initial speed bump — the front geometry wasn't tall enough to ride over it cleanly.

Bridge concept 3

Long bridge, redesigned

Reworked the bridge geometry with a stationary center section and lowering ends on brackets, aiming for a more stable placement across the gap.

Bridge build, final

Final bridge robot

Added a much larger cylindrical front wheel and lowered the servo mounts to clear the bump reliably, with dual servos handling deployment.

Rescue concept 1

Arm rescue bot

Used a pull-bar and conveyor system to move a walled passenger box toward the stranded figure. Mechanically complex for what it needed to do.

Rescue concept 2

Claw rescue bot

Swapped to a servo-driven claw meant to grip and lift figures directly. Cardboard and popsicle-stick mockups gave way to a 3D-printed version.

Rescue build, failed

Claw prototype, built and tested

Alignment and claw width made pickup unreliable in practice — the mechanism worked on paper but not consistently on the course.

Rescue concept 3

Swivel-mount attachment

Explored a hammer-style attachment on a swiveling base, weighted for balance. Ultimately too complex to fabricate reliably in the time we had.

Rescue build, failed

Arm bot, first build

A rotating tower with an arm to hold and secure the figure. Functional concept, but the pickup mechanism still wasn't consistent enough to trust in competition.

Rescue build, final

Sweep-arm rescue bot

Replaced the lifting/gripping approach entirely with a simpler servo-driven sweep arm that pushed figures into a laser-cut basket. Less elegant, far more reliable — this is what we ran in competition.

For DP2, a separate robot ran the maze with zero manual input. Control logic ran on three Sharp IR sensors with failsafes and a turn counter, mounted through a custom SolidWorks bracket.

• Mechanical: 3D-printed wheel and mechanism parts; laser-cut bridge and basket components
• Controls: Raspberry Pi Pico, Kitronik motor shield, Wi-Fi connection, phone-based remote control
• Software: Python scripts mapping controller inputs to wheel and servo behavior
• Validation: SolidWorks CAD and Excel motor-speed analysis before every build

Scoring rewarded speed — full marks up to 30 seconds, decreasing steadily after. Our 1:04 total run cost points on time even with every objective completed.

Raw footage from the competition run, not a polished highlight reel — includes the full sequence from bridge deployment through figure retrieval.

• Turned failed prototypes into simpler, more reliable mechanisms rather than adding complexity to fix them
• Combined CAD, digital fabrication, electronics, and software into one working system
• Used weighted screening criteria instead of picking concepts by gut feel
• Led a team of four through iteration, testing, and a live competition presentation