Autonomous robotics · Third Prize

FAST Maze Runner

An autonomous maze-solving differential-drive robot utilizing ultrasonic sensor arrays, dedicated power regulation, and dynamic pathfinding algorithms to solve unmapped mazes.

Role
Robotics Team Lead · The Physicians
When
April 2025
Mode
Physical systems
FAST Maze Runner autonomous robot prototype with ultrasonic sensors and electronics assembly.
The completed maze-solving robot before official trial runs.

01 / Context

The problem

Solving an unmapped physical labyrinth requires real-time sensing, precise turning actuation, and collision-free path decision logic under aggressive battery and timing constraints.

Approach

Our team built a differential-drive robot chassis with an Arduino controller, high-torque geared motors, and three ultrasonic sensors. The control software samples obstacle clearances, detects wall junctions, and executes a modified wall-following and junction-priority traversal algorithm.

02 / Contribution

What I contributed

  1. 01

    Led Team “The Physicians” across hardware integration, electronic testing, and competition trials.

  2. 02

    Designed the multi-sensor placement layout and motor driver wiring to eliminate electromagnetic noise.

  3. 03

    Developed and calibrated the maze exploration and junction-selection state machine in C++.

  4. 04

    Secured Third Prize at the FAST Robotics Maze Solver Challenge 2025 at Danang University of Technology.

03 / Decisions

Technical decisions

01

Separate motor and logic power

Used dual power lines and filtering capacitors to prevent MCU brownout resets when DC motors draw peak current during quick turns.

02

Three-point sensor array

Positioned front, left, and right ultrasonic sensors at 45-degree angled axes to detect diagonal corridors before committing to turns.

03

Dynamic speed throttling

Reduced forward drive speed when corridor margins narrowed below safe thresholds to maintain heading accuracy.

05 / Evidence

Outcome

  • Awarded Third Prize at FAST Robotics Maze Solver Challenge 2025.
  • Achieved zero collision faults across all official competition runs.
  • Validated end-to-end autonomous navigation combining physical assembly, sensing, and real-time control.
Next system / 05Vibe Coding Platform↗