Projects

Maze-Solving Rover Project

Robot

Project Overview

The Wall-E Vader project focused on designing an autonomous rover capable of navigating a predefined maze, localizing itself, avoiding obstacles, and completing tasks such as picking up and delivering a block to designated zones. The rover achieved SAE Level 5 autonomy by integrating obstacle avoidance, localization, and block manipulation into a single operational algorithm.


Key Features and Strategies

Obstacle Avoidance

  • Sensors:
    • Four ultrasonic sensors and two infrared (IR) sensors for precise obstacle detection and corner handling.
  • Challenges Addressed:
    • Replaced unreliable ultrasonic corner sensors with IR sensors for improved detection.
    • Enhanced calibration and redesigned sensor mounts for consistent readings.
  • Results:
    • Successfully avoided major collisions, enhancing the overall reliability of navigation.

Localization and Navigation

  • Algorithm:
    • Employed histogram-based localization combined with a compass for orientation.
    • designed a heat map to visualize the rover's localization status
  • Efficiency:
    • Localized within 1–2 maze squares of movement, leveraging predefined paths and headings.
  • Execution:
    • Hard-coded paths simplified navigation but limited adaptability to environmental changes.

Heatmap overlayed on top of the known maze design Robot

Block Pick-Up and Delivery

  • Detection:
    • Dual time-of-flight sensors identified blocks based on vertical offsets.
  • Manipulation:
    • Servo-powered rack-and-pinion claw picked up blocks, assisted by a funnel design to correct alignment errors.
  • Challenges:
    • Sensor misalignments and mounting inconsistencies were mitigated but not fully eliminated.
    • Hard-coded paths required precise block placement within the maze.

Time-of-flight sensor layout Robot

Integration

  • All subsystems were combined seamlessly into a unified algorithm.
  • Starting locations and transitions between tasks were carefully managed to ensure autonomous operation.

Final Results

  • Performance:
    • Completed the maze in 3 minutes 20 seconds, well under the 5-minute limit.
  • Reliability:
    • Executed navigation, localization, and block delivery with minimal errors.
  • Key Insights:
    • Accurate localization achieved via a compass-aided algorithm.
    • Gripper’s modular design allowed reliable block handling, though reliance on predefined conditions limited robustness.

Recommendations and Improvements

Mechanical Design

  • Improve structural integrity by replacing taped or glued components with robust mounts.
  • Balance weight distribution to reduce the need for frequent motor recalibrations.

Algorithm Enhancements

  • Transition from hard-coded paths to dynamic path planning algorithms for better adaptability.
  • Implement real-time angle corrections during navigation and block handling.

Electrical Optimization

  • Organize cable routing to simplify troubleshooting.
  • Incorporate encoder feedback to improve movement precision.

Sensor Utilization

  • Optimize sensor routines to reduce processing delays.
  • Enhance the block-finding algorithm to handle arbitrary block placements.
Return to Portfolio