SLAM plot from Apriltag detected Square and Octagon motion
Zoom In
Planned path for Maximum Safety and Minimum Distance
Zoom In
Performance comparison between actual and planned trajectories for Roomba-like system
Zoom In

Project information

  • Category: Robotics & Autonomy
  • Project duration: Sep '25 - Dec '25
  • Team size: 2
  • Key Areas: SLAM, Navigation, Path Planning
  • Github URL: WaveRover Repository

WaveRover: Navigation, SLAM and Coverage

Technology(s) Used: EKF (Kalman Filter), A* Search, YOLO, AprilTags, Computer Vision, Subsumption Architecture

  • Waypoint Navigation & Visual Servoing: Developed a closed-loop controller for a wheeled robot using visual feedback. Performed camera calibration and distortion correction to improve pose estimation accuracy using AprilTags.
  • SLAM Implementation: Built an EKF Kalman filter-based SLAM system to estimate robot pose and YOLO-detected landmark positions, utilizing Mahalanobis gating for data association.
  • Path Planning: Designed grid-based A* algorithms focused on two distinct behaviors: Maximum Safety and Minimum Time/Distance.
  • Autonomous Coverage: Engineered a "Roomba-like" modular subsumption-based architecture for efficient area coverage and obstacle avoidance.