2. Technical Specifications

2.1 Robot Platform

Hardware platform: TurtleBot 4 (Create 3 differential-drive base) with integrated sensing stack.

  • Create 3 base: wheel encoders, IMU, hazard sensors (bump, cliff, wheel-drop, etc.)
  • 2D LiDAR: for obstacle detection and 2D mapping inputs
  • RGB-D camera: for OOI detection/tracking and depth-based range estimation
  • Compute: onboard (Raspberry Pi / platform default), with optional offboard workstation if explicitly documented

Simulation environment (for early integration):

  • TurtleBot 4 simulation in Gazebo/Streamlit (used for early pipeline verification and regression tests)

2.2 Kinematic Model

Declared model: Differential Drive (unicycle-equivalent planar model)

Let robot pose be x = [x, y, θ]ᵀ and control inputs be u = [v, ω]ᵀ:

  • ẋ = v cos(θ)
  • ẏ = v sin(θ)
  • θ̇ = ω

Wheel mapping (for reference; r = wheel radius, L = wheel separation):

  • v = (r/2)(ω_R + ω_L)
  • ω = (r/L)(ω_R − ω_L)

2.3 Perception Stack (Sensors Used)

Required baseline sensors:

  • 2D LiDAR: obstacle geometry, local collision checking, 2D mapping support
  • RGB-D camera: OOI detection/tracking + depth-based range to target
  • IMU: angular rate/orientation aiding for stable heading estimation
  • Wheel encoders/odometry: incremental motion estimation for navigation and tracking stabilization

Safety and hazard sensing:

  • Bumper switches: contact detection → immediate stop / back-off logic
  • Cliff sensors: edge detection near docks/ramps → stop and retreat
  • Wheel-drop / stall / slip / kidnap detection: triggers safe halt and recovery protocol

2.4 Additional Sensors / Safety Additions

  • Fiducial markers (AprilTag) on OOI:
    • Justification: robust ID + relative pose for reliable follow control under occlusion/lighting variability [1].

    • Justification: non-line-of-sight ranging cues to improve reacquisition in shelving occlusions.

  • Short-range ToF/proximity sensors around blind spots (optional):
    • Justification: mitigates single-plane LiDAR limitations (low/high obstacles, near-base blind regions).
  • Operator annunciation (LED/beeper) (operational safety enhancement):
    • Justification: improves human awareness when robot enters follow/chase modes.

References

[1] E. Olson, “AprilTag: A robust and flexible visual fiducial system,” Proc. IEEE ICRA, 2011.


This site uses Just the Docs, a documentation theme for Jekyll.