VoxNav: Voice→Intent Digital Twin for Differential-Drive Robots

A modern, software-only digital twin of Arduino voice-controlled robot. It provides:

• A 2D simulator (pygame) for a differential-drive robot with simple kinematics
• A natural-language intent parser for commands like forward 30 cm, left 90, go to 6,3
• A grid-based A* path planner with obstacle map loading from YAML
• A HAL-style "motor interface" so your old Arduino MoveForward/SpinLeft/... semantics map to the sim
• Tests + GitHub Actions CI

Quick start

python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
python -m voxnav.ui.pygame_app --world examples/worlds/basic_10x10.yaml

• Click the window, then press / to open the command box. Type: forward 30 cm or left 90 or goto 8,2.
• Arrow keys also drive the robot (for quick manual checks).

Repo layout

src/voxnav/
  __init__.py
  control/intents.py         # text→intent parser
  planning/astar.py          # grid A* with 4/8-connectivity
  sim/robot.py               # kinematics + HAL (set_motor, move_cm, turn_deg)
  sim/world.py               # grid world, YAML loader
  ui/pygame_app.py           # pygame visualizer + command console
examples/worlds/basic_10x10.yaml
firmware/Aswin_VoiceRobot.ino
assets/                         # images/pdf for README
tests/test_astar.py
.github/workflows/ci.yml

Features

• Digital twin: demonstrate without hardware; add sensors (line, lidar) in software and test algorithms
• NLP-to-motion: from voice or typed phrases → structured intents → motion primitives / paths
• Planning-aware: goto x,y runs A* over a map and follows waypoints with proportional control
• Reproducible: deterministic sim loops, unit tests, CI

Next steps (ideas for issues)

• Integrate offline STT (Vosk/Whisper) and phrase grammar
• Add virtual sensors: line sensors for line-following, range rays for obstacle avoidance (Bug/DWA)
• Record telemetry to CSV + live plots
• Export ROS2 bridge (publish /cmd_vel from intents)

---