Feature/simulated perception

[blankey1337]

Summary

This PR transforms the AlohaMini simulator into a household robot development environment by adding three major capabilities: Simulated Perception, Semantic Navigation, and Task Sequencing.

Key Features

Simulated Perception (standalone_sim.py)

• 2D World Rendering: Replaced noise-based camera simulation with a rendered 2D top-down view of the robot and environment.
  Simulated Objects: Added support for virtual objects (Soda Can, Trash, Charger) with specific locations and colors.
• Perfect Detector: Implemented an "oracle" object detector that publishes ground-truth bounding boxes and labels for objects within the robot's field of view.

Room Navigation (navigation.py, nav_service.py)

• Semantic Map: Defined a map of rooms (Kitchen, Bedroom, Living Room, Hallway) with coordinate targets.
• Navigation Controller: Implemented a PID-based controller to drive the robot to target (x, y) coordinates while managing heading.
• Navigation Service: Created a ZMQ-based service that accepts high-level commands like go_to_room("Kitchen").

Chore Sequencer (chore_sequencer.py)

• Task Orchestration: Added a state machine to sequence complex multi-step tasks.
• "Clean Up" Chore: Implemented a demo task where the robot navigates to the kitchen, scans for trash, picks it up, and disposes of it in the hallway bin.

Dashboard Improvements (dashboard/)

• Visualizations: Updated the web dashboard to draw bounding boxes on the live camera feed.
• State Inspection: Improved the status panel to pretty-print complex JSON data (like detections and room coordinates) and handle image data gracefully.

Testing

• Verification: Ran the full stack locally (standalone_sim.py, app.py, chore_sequencer.py).
• Results: Confirmed the robot successfully navigates between rooms, detects the "trash" object, and updates its status on the dashboard in real-time.

alohamini_vision.mov