How we-worker/robot_localization_in_stationary_scenario works
Main Components
a. Hardware
• LIDAR Sensors (LD06, T-mini-Pro)
• STM32 microcontroller (driving wheels, stepper motors, robotic arm)
b. Software Processes (all Python scripts)
• Sensor Acquisition
• LIDAR_LD06_python_loder/CalcLidarData.py
• T_mini_pro_loder/CalcLidarData.py
• Data Processing & Localization
• LikelihoodField.py
• Location.py
• Location_simulation.py (log-replay mode)
• Communication
• Socket_fun.py (inter‐process IPC via TCP/UDP sockets)
• contorller.py (UART interface to STM32)
• Orchestration & Logic
• Main.py (high-level robot logic, ties everything together)
• Configuration & Data Stores
• config.yaml (parameters)
• maps/ (map files)
• logs/ (recorded LIDAR/serial logs)
• Helpers & Scripts
• run.bat, stop.bat (launch/terminate processes)
Relationships & Data Flows
• Sensor → Acquisition:
LIDAR devices stream raw bytes over serial → CalcLidarData modules parse into “ranges”, “angles”, “is_new” → push into a shared dictionary (multiprocessing.Manager).
• Acquisition → Localization:
CalcLidarData writes to Manager dict → Location.py and LikelihoodField.py pull data to compute robot pose.
• Localization ↔ Simulation:
Location_simulation.py uses saved logs + same localization pipeline for debugging; no live serial.
• Localization → Orchestration:
Estimated pose → forwarded to Main.py via socket_fun (TCP/UDP).
• Orchestration → Actuators:
Main.py decisions → contorller.py (UART) → STM32 → motor/arm hardware.
• Config & Maps:
All processes read config.yaml; localization loads maps/map1.pixil.
• Logging:
Every module optionally logs to logs/*.json for replay or debugging.
Architectural Patterns & Design Principles
• Modular, multi‐process pipeline: each major function runs as an independent Python process for isolation.
• Publisher/Subscriber style via shared Manager dict for sensor data and via socket_fun for inter‐process messaging.
• Separation of concerns: acquisition, estimation, decision logic, and hardware interfacing are distinct components.
• Config‐driven behavior: parameters in config.yaml steer algorithms without code changes.
• Simulation mode: reuse pipeline for offline debugging (Location_simulation.py).
Final Recommendations
• Be exhaustive: include every Python script and its role.
• Related modules together but keep data flow obvious.