Python library for lightweight robot simulation, designed for navigation, control, and reinforcement learning tasks. 🎧

Most robotics simulators are powerful but quite heavy. Sometimes you just want a minimal environment where you can test navigation, control, or RL without a full-fledged physics engine.

IR-Sim exactly meets this need if you're working with robot navigation or AI.

Why it's worth paying attention to:

• Quick installation and simple configuration
• Scenarios are defined using regular YAML
• Real-time visualization via Matplotlib
• Built-in collision detection
• Suitable for AI and RL pipelines

What you can quickly prototype:

– Multi-robot collision avoidance
– Lidar-based navigation
– Dynamic scenes with moving obstacles
– Testing high-level behavior before moving to a heavy simulator

The project is open-source under MIT, actively developed, and already used in a number of scientific works on navigation and planning.

GitHub: https://github.com/hanruihua/ir-sim
Documentation: https://ir-sim.readthedocs.io/en

If you're a student, researcher, or just studying navigation algorithms, IR-Sim provides a convenient sandbox where you can quickly test ideas without installing bulky dependencies.

👉 @DataScienceN

The tutorial on Basketball AI for YouTube is finally out. How to analyze one play frame by frame using a CV pipeline for sports analytics. 
🤟 

More than 1000 hours of work condensed into 37 minutes. 

- player detection via RF-DETR 
- player tracking via SAM2 
- team clustering via SigLIP and K-means 
- number recognition via SmolVLM2 

The system can be reused as a base for your own projects (analytics, highlights, broadcasts, etc.). 

Watch here 

👉  @DataScienceN ❤️