This project aims to develop a performant and adaptable controller for an inverted pendulum system using a digital twin representation.
The initial goal is to create a basic proportional-integral-derivative (PID) controller that can balance the inverted pendulum model in a physics simulation (Phase 1).
After the core control functionality is implemented, the focus shifts to improving the system architecture and flexibility (Phases 2 and 3). Cross-compilation support will be added to deploy the controller on physical hardware like a Raspberry Pi. The communication protocols will be enhanced to support more advanced interactions between the controller and simulation. Multiple control instances will be coordinated for increased robustness.
Later phases will further expand the capabilities of the system (Phases 4 and 5). Distributed and adaptive control algorithms will be introduced to make the controller intelligent and resilient. The communication protocols and interfaces will grow to support complex multi-agent control scenarios.
Throughout the development, the fidelity of the digital twin will improve to better replicate the dynamics of real inverted pendulum systems. This will ensure the controller can be seamlessly transferred from simulation to physical applications.
The project is currently in Phase 1. See ROADMAP.md for more details of what's to come
We welcome contributions to help advance the project! Please see the organization's README to learn more.