Changelog¶
All notable changes to the OpenSMC project will be documented here.
[2.0.0] - 2026-03-16¶
Added¶
- Reinforcement Learning: Full integration with Gymnasium for RL-SMC surface discovery and parameter optimization.
- Gymnasium Envs: 5 pre-built control environments for RL training.
- New Controllers: Added FixedTimeSMC, NFTSMC, and IncrementalHSMC.
- Estimators: Added Levant's differentiator and RBF-ELM observers.
- MkDocs Site: A comprehensive documentation site with auto-generated API reference.
Improved¶
- MATLAB Parity: Synchronized all Python classes with the corresponding MATLAB
+package. - Performance: Optimized the simulation core (RK4) for faster execution.
- Examples: Expanded the notebooks gallery with real-world trajectory tracking cases.
[1.0.0] - 2024-05-15¶
Added¶
- Core Framework: Base classes for SlidingSurface, ReachingLaw, Plant, and Controller.
- Standard Surfaces: Linear, Terminal, and Integral sliding surfaces.
- Standard Reaching Laws: Constant, Exponential, and Power-rate.
- Plant Models: Double Integrator, Inverted Pendulum, and Crane models.
- Metrics: Initial suite of performance metrics (ISE, IAE, ITAE).
- Dual Support: Initial support for both Python and MATLAB.
Initial Release¶
- The first open-source release of OpenSMC for community use.