Model-based Reinforcement Learning

A toolbox for a Model-based Reinforcement learning technique.

Some features included:

• Database management and analysis
• A library of basis functions to be used in data-fitting stage
• 3D Graphical simulation for some benchmark examples
• Different model identification techniques

Model Learning

The toolbox also includes a benchmarking of different learning algorithms with various settings on multiple nonlinear systems. The resutls are generated for different nonlinear systems including:

• Dynamic vehicle system
• Quadrotor system
• Chaotic Lorenz system
• ...