The onset of Industry 4.0 has created a huge buzz around autonomous systems, particularly in the field of control design. Reinforcement learning has made a comeback in recent years due to the availability of better computing resources, and is making massive strides in research, especially with the surge of deep learning. Taking its origins from control theory, reinforcement learning is an area of machine learning concerning how a software agent interacts with the environment around it and attempts to maximise a cumulative reward.
The Reinforcement Learning Toolbox is a brand-new MATLAB add-on product to help you get started on your reinforcement learning journey. The toolbox ships with great documentation, and loads of examples for implementing controllers and decision-making algorithms for autonomous systems. Using MATLAB and Simulink’s rich development environment, you can design, test and deploy your AI-based control system to PLCs and Edge computing devices.
Watch our two-part video series created by our Vacation Work Students Daniel and Vicky to find out more about how reinforcement learning works and how to implement a practical controller with a reinforcement learning algorithm in MATLAB.