So far, we have been looking at how a robot or a group of robots can handle a certain application. We began by creating robots, defining links, and programming robots to handle tasks in an appropriate manner for a certain application. We also learned how to handle multiple such robots for that application. What if we gave our robots intelligence like a human being so that they can sense, think, and act by understanding the actions they carry out in an environment? This chapter deals with one of the most important topics of machine learning called reinforcement learning, which might pave the way for artificial intelligence-based robot solutions.

In this chapter, you will be introduced to reinforcement learning, its usage with robotics, the algorithms used (such as Q-learning and SARSA), ROS-based reinforcement learning packages, and examples with robots (such as TurtleBot) in simulation.

In this chapter, we will cover the following topics:

• Machine learning
• Understanding reinforcement learning
• Markov Decision Process (MDP) and Bellman equation
• Reinforcement learning algorithms
• Reinforcement learning in ROS