Chapter 14: Solving Robot Learning
So far in the book, we have covered many state-of-the-art algorithms and approaches in reinforcement learning. Now, starting with this chapter, we will see them in action to take on real-world problems! We start with robot learning, an important application area for reinforcement learning. To this end, we will train a Kuka robot to grasp objects on a tray using PyBullet physics simulation. We will discuss several ways of solving this hard-exploration problem and solve it both using a manually crafted curriculum as well as using the ALP-GMM algorithm. At the end of the chapter, we will present other simulation libraries for robotics and autonomous driving, which are commonly used to train reinforcement learning agents.
So, this chapter covers:
- Introducing PyBullet
- Getting familiar with the Kuka environment
- Developing strategies to solve the Kuka environment
- Using curriculum learning to train the Kuka robot
- Going beyond PyBullet...
