Deep reinforcement learning tries to imitate the way a human or other intelligent being might interact with a new environment: trial and error. It is born out of the culmination of the research of many fields such as computer science, psychology, neuroscience, mathematics, and more. Though it is uncommon to see RL in industry today, it’s potential for impact is huge.
Everything starts with training. Camera’s and other sensors observe the environment. Neural networks are trained to recognize the relevant objects, their status and their location. The training also includes learning how the environment responds to actions, this is the reinforcement learning part. Training can be done via simulation or in the real environment, for example by the observation (with cameras) of workers who are executing tasks that have to be robotized.
After deployment the neural network will improve itself through continuous learning from the resulting observations and rewards of the actions it demands and from possible changes in the behavior of the environment.