Learning to drive as humans do: Reinforcement learning for autonomous navigation

Today's article comes from the International Journal of Advanced Robotic Systems. The authors are Lun et al., from the Beijing University of Posts and Telecommunications. This paper has been a long-time coming. It looks like it's been in peer-review for two full years and was just finally accepted and published online at the end of October. In it the authors describe an alternative way to think about autonomous-car training. Their question is simple: How can you make the learning-process that autonomous-cars experience more like the way that humans learn to drive? Let's get into it.

DOI: 10.1177/17298806241278910