My own plan, from a similar starting position, is to buy one of these[1], and start tinkering with it until I have a good set of "hard problems" I've run into. Then, read the textbooks, with a motivated eye to solving those problems. I won't necessarily solve any of the Hard Problems, but I'll hopefully get a journeyman-level ability to converse with others in the field about those problems--and from there, it's just a hop, skip and a jump to making friends in the field and being tossed pithy journal articles to read.
The two obvious paths I can see are: a) quit job and get masters or even PHD in robotics; b) become rich and found a company like Elon Musk.