http://arstechnica.com/science/news/...ots-think.ars/
This gives a fairly good intro into Bayesian probability and how robots can use it to navigate and make decisions about the world. It is fairly accessible. Also, if you don't understand the P(X = x | Y = y) stuff at the start, just keep reading. Though it is important to understand if you're trying to make a robot using those concepts, it's not critical to understand the rest of the article.