The basic concept of an information space can be traced back to work of Kuhn  in the context of game trees. There, the nondeterministic I-state is referred to as an information set. After spreading throughout game theory, the concept was also borrowed into stochastic control theory (see [95,564]). The term information space is used extensively in  in the context of sequential and differential game theory. For further reading on I-spaces in game theory, see [59,759]. In artificial intelligence literature, I-states are referred to as belief states and are particularly important in the study of POMDPs; see the literature suggested at the end of Chapter 12. The observability problem in control theory also results in I-spaces [192,308,478,912], in which observers are used to reconstruct the current state from the history I-state. In robotics literature, they have been called hyperstates  and knowledge states . Concepts closely related to I-spaces also appear as perceptual equivalence classes in  and also appear in the information invariants framework of Donald . I-spaces were proposed as a general way to represent planning under sensing uncertainty in [68,604,605]. For further reading on sensors in general, see .
The Kalman filter is covered in great detail in numerous other texts; see for example, [226,564,912]. The original reference is . For more on particle filters, see [45,293,350,536].
Steven M LaValle 2012-04-20