Information invariants in robotics. I. State, communication, and side-effects
Bruce Randall Donald · 2002
The basic tools and concepts behind information invariants are described. The authors illustrate by example how such invariants can be analyzed and derived. An analogy between information invariants and tradeoffs is developed. This leads to the consideration of kinodynamic situations, in which performance measures, planning complexity, and robustness (in the sense of resistance to control uncertainty) are traded-off. Without a performance (cost) measure, it is more difficult to develop information invariants. Measures of the information complexity of a robotic task are reported. Automata-theoretic results are used to explore invariants that trade-off internal state, communication, and external state. A precise way to measure the information that a compass gives an autonomous mobile robot is reviewed.>