Information Spaces: They’re Everywhere!
Department of Computer Science
Time and Place
Mauldin Auditorium (NSH 1305)
Refreshments 3:15 pm
Talk 3:30 pm
Planning and control problems that involve sensing uncertainty naturally live in an information space. The concept arose in the context of game theory and stochastic control, and has appeared as belief spaces in AI literature (POMDPs). Just as configuration spaces have been important for unifying virtually all path planning problems and approaches, information spaces can serve the same purpose for problems that involve sensing uncertainty. This talk will present a unified perspective on information spaces. Following this, many examples will be drawn from our work, and the works of others, to illustrate the power and importance of reasoning in terms of information spaces. The key to numerous successful approaches has been to characterize and simplify the information space in some way. Examples include sensorless manipulation, the Kalman filter, pursuit-evasion, and on-line navigation.
Steven M. LaValle is an Associate Professor in the
Dept. of Computer Science at the
For appointments, please contact Jean Harpley.