Planning for Agents with Changing Goals

One of the most important applications of planning technology is guiding robotic agents in an autonomous fashion through complex problem scenarios. Increasingly, real-world scenarios are evolving in ways that require intensive interaction between human actors and the robotic agent, mediated by a planning system. We propose to demonstrate an integrated system in one such problem that falls under the aegis of an urban search and rescue (USAR) scenario. We show a simulation of a run through one such problem where a mobile robot is given certain goals to achieve in a layout of interest, and discuss how the various capabilities of the planner are instrumental in achieving the agent’s goals.