A model of FSM-based planner and dialogue supporting system for emergency call services

Abstract

The objective of an emergency call is to provide the operator with as much pertinent information about the emergency as possible, so that the best course of action can be taken toward its resolution. In this paper, we propose a dialogue planner to support operators who take emergency calls. The proposed system helps the operators to respond appropriately to a caller's report by planning potential follow-up dialogue. The decision-making process used by the planner to select from among the proposed actions is implemented as a finite-state machine. The planner assigns a weight to each path between the states, evaluates the current dialogue model, and then adjusts the weights accordingly. Therefore, the weights assigned to potential actions vary dynamically as the dialogue progresses. The proposed model guides the operators' decisions throughout each conversation by informing them regarding the action that has the highest weighting.