Automatic Generation of Best Emergency Routes and Procedures on a Brazilian Frigate

Abstract

We describe prototype software we have written that addresses the serious problem of management of damage control on naval vessels. Human reasoning and judgment may be strongly affected by stress and panic when several simultaneous events occur in a damaged-ship environment. Thus it would be helpful in times of emergency to dispassionately calculate the best courses of action to follow and suggest them to personnel. We describe two tested modules for the decision-aid system of the new Niteroi-Class Frigate Damage Control System project: an emergency-route generator for people and casualties, and a smoke-extraction procedure generator. Our modules gather and organize sensor data and reports about rooms, gateways and equipment. They reason automatically to determine the best routes and procedures, and can present the damage-related information in a concise manner to users both in the command center and throughout the ship. This system is implemented as an expert system using the CLIPS shell. The ship is represented as a graph where the compartments are nodes and the doors and hatches are the edges between them; costs reflect the difficulty in passing from a compartment to another. For finding emergency routes for people, a costminimizing search is done using the current ship data. A different search is done for smoke extraction to determine routes for ventilation of the smoke, with processing that checks conditions and fills in details.