AIDRA: A Disaster Rescue Coordination System Integrating Multi-Agent Technology

AIDRA is a Python-based hybrid artificial intelligence framework designed specifically for high-risk disaster rescue scenarios. It integrates multiple AI technologies including A* search, constraint satisfaction problems, fuzzy logic, and probabilistic reasoning to implement functions such as ambulance route planning, casualty priority assignment, and resource scheduling.

Disaster rescue scenarios are highly uncertain: roads may be blocked at any time, fires may spread rapidly, and aftershocks may occur suddenly. These factors require rescue dispatch to make quick decisions with incomplete information. The design goal of AIDRA is to build an intelligent rescue coordination framework that can operate adaptively in extremely uncertain environments.

This project was completed by Abu Bakar and Adeel Yaqoob in the Artificial Intelligence Course (AIC 201) under the guidance of Dr. Arshad Farhad. The project was developed in Python, using Tkinter to build the graphical interface, and implementing a modular agent architecture through object-oriented programming.

The uniqueness of AIDRA lies in that it does not rely on a single algorithm, but organically integrates multiple classic AI technologies to form a collaborative intelligent decision-making system.

In rescue scenarios, time is life. AIDRA uses the A* algorithm to calculate the optimal path from the ambulance to the casualty, while dynamically avoiding dangerous areas and blocked roads. The algorithm follows the classic evaluation function f(n) = g(n) + h(n), where g(n) represents the actual cost already incurred, and h(n) is the heuristic estimate of the distance to the target. This combination ensures that the system can not only find the shortest path but also re-plan the path in real time in response to environmental changes.

Rescue resources are always limited—there are a limited number of ambulances, a limited number of medical kits, and a maximum road capacity. AIDRA models resource allocation as a constraint satisfaction problem, seeking the optimal resource allocation plan under the premise of satisfying all hard constraints. The constraints considered by the system include multiple factors such as ambulance availability, medical supplies inventory, road traffic status, and rescue priority.

The risk of disaster environments is difficult to describe with precise numerical values. AIDRA introduces fuzzy logic to handle this uncertainty. The system outputs three risk levels—LOW, MEDIUM, and HIGH—based on fuzzy inputs such as smoke concentration, structural vibration amplitude, and environmental hazard level. This processing method is more in line with the experiential judgment of human experts, enabling the system to make reasonable decisions even when information is incomplete.

In the face of unknown disaster evolution trends, AIDRA uses Bayesian-style probabilistic reasoning for risk assessment and disaster prediction. The system updates its beliefs about road safety and environmental risk based on Bayes' theorem P(A|B) = P(B|A)P(A)/P(B), providing probabilistic support for rescue decisions.

AIDRA has built-in lightweight prediction logic to assess the survival probability of casualties and determine rescue priorities. The system automatically calculates the rescue priority of each casualty by comprehensively considering the severity of the injury, rescue delay time, distance, and environmental risk level. This data-driven priority ranking helps rescue forces allocate limited resources to where they are most needed.