Argus: A modular multi-agent coordination framework for mission assignment, path planning, and simulation

Abstract

Coordinating multiple autonomous agents across simultaneous missions is difficult when agents differ in capability, position, battery, and health, and missions differ in priority, target, and required capability. This paper presents ARGUS, a modular software framework for multi-agent coordination, mission assignment, path planning, and tick-based simulation. ARGUS assigns agents to missions with a priority-ordered, capability-filtered greedy rule, plans obstacle-aware routes with an eight-directional A* search, and executes missions in a discrete-tick engine that enforces battery feasibility and prevents agent collisions. It supports point and bounded-area coverage missions, builds scenarios from OpenStreetMap building footprints, and exposes a FastAPI backend to a React frontend. The framework is verified by 485 automated unit, integration, and API tests. Results are functional rather than performance benchmarks. ARGUS is a centralized simulation, not a physical drone control or distributed system.