To address key issues in the design of autonomous combat systems for drone swarms, a design method based on Multi-Agent systems is proposed. An Agent model for each node in the drone swarm is established along with its inference rules; in response to the need for modularization and generalization of the simulation system, an interoperable interface and the interactive relationship of autonomous combat for drone swarms are designed; simulation and inference validation of the drone swarm system are conducted. The simulation results indicate that the proposed design not only effectively conducts and completes a dynamic demonstration and validation of the entire process of autonomous combat network generation, swarm evolution, and performance evaluation, but also further assesses the collaborative combat effectiveness of the drone swarm through repeated random trials, summarizing the strategies and experiences of collaborative combat.In recent years, the construction of autonomous drone combat swarms using a large number of low-cost, lightweight [1] small and medium-sized drones has become an important development direction for modern drone swarms [2]; they can carry various electronic devices or weapon units, replacing single platforms, and overcoming the adaptability flaws to dynamic and complex combat environments through tightly coupled collaboration among individuals [3]. As the transformation of military information continues to deepen, various combat styles represented by network-centric warfare [4] and distributed lethality [5] are emerging, which have brought a tremendous impact on the traditional command and control methods of swarms, making research on command and control of drone swarms a hot topic in the military field. While using real wars to study related issues in command and control is ideal, the costs and consequences may be immeasurable and unbearable. Therefore, utilizing simulations to conduct in-depth research on organizational design, mission planning, and other command and control-related issues has become an effective means to solve the above problems. For any system simulation, the issue of model establishment must first be addressed; the command and control system and its operational environment are complex systems that span physical, informational, and cognitive domains, and how to model and simulate such a complex system has become a pressing issue. Traditional command and control systems, represented by organizational theory, are primarily designed based on effectiveness-oriented combat thinking [6], lacking the dynamic strategy design for the collaboration and confrontation among various elements within the system, simplifying the division of combat tasks to resource matching problems, and lacking consideration for actual command authority and command processes [7]. New command and control systems represented by complex networks [8] and intelligent agents [9] have provided a certain degree of networked description of organizational relationships but lack the design and description of command systems, command processes, and organizational rules of intelligent agents within the organizational structure. Therefore, to realize the autonomous combat of drone swarms, a distributed intelligent command and control system needs to be supported, which mainly includes intelligent technologies such as situational awareness, combat planning and decision-making, action control, simulation and training, and human-computer interaction. This article considers the particularity of the drone swarm algorithm and software deployment architecture; due to the distributed and decentralized characteristics [10] of drone swarms, their characteristics align with those of artificial intelligence Agents [11], making it easiest to model the distributed characteristics of drone swarms using Agent concepts. Therefore, this article proposes an Agent simulation modeling method suitable for simulated entities through the study of the characteristics of various entities in the combat space and explores and attempts how to manage and schedule Agent entity models, then builds a distributed Agent command and control simulation environment to study specific issues in the command and control field, verifying the feasibility of the simulation entity modeling method and management scheduling technology through the construction of an autonomous combat system for drone swarms.
Quick Access to Resources
Convenient Download, please follow theSpecial Knowledge public account (click the above blue Special Knowledge to follow)
Reply or send a message “MAUA” in the backend to obtain the download link for “Design of Autonomous Combat System for Drone Swarm Based on Multi-Agent” Special Knowledge
Click “Read Original” to learn about usingSpecial Knowledge, and access100000+ AI-themed knowledge resources