ItemBook Chapter

An Adversarial-Resilient Multi-Agent AI Framework for Autonomous Robotic Warfare Defense

Title
An Adversarial-Resilient Multi-Agent AI Framework for Autonomous Robotic Warfare Defense
Creator
Srividya, R.; Joseph, Rose Bindu; Dara, Vijaya Lakshmi; Bunruangses, Montree
Description
Next-generation AI-enabled defense is essential to deter enemy drones, swarms, and autonomous vehicles in contested and deceptive environments, with modern battlefields becoming increasingly dependent on autonomous robot platforms. To discover and dissect and eliminate hostile autonomous threats in real-time, this study presents an integrated Adversarial-Resilient Swarm Defense AI Framework (AR-SDAI) with a Spatio-Temporal Transformer, a Multi-Agent Reinforcement Learning Countermeasure Module, and a Hybrid Graph Attention Network. To identify hidden or counterfeit threats and improve defense against enemy attacks, the system begins with applying a Transformer-based situational awareness model to merge multi-sensor battlefield data to fuse. Autonomous defense drones are then controlled by a multi-agent reinforcement learning framework to perform actions of dynamic electronic jamming and optimal interception maneuvers in a swarm environment. Finally, the system can find unusual patterns and create human-understandable counter-strategies for human-in-the-loop control with the help of a graph-based explainability layer that models the interactions of adversary swarms as dynamic graphs. Compared to traditional rule-based and CNN-RNN baselines using experiments on a simulated Red-Blue drone warfare test benchmark, the suggested AR-SDAI is better by 23% in threat detection accuracy, 31% in swarm interception success rate, and 19% in response latency. With its provision of robust, explainable, and flexible AI capability for next-generation robotic warfare settings, the paper in general enhances the state of autonomous defense operations. 2026 Saurav Mallik, Sandeep Kumar Mathivanan, Basu Dev Shivahare.
Source
Robotics in Weaponry using Machine Learning and Engineering;pp.451-468
Date
01-01-2026
Publisher
CRC Press
Coverage
Srividya R., Department of Computer Science, M.S. Ramaiah College of Arts Science and Commerce Autonomous, India; Joseph R.B., Department of Mathematics, Dayananda Sagar College of Engineering, Bangalore, India; Dara V.L., School of Business and Management, CHRIST (Deemed to be University), Yeshwanthpur Campus, Bangalore, India; Bunruangses M., Department of Computer Engineering, Rajamangala University of Technology Phra Nakhon, Bangkok, 10300, Thailand
Rights
Restricted Access; Hardcopy may be available in the library
Relation
ISBN: 978-104094368-7; 978-104107444-1;
Format
online
Language
English
Type
Book chapter

Citation

Srividya, R.; Joseph, Rose Bindu; Dara, Vijaya Lakshmi; Bunruangses, Montree, “An Adversarial-Resilient Multi-Agent AI Framework for Autonomous Robotic Warfare Defense,” CHRIST (Deemed To Be University) Institutional Repository, accessed June 17, 2026, https://archives.christuniversity.in/items/show/24461.

Output Formats