(IIT Patna College, India.)
Phanideep Karnati is associated with the Department of Information Technology at IIT Patna College and serves as a CEO and GenAI Solution Architect. His professional interests include Information Technology, Generative AI, and intelligent solution architecture. He contributes to the development of autonomous AI approaches for disaster response and emergency management.
Phanideep Karnati, Pushpendra Kumar Sharma. In: Agentic Artificial Intelligence: Intelligent Autonomous Systems and Real-World Applications — ISBN: 978-81-69935-67-8. Pages: 183 - 206
This section offers an extensive structure for self-governing entities in crisis intervention and emergency administration. It analyses disaster-response frameworks, situational awareness, integration of multimodal data, damage evaluation, support for search-and-rescue operations, resource distribution, planning of emergency routes, assistance with evacuations, management of communications, coordination of healthcare, logistics, monitoring of infrastructure, unmanned aerial vehicle and robotic systems, edge-cloud frameworks, and coordination among multiple agents. The chapter also explores human-in-the-loop control, interpretability, uncertainty management, cybersecurity, privacy, ethical implications, resilience, and assessment approaches. A conceptual multi-agent framework is established wherein specialised agents cooperate in the domains of sensing, evaluation, planning, logistics, communication, and emergency response. Special attention is given to constrained autonomy due to the substantial impact that emergency decisions can have on human safety. Future investigative pathways concerning disaster digital twins, multimodal foundational models, autonomous rescue robotics, federated learning, robust communication, and cooperative multi-agent emergency systems are also examined. The section illustrates how self-governing agents can enhance emergency management effectiveness by facilitating information synchronisation, alleviating regular tasks, and aiding prompt, data-driven operational choices.
