Research Projects

Extended Reality (XR) technologies like AR and VR impose high demands on network infrastructure. Next generation 6G/7G networks aim to address these with advanced communication paradigms. This project seeks to develop trustworthy AI-driven protocols for XR services that meet stringent quality-of-service (QoS) demands, ensuring reliability, scalability, and optimal resource utilization. Rationale: Current AI protocols lack robustness, real-time adaptability, and causal reasoning, failing to meet XR's dynamic QoS needs. By integrating reinforcement learning and causal inference, this project aims to create explainable, trustworthy, and efficient solutions, aligned with the 6G/7G vision for adaptive and reliable systems. Scientific Objectives: Develop a Testbed for XR Protocol Evaluation: Design a working setup to simulate XR traffic over network simulators and emulators, enabling performance benchmarking of existing protocols. Design AI-Driven Protocols: Leverage modern AI tools involving causality and reinforcement learning, to create scalable, context-aware protocols for XR services. Adaptive QoS Frameworks: Develop mechanisms that can flexibly prioritize latency-sensitive and bandwidth-intensive XR services, ensuring consistent performance in dynamic environments. Integrate Causal Models: Embed causal inference techniques into protocol design to identify and leverage cause-effect relationships, enabling transparent, explainable, and trustworthy AI systems. Hypothesis and Models: Causal AI protocols can enhance XR service efficiency and reliability by adapting to network dynamics. Models using network calculus, queuing theory, and reinforcement learning will predict system behavior and optimize resource allocation for dynamic environments. Main Experiments: Simulation Studies: Conduct extensive simulations on platforms like NS-3 and OMNeT++ to evaluate the performance of existing XR protocols under realistic traffic patterns, including stochastic and deterministic variations. AI Model Training: Implement causal reinforcement learning algorithms integrated with causal reasoning to handle uncertainties and optimize resource utilization. Prototype Development: Build a live testbed using open-air-interface (OAI) with software-defined networking (SDN) capabilities and real-world XR devices, such as Insta360 cameras, to validate protocols. Performance Evaluation: Compare and contrast developed solutions against existing ones. Significance : This research will transform communication systems, enabling seamless XR experiences in telemedicine, education, and automation. It bridges critical gaps in AI-driven systems and provides a roadmap for designing robust, scalable, and adaptive protocols for the 6G era. By integrating causality, the project establishes a new paradigm for trustworthy AI-driven communication systems.