×

img Accessibility Controls

Research Projects Banner

Research Projects

Reconfigurable Intelligent Surfaces aided Integrated Sensing and Communication using Machine Learning Algorithm for Next-Gen 6G Communications

Implementing Organization

Principal Investigator
Dr. Valmik Tilwari
Indian Institute Of Information Technology, Guwahati , Assam
valmik@iiitg.ac.in
CO-Principal Investigator
Nil

Project Overview

The evolution toward 6G wireless networks is destined to promise unprecedented advancements in connectivity, including the goal of ultra-reliable, high-speed communication and potential applications. the combination of Intelligent Reflecting Surfaces (IRS) and Non-Orthogonal Multiple Access (NOMA) has arisen as a revolutionary method. IRS can dynamically change the phase shifts of its reflected signals, giving it the capability to reconfigure the wireless propagation environments and boost signal power and coverage without demanding extra power-consuming equipment. By integrating with NOMA, which allows multiple users to allocate the same frequency and time resources through power-domain multiplexing, this integration greatly improves the spectral efficiency of future 6G networks. In combination, IRS and NOMA enable scalable and energy-efficient communication systems, especially in areas where traditional line-of-sight links are not possible or reliable. These developments are further enhanced by Integrated Sensing and Communication (ISAC), which makes it possible for sensing and communication to coexist in a single system. ISAC is significantly more effective in localizing users and adjusting communications to their exact locations when combined with NOMA. By dividing users in the power domain, NOMA allows several users to share a spectrum, allowing ISAC to distinguish between users according to their spatial characteristics and power levels. The integration of IRS with NOMA improves ISAC's capacity to locate users and adjust the environment for optimal communication performance. the use of machine learning (ML) named Deep Randomized Neural Networks (RandNN) is essential to the optimization of these complex technologies. Through learning from the dynamic environment of the network, RandNN-driven algorithms can improve the accuracy of channel estimates, optimize NOMA power allocation, and adaptively control IRS setups. Overall, the proposed framework performance will be evaluated for average estimation error in user position for different values of received SNR, outage probability, achievable ergodic rate, throughput, and energy efficiency.
Funding Organization
Quick Information
Area of Research
Engineering Sciences
Focus Area
Electrical, Electronics & Computer Engineering
Start Date
27 Mar 2025
End Date
26 Mar 2028
Status
ongoing
Output
No. of Research Paper
00
Technologies (If Any)
00
No. of PhD Produced
00
Publications
00
No. of Patents
Filed : 00
Grant : 00
arrowtop
Latest Updates
Loading…