AI-Driven Real-Time Acoustic Modelling for Better Audio Perception in Dynamic Environments

This paper proposes an AI-driven framework for real-time acoustic modelling that enhances audio perception in dynamic environments. The system combines feedback microphones, deep learning models, and adaptive acoustic panels to monitor and optimize room acoustics continuously. Convolutional and recurrent neural networks estimate reverberation and clarity metrics, while a reinforcement learning controller adjusts panel states for optimal intelligibility. Unlike static treatments, this closed-loop approach adapts to changing occupancy, noise, and source locations. The expected outcome is a robust, intelligent acoustic system with significant applications in education, healthcare, and immersive audio experiences.