A Low-Latency Framework for Channel Estimation and Reconfiguration

Reconfigurable Intelligent Surfaces (RIS) have emerged as a key enabler for future mobile networks by facilitating highly controllable and energy-efficient wireless environments. However, conventional implementations rely on acquiring full Channel State Information (CSI), a process that incurs significant latency and overhead-especially in scenarios involving user mobility. This bottleneck poses a major obstacle to achieving real-time, high-gain beamforming. Motivated by this challenge, we pose a fundamental question: Can we bypass high-dimensional channel estimation by directly sensing low-dimensional user angles to enable real-time beamforming? We provide an affirmative answer by introducing a novel lowlatency framework that transforms the problem from channel estimation to parametric angle sensing. Leveraging a Maximum Likelihood Estimation (MLE) approach, the framework enables rapid and accurate user angle detection for real-time RIS configuration. To overcome the computational burden typically associated with MLE, we introduce a high-precision algorithm augmented with a coarse-to-fine search strategy, significantly reducing complexity while maintaining estimation accuracy. Moreover, we establish a complete sensing-to-communication performance validation loop that quantitatively links angle sensing accuracy to communication quality. Extensive simulations show that our framework achieves communication rates nearly matching the theoretical upper bound under perfect CSI and consistently outperforms conventional methods-particularly in highly dynamic scenarios with high user mobility and low Signal-to-Noise Ratio (SNR).