ISAC-Aided Channel Prediction in Limited Feedback OFDM System

Abstract

Integrated sensing and communication (ISAC) is a defining technology of sixth-generation (6G) wireless systems, enabling base stations to simultaneously serve user equipment on the downlink while sensing the propagation environment through monostatic echo reflections. In frequency division duplex (FDD) systems, the base station relies on periodic channel state information (CSI) feedback to perform precoding. However, between CSI feedback instants, the transmitter has no observations, leading to prediction errors that compound rapidly. We propose a sensing-assisted channel prediction framework that fuses sparse, quantized discrete Fourier transform coefficients from CSI feedback with dense echo measurements available at every slot, without UE cooperation, using a Kalman filter driven by an autoregressive state-space model. Our simulations across 3GPP-compliant geometric channel scenarios demonstrate that the echo-augmented channel predictor achieves higher prediction accuracy than CSI-only feedback when CSI feedback is sparse, establishing the ISAC-aided approach as a practical, UE-transparent mechanism for reliable channel prediction in next-generation FDD systems.