Brief

Researchers have developed a new framework called Point-Cloud-Assisted Tangent Gaussian Splatting (PC-TGS) to improve channel prediction in wireless networks. This method integrates sparse radio measurements with dense LiDAR-based geometry to extrapolate channel information to unmeasured locations. PC-TGS represents environmental scatterers as anisotropic 3D Gaussians and uses a tangent-plane projection for angular domain mapping, achieving better prediction performance and faster inference times compared to existing methods. AI

Researchers have developed a new model called LiQSS (Linear Quantum-Inspired State-Space) that aims to improve real-time forecasting for 6G networks. This post-Transformer design uses quantum-inspired tensor networks to achieve linear-time sequence modeling, significantly reducing parameter count and increasing inference speed compared to Transformer-based models. The LiQSS model was evaluated on a dataset for predicting Reference Signal Received Power (RSRP) and demonstrated substantial efficiency gains without compromising accuracy. AI