Abstract:In recent years, the application of neural networks (NNs) in wireless communication has garnered widespread attention and proven successful. However, conventional learning-based NNs often suffer from poor calibration, meaning that they struggle to reliably quantify prediction confidence and lack proper uncertainty estimation. This limitation becomes especially critical for next generation communication systems, particularly in complex industrial scenarios with stringent reliability requirements. Previous efforts to enhance model calibration have primarily centered on modifying NNs’ training processes. However, these methods often demand significant computing resources, making them impractical for resource-constrained scenarios. In this paper, we investigate a distributed wireless communication system involving multiple users and propose a novel approach to improve model calibration. Our method focuses on enhancing calibration during the inference stage of NNs by introducing a power control mechanism. Notably, existing research indicates that many NNs exhibit overconfidence, i.e., the NN’s confidence exceeds its actual accuracy. Leveraging this insight, we exploit the inherent noise and fading in wireless systems to naturally reduce the NN’s confidence while preserving accuracy. To achieve this, we employ linear relaxation-based perturbation analysis (LiRPA) to approximate the relationship between the perturbed output and the input perturbation of the NN. Subsequently, we devise an optimization problem by leveraging the analyzed relationship and the definition of perfect calibration. It is aimed at finding the input perturbation that maximizes the probability of the model achieving perfect calibration. Finally, considering different channel conditions and a given specific modulation method, we derive the optimal transmission power based on bit error rate (BER) formula. Simulation results demonstrate that our proposed power control method exhibits significant advantages in model calibration compared to several traditional approaches.

Influence of Antenna Calibration Error on the Performance of CoMP Systems

Modeling of Antenna Calibration Error and Its Impact on Open-Loop Beamforming

Inter-Cell Antenna Calibration for Coherent Joint Transmission in TDD System

A Simple Over-The-Air Hardware Calibration Procedure In Tdd Systems

Semi-Blind Calibration and Performance Analysis for Cdma Communications with Smart Antennas

Retrieving Channel Reciprocity for Coordinated Multi-Point Transmission with Joint Processing

Coordinated Multi-Point Transmission With Non-Ideal Channel Reciprocity

Power Control for NN-based Wireless Distributed Inference with Improved Model Calibration

Relaying Systems with Reciprocity Mismatch: Impact Analysis and Calibration.

Coordinated Multi-Point Transmission Strategies for TDD Systems with Non-Ideal Channel Reciprocity

Impact and Calibration of Nonlinear Reciprocity Mismatch in Massive MIMO Systems

Performance Analysis of TDD Reciprocity Calibration for Massive MU-MIMO Systems with ZF Beamforming

Network Mimo System Calibration with Non-Ideal Channel Reciprocity

Performance of Cell-Free Massive MIMO Under Imperfect Channel State Information and Reciprocity Calibration

Effect of Amplitude-and-Phase Errors on the Performance of Anti-jamming Antennas and Its Channel Calibration

Hierarchical-Absolute Reciprocity Calibration for Millimeter-Wave Hybrid Beamforming Systems

RF Mismatches and Nonlinear Distortions in Cell-Free Massive MIMO: Impact Analysis and Calibration Performance Analysis

Mutual Coupling Calibration For Smart Antennas In Cdma Communication Systems

Low Complexity Channel Estimation in TDD Coordinated Multi-Point Transmission Systems.

Interconnection Strategies for Self-Calibration of Large Scale Antenna Arrays

Performance Analysis of Large Scale Antenna System With Carrier Frequency Offset, Quasi-Static Mismatch and Channel Estimation Error.