Neural Network Based Framework for Passive Intermodulation Cancellation in MIMO Systems

2025-09-21

AI Summary

To address severe self-interference caused by passive intermodulation (PIM) in 5G and future MIMO-OFDM systems, this paper proposes a lightweight deep learning–based PIM suppression framework. Methodologically, it innovatively integrates depthwise separable convolution with dilated convolution to efficiently model nonlinear couplings between antennas and subcarriers; additionally, cyclical learning rates and gradient clipping are employed to enhance training stability and generalization. The resulting model contains only 11K parameters, significantly reducing computational overhead. Evaluated on a controllable MIMO testbed, it achieves effective suppression of third-order PIM, yielding an average power error reduction of 29 dB. To the best of our knowledge, this is the first work to systematically apply a compact neural architecture to PIM mitigation—achieving high accuracy, low complexity, and strong scalability—thereby offering a practical solution for real-world deployment.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Passive intermodulation (PIM) has emerged as a critical source of self-interference in modern MIMO-OFDM systems, especially under the stringent requirements of 5G and beyond. Conventional cancellation methods often rely on complex nonlinear models with limited scalability and high computational cost. In this work, we propose a lightweight deep learning framework for PIM cancellation that leverages depthwise separable convolutions and dilated convolutions to efficiently capture nonlinear dependencies across antennas and subcarriers. To further enhance convergence, we adopt a cyclic learning rate schedule and gradient clipping. In a controlled MIMO experimental setup, the method effectively suppresses third-order passive intermodulation (PIM) distortion, achieving up to 29dB of average power error (APE) with only 11k trainable parameters. These results highlight the potential of compact neural architectures for scalable interference mitigation in future wireless communication systems.

Problem

• Addressing passive intermodulation interference in MIMO-OFDM systems
• Overcoming limitations of complex nonlinear models with high computational cost
• Developing scalable PIM cancellation for 5G and future wireless systems

Innovation

• Lightweight deep learning framework for PIM cancellation
• Uses depthwise and dilated convolutions for nonlinear dependencies
• Employs cyclic learning rate and gradient clipping