Lightweight LSTM-Based Adaptive CQI Feedback Scheme for IoT Devices.

Area of Science:

• Wireless Communication
• Machine Learning
• Internet of Things (IoT)

Background:

• Exponential growth of IoT devices necessitates efficient radio resource management.
• Base stations require Channel State Information (CSI) for effective resource allocation.
• Current Channel Quality Indicator (CQI) feedback mechanisms increase feedback overhead.

Purpose of the Study:

• To propose a novel CQI feedback scheme for IoT devices using Long Short-Term Memory (LSTM) networks.
• To reduce feedback overhead and computational complexity for resource-constrained IoT devices.
• To enhance the efficiency of radio resource management in dense IoT environments.

Main Methods:

• Developed an LSTM-based model for aperiodic CQI feedback utilizing channel prediction.
• Designed a lightweight LSTM architecture to minimize memory and processing requirements.
• Conducted simulations to evaluate the performance of the proposed scheme against traditional periodic feedback.

Main Results:

• The proposed lightweight LSTM-based scheme significantly reduces feedback overhead compared to periodic reporting.
• The model achieves substantial complexity reduction without compromising performance.
• Demonstrated the effectiveness of aperiodic CQI reporting through LSTM-based channel prediction.

Conclusions:

• The lightweight LSTM-based CQI feedback scheme offers an efficient solution for managing radio resources in IoT networks.
• This approach mitigates the feedback overhead challenge associated with increasing IoT device density.
• The study highlights the potential of optimized machine learning models for resource-constrained IoT applications.