DDPG-Based Throughput Optimization with AoI Constraint in Ambient Backscatter-Assisted Overlay CRN

Area of Science:

• Wireless Communications
• Network Optimization
• Data Freshness Metrics

Background:

• Ambient Backscatter (AB) communications (ABCs) and RF-powered Cognitive Radio Networks (CRNs) address energy and spectrum challenges.
• Growing demand for high-throughput and timely data in wireless networks necessitates advanced optimization techniques.
• Age of Information (AoI) is a critical metric for data freshness in dynamic network environments.

Purpose of the Study:

• To optimize long-term throughput in AB-assisted overlay CRNs (ABO-CRNs) under an Age of Information (AoI) constraint.
• To investigate the trade-offs between throughput and data freshness in integrated ABCs and CRNs.
• To develop a robust optimization framework for dynamic wireless network environments with incomplete information.

Main Methods:

• Deep Reinforcement Learning (DRL), specifically the Deep Deterministic Policy Gradient (DDPG) algorithm, was employed for throughput optimization.
• Novel reward functions were developed considering time and energy allocation when AoI constraints are violated.
• Analysis of the impact of minimum throughput requirements and maximum allowable AoI on network performance.

Main Results:

• The proposed DRL-based scheme effectively optimizes throughput in ABO-CRNs while adhering to AoI constraints.
• Simulation results demonstrate that the ABO-CRN achieves throughput close to the throughput-optimal (T-O) baseline.
• The achieved AoI in the ABO-CRN closely approximates the AoI-optimal (A-O) baseline.

Conclusions:

• The DRL-based approach provides an effective solution for balancing throughput and data freshness in complex wireless networks.
• The integrated ABO-CRN framework demonstrates superior performance compared to traditional schemes, offering near-optimal throughput and data timeliness.
• This research contributes a novel method for optimizing wireless networks facing energy and spectrum limitations.