[Paper Review] Trajectory Optimization and Phase-Shift Design in IRS Assisted UAV Network for High Speed Trains

Yu Min Park, Yan Kyaw Tun+2 more|arXiv (Cornell University)|Jul 2, 2021

Advanced Wireless Communication Technologies15 references4 citations

TL;DR

This paper proposes an intelligent reflecting surface (IRS)-assisted UAV network to enhance wireless connectivity for high-speed trains (HSTs) by jointly optimizing UAV trajectory and IRS phase shifts to maximize the minimum achievable data rate. Using Binary Integer Linear Programming (BILP) for phase-shift optimization and Soft Actor-Critic (SAC) reinforcement learning for trajectory control, the framework achieves up to 19.9% higher data rates than fixed IRS and 4% higher than fixed phase-shift schemes, demonstrating superior performance in dynamic HST environments.

ABSTRACT

The recent trend towards the high-speed transportation system has spurred the development of high-speed trains (HSTs). However, enabling HST users with seamless wireless connectivity using the roadside units (RSUs) is extremely challenging, mostly due to the lack of line of sight link. To address this issue, we propose a novel framework that uses intelligent reflecting surfaces (IRS)-enabled unmanned aerial vehicles (UAVs) to provide line of sight communication to HST users. First, we formulate the optimization problem where the objective is to maximize the minimum achievable rate of HSTs by jointly optimizing the trajectory of UAV and the phase-shift of IRS. Due to the non-convex nature of the formulated problem, it is decomposed into two subproblems: IRS phase-shift problem and UAV trajectory optimization problem. Next, a Binary Integer Linear Programming (BILP) and a Soft Actor-Critic (SAC) are constructed in order to solve our decomposed problems. Finally, comprehensive numerical results are provided in order to show the effectiveness of our proposed framework.

Motivation & Objective

• To address the challenge of intermittent and unreliable connectivity in high-speed train (HST) communications due to frequent handovers and non-line-of-sight (NLOS) blockages.
• To enhance wireless coverage and data rates for HST users by leveraging intelligent reflecting surfaces (IRS) mounted on unmanned aerial vehicles (UAVs).
• To jointly optimize UAV trajectory and IRS phase shifts to maximize the minimum achievable rate across all HST users, ensuring fairness and reliability.
• To overcome the non-convex and NP-hard nature of the joint optimization problem through decomposition into tractable subproblems.
• To enable real-time, adaptive response to the rapidly changing HST environment using reinforcement learning and efficient phase-shift assignment.

Proposed method

• The system models a single-input multiple-output (SIMO) uplink scenario where a UAV equipped with IRS reflects signals from a base station (BS) to multiple HSTs, with perfect CSI assumed.
• The joint optimization problem is decomposed into two subproblems: IRS phase-shift optimization via a Transportation Problem (TP) formulation solved using Binary Integer Linear Programming (BILP).
• UAV trajectory optimization is formulated as a Markov decision process (MDP) and solved using the Soft Actor-Critic (SAC) reinforcement learning algorithm to maximize long-term reward (minimum data rate).
• The BILP model assigns reflectors to HSTs based on channel quality and path loss, ensuring global optimization of phase-shifts under resource constraints.
• SAC is trained to learn optimal UAV flight paths that dynamically adapt to HST movement, improving rate fairness and coverage.
• The framework integrates real-time phase-shift assignment with adaptive UAV mobility, enabling dynamic, low-latency response to HST mobility.

Experimental results

Research questions

• RQ1 How can IRS-aided UAVs improve the minimum achievable data rate for high-speed train users in non-line-of-sight environments?
• RQ2 What is the relative impact of UAV trajectory optimization versus IRS phase-shift design on system performance in HST communications?
• RQ3 Can reinforcement learning effectively optimize UAV trajectories in real time for fast-moving HSTs?
• RQ4 How does the number of IRS reflectors affect the system’s capacity and computational complexity?
• RQ5 To what extent does joint optimization of UAV trajectory and IRS phase shifts outperform fixed or random configurations?

Key findings

• The proposed method achieves a 19.9% higher minimum data rate compared to a system with a fixed IRS, demonstrating the effectiveness of dynamic phase-shift optimization.
• The system outperforms a fixed-phase-shift configuration by 4%, highlighting the importance of adaptive IRS beamforming.
• Reinforcement learning-based UAV trajectory optimization significantly improves performance compared to fixed-altitude UAVs, especially in high-mobility scenarios.
• The BILP-based phase-shift assignment ensures global optimality in reflector allocation, though computational cost increases with the number of reflectors.
• Optimal UAV altitude varies dynamically with HST position and path loss, with 200m not being optimal across all scenarios, confirming the need for adaptive control.
• The proposed algorithm achieves convergence in SAC training, with reward increasing and loss decreasing over episodes, indicating effective policy learning.