RIS-Assisted Multi-User Localization in UAV-Enabled mmWave Wireless Networks

Localization techniques based on time of arrival (TOA) and angle of arrival (AOA) have been widely used in mobile communication systems and the localization accuracy can suffer from severe path loss and blockage. The mobility of UAVs and the signal redirection capabilities of reconfigurable intelligent surfaces (RISs) can be utilized to reduce these negative effects. In this paper, we propose a multi-user localization scheme in UAV-enabled millimeter-wave wireless networks, where the UAV localizes the chosen ground user by receiving the positioning reference signal from both the direct path and the reflected path via a RIS in each time instance. We derive the positioning error bound (PEB) based on the Fisher information matrix (FIM) of the unknown channel parameters and location parameters. Then we propose an alternating algorithm to minimize the maximum PEB among all users, by optimizing the UAV trajectory, the user scheduling, the UAV beamforming, and the RIS phase shifts iteratively until convergence. Furthermore, we formulate a robust optimization problem with imperfect knowledge of location parameters to minimize the maximum worst-case PEB, which can also be solved by the alternating algorithm. Numerical results show that the proposed alternating algorithm can improve the localization accuracy by more than twice compared to the scheme with a fixed base station and the scheme without RIS deployment.

I. Introduction

With the upgrading of wireless communication, localization is no longer a by-product of communication, but a key technology that enables advanced services in wireless communication systems [1]. On the one hand, higher carrier frequencies such as millimeter-wave (mmWave) and Terahertz, wider bandwidths, and larger antenna arrays in 5G and future 6G communications facilitate high-precision localization [2]. On the other hand, high-precision localization plays an important role in emergency rescue [3], automatic driving, and multiple verticals such as assisted living [4], industrial smart warehousing [5], and dynamic guidance in shopping malls [6]. The positioning technologies specified in the Third Generation Partnership Project (3GPP) Release 16 mainly rely on the time of arrival (TOA), time difference of arrival (TDOA), and angle information including the angle of arrival (AOA) and angle of departure (AOD), which can achieve sub-meter positioning accuracy [7]. At least two base stations (BSs) are required to participate and cooperate in these techniques. To reduce the cost of BS deployment, there has been much research in which localization is performed with one single BS, including localization methods based on received signal strength (RSS) fingerprints [8], [9] and methods exploiting the multipath in the radio environment [10], [11], [12]. However, the RSS fingerprint-based methods have a high operating cost when establishing the fingerprint database, and the RSS values of adjacent locations can be similar, which leads to limited localization accuracy. As for methods exploiting multipath, uncontrollable multipath and severe path loss under high carrier frequency have a negative effect on localization performance.