Self-Interference-Resistant IEEE 802.11ad-Based Joint Communication and Automotive Radar Design

Aimin Tang,Songqian Li,Xudong Wang

DOI: https://doi.org/10.1109/jstsp.2021.3118888

IF: 7.695

2021-11-01

IEEE Journal of Selected Topics in Signal Processing

Abstract:The IEEE 802.11ad based joint communication and radar sensing (JCRS) has attracted great attention for vehicles in recent years. Existing studies mostly assume full-duplex radio for radar receiver based on perfect self-interference cancellation. However, the self-interference may not be fully cancelled in practice. In this paper, imperfect self-interference cancellation is considered. To combat the strong self-interference, a novel JCRS is developed based on IEEE 802.11ad orthogonal frequency division multiplex physical-layer (OFDM PHY) for multi-target sensing. More specifically, the received radar signal is designed to be synchronized with self-interference, and two radar sensing functions are developed for short-range sensing and long-range sensing separately. The synchronization with self-interference resolves the difficulty in self-interference calibration in short-range sensing and prevents the inter-subcarrier-interference (ICI) from self-interference in long-range sensing. The Golay sequences in the preamble are leveraged to combat self-interference for short-range sensing, thanks to the relatively strong reflection signal. For long-range sensing, the pilot signal and a few self-interference-free OFDM symbols at the end of the data frame are leveraged to combat self-interference. Particularly, a novel pilot signal design is first developed to achieve subcarrier-wise pulse radar for accurate velocity and coarse range estimation, and then the self-interference-free OFDM symbols are utilized to achieve accurate range estimation. A strategy of frame transmissions is finally developed to integrate these two functions while balancing the performance between communications and sensing. Simulation results show that cm/s level velocity estimation and cm level range estimation can be achieved for up to 200-meter radar sensing.