A Low-Complexity Method for FFT-based OFDM Sensing

OFDM sensing is gaining increasing popularity in wideband radar applications as well as in joint communication and radar/radio sensing (JCAS). As JCAS will potentially be integrated into future mobile networks where OFDM is crucial, OFDM sensing is envisioned to be ubiquitously deployed. A fast Fourier transform (FFT) based OFDM sensing (FOS) method was proposed a decade ago and has been regarded as a de facto standard given its simplicity. In this article, we introduce an easy trick — a pre-processing on target echo — to further reduce the computational complexity of FOS without degrading key sensing performance. Underlying the trick is a newly disclosed feature of the target echo in OFDM sensing which, to the best of our knowledge, has not been effectively exploited yet. Index Terms OFDM, radar sensing, multi-carrier, DFT, FFT, Decimation I. BACKGROUND AND MOTIVATION Orthogonal frequency-division multiplexing (OFDM) sensing has been a hot topic over the past decade. Given its wideband nature, OFDM sensing is attractive in many radar applications that require high range resolution, such as radar imaging [1]. Due to its high flexibility in waveform design and reconfiguration, OFDM is also a popular choice for software-defined radar [2]. Moreover, OFDM sensing is able to exploit the frequency diversity to survive interferencelimited scenarios, e.g., automotive radar networks. In a similar way, OFDM is known to facilitate multiple-input and multiple-output (MIMO) radars in achieving high angular resolution [3]. In fact, the initial motivation of introducing OFDM to radar sensing is to perform joint communication and radar sensing (JCAS) [4]. The proliferation of wireless applications and ever-critical spectrum crowdedness make JCAS highly popular nowadays. In a recently proposed perceptive mobile network (PMN) [5], JCAS is integrated into ubiquitous mobile networks. Thus, PMN is envisioned to achieve ubiquitous sensing, or more specifically ubiquitous OFDM sensing, given the crucial role of OFDM in modern mobile networks [6]. A fast Fourier transform (FFT) based OFDM sensing (FOS) method was proposed a decade ago by Sturm etc. [7]. Due to its low complexity and flexibility in accommodating classical radar sensing algorithms/theories, FOS has been regarded as a de facto standard for OFDM sensing since its publication. In short, FOS collects M consecutive OFDM symbols, each having N subcarriers. After some pre-processing, two batches of FFTs are performed, one batch along subcarriers and the other over symbols. A range-Doppler matrix (RDM) is then achieved, enabling target detection and parameter estimation. More details about FOS will be presented in Section