Objective Linear chirped microwave (LCMW) signals, whose instantaneous frequency chirps linearly with time, have attracted significant attention due to their large temporal duration and bandwidth when generated using photonic techniques. Such signals have found widespread applications in radar, sensing, and biomedical systems. Among various approaches, the generation of LCMW signals by beating a seed laser with an optical frequency comb (OFC) produced by an acousto-optic frequency shifting loop (AO-FSL) has emerged as an active research topic. However, the LCMW signals generated by AO-FSL suffer from instantaneous frequency oscillation (IFO) on the order of hundreds of megahertz (MHz), primarily arising from two factors of the OFC: its non-flat spectral power that decreases progressively, and its sharp spectral edges that give rise to straight-edge diffraction and phase-capture effects. The IFO severely degrades the signal quality and significantly limit the practical applicability of LCMW signals. A novel approach combining spectral smoothing filtering with power suppression at specific spectral locations is proposed, and its effectiveness in suppressing IFO over the entire chirp period of the LCMW signal is numerically demonstrated.

Methods First, the time–frequency characteristics of the OFC generated by an AO-FSL and the physical mechanisms responsible for the IFO in the generated LCMW signals are investigated. Subsequently, a novel IFO suppression method combining spectral smoothing filtering with power suppression at certain specific spectral location is implemented in numerical simulations. Specifically, optical filtering is first employed to smooth the sharp spectral edges of the OFC, thereby mitigating the severe frequency oscillations at the starting and end of the chirp period caused by straight-edge diffraction and phase-capture effects. Then, power suppression is applied at the starting portion of the OFC spectrum to reduce the amplification factor of frequency oscillations in the mid-period segment of the LCMW signal. Finally, the IFO suppression performance of different filters with various parameter configurations, including tunable band-stop filters (TBSFs), fiber Bragg gratings (FBGs), and tunable band-pass filters (TBPFs), is investigated via simulations using OptiSystem and MATLAB.

Results and Discussions Firstly, the suppression of IFO in the LCMW signals generated by an AO-FSL is investigated(Fig.2). The LCMW signal exhibits a bandwidth of 40.000 GHz and a temporal duration of 12.500 ns, with an initial IFO of 520.109 MHz. Simulation results show that the IFO can be suppressed to 205.506 MHz by employing a 1st-order TBSF with a center frequency of 193.100 THz and a bandwidth of 10 GHz. Furthermore, the influences of the TBSF center frequency, bandwidth, and filter order on the IFO suppression performance are investigated (Fig.3-Fig.5). To reduce the fabrication complexity and cost of the filter, a FBG with a length of 100 mm, an effective index of 1.450, and a modulation depth of 10^-4 is designed and simulated for IFO suppression, achieving an IFO suppression down to 195.556 MHz with performance comparable to that of the TBSF (Fig.6). To further examine the effect of spectral smoothing filtering, cascaded TBSFs are employed to simultaneously smooth both edges of the OFC, resulting in an IFO reduction to 95.509 MHz (Fig.7). In addition, a 10th-order TBPF with a center frequency of 193.120 THz and a bandwidth of 30 GHz, exhibiting spectral characteristics similar to those of the cascaded TBSFs is investigated, with the IFO suppressed to 81.294 MHz (Fig.8). The influences of the TBPF bandwidth and filter order on the IFO suppression are also investigated (Fig.9, Fig.10). Finally, the effect of filter center-frequency deviation on signal IFO suppression is investigated (Fig.11). The TBSF and TBPF are barely affected at a frequency offset of 0 - ±1 GHz, but their IFO suppression performance degrades significantly with increasing offset, while the FBG maintains stable IFO suppression within a ±5 GHz offset range.

Conclusions A novel method for suppressing the IFO of LCMW signals is proposed. The method combines spectral smoothing filtering of the OFC generated by an AO-FSL with power suppression at specific spectral locations. By simulating the influences of the optical filter type, center frequency, bandwidth, and filter order through numerical simulations, the IFO of the LCMW signal is suppressed from 520.109 MHz to a minimum of 81.294 MHz. These results demonstrate that the proposed method can effectively suppress the IFO of LCMW signals and improve instantaneous frequency accuracy, thereby facilitating the practical application of LCMW signals.