Search

The laser frequency stabilization technology is very important in the development of laser applications, and the improvement of this technology has tend to further expectations for portability and low cost, in addition to the realization of strict frequency control. In this context, we construct a dual-cavity laser frequency stabilization system based on an intracavity dual-mode self-reference mechanism. With this approach, we achieve remarkable results with a laser frequency drift of about 6.6 kHz and an Allen deviation (ADEV) of 10−13 levels over an integration time scale from 0.01 s to 1000 s. Notably, the solution achieves one of the best results without the need for an external reference structure. This work provides a way to realize wavelength-stabilized laser sources by interlocking two DFB lasers. This approach is distinguished by its simplicity and cost-effectiveness, making it a promising option for a variety of applications outside of the laboratory environment.

An absolutely consistent fiber-optic phase synchronization scheme based on fixed-phase-reference optical active compensation is proposed. By measuring the time interval of time pulse and the phase difference of frequency signal, and then using the optical delay line to compensate the phase change in real time according to a fixed phase reference, the phase period ambiguity and the phase-lock control initial randomness are eliminated, realizing a stable and repeatable fixed phase difference between different sites. Then the phase synchronization performances of stability and absolute consistence have been demonstrated through experiments in different scenarios. The phase difference fluctuation within 10000 s is only ±0.0006 rad, showing an ultra-high stability. Under the total of 5 different operations, including the signal source restart, system restart, adding 20 m additional fiber, adding 1 km additional fiber, and adding 2 km additional fiber, the maximum inconsistency of the mean phase difference is about 1% of one full cycle. The proposed and verified absolutely consistent phase synchronization method can well maintain the coherence in transferring frequency signal, having important application prospects in radar network and other time-frequency-phase synchronization-based coherent detection.

Optical amplification, which is necessary for extending the distance of optical frequency transfer, has a significant impact on the transmission performance. In this paper, we propose a high-gain optical injection locking amplifier (OILA) with low phase noise based on an optical injection phase-locked loop for the phase-coherent optical frequency transmission. The high-gain OILA can provide more than 75-dB gain and ensure that the input carrier frequency fractional stability can reach 8.2 × 10-20 at an averaging time of over 100 s. We then transfer a commercial narrow-linewidth laser over a 220-km fiber link with only one amplification step placed at the remote end. After eliminating the noise induced in the fiber link, the fractional frequency instability of the transferred frequency can reach 9.0 × 10-16 at 1 s and 7.1 × 10-20 at 20 000 s.

A novel slow-light reflector composed of a double-ring resonator with a uniform fiber Bragg grating (FBG) incorporated is proposed in this paper, which can be used as a mirror of narrow-linewidth fiber laser, based on its enlarged group delay, and as a high-sensitivity sensor. Its characteristics are investigated theoretically and experimentally in this paper. Compared with the single fiber ring the amplitude of resonance is modulated, the effective mode spacing is expanded greatly, due to the Vernier effect, giving much better mode selectivity. Mode hopping is believed to be suppressed effectively in applications of narrow-linewidth lasers, and resolution will be enhanced in sensor applications.

Neural architecture search (NAS) has been widely studied to design high-performance network architectures automatically. However, existing approaches require more search time and substantial resource consumption due to their intensive architecture evaluations. Moreover, recently developed NAS algorithms are noncompetitive when combining multiple competing and conflicting objectives, e.g., the test accuracy and the number of parameters. In this paper, a low-cost NAS (LoNAS) method is proposed to address these problems. First, a variable-architecture encoding strategy based on a novel Reg Block is designed to construct high accuracy network architectures with few parameters. Second, a training-free proxy based on the neural tangent kernel (NTK) is proposed to accelerate the search process efficiently. Finally, a three-stage evolutionary algorithm (EA) based on multiple-criteria environmental selection and a set of block-based mutation operators are designed to balance exploration and exploitation better. The experimental results show that LoNAS finds network architectures with competitive performance compared to the state-of-the-art architectures in test accuracy and the number of parameters. Moreover, LoNAS uses less search time and fewer computational resources, consuming only 0.02 GPU Days with one GPU on CIFAR-10 and CIFAR-100. Furthermore, the architectures found by LoNAS on CIFAR-10 and CIFAR-100 exhibit good transferability to ImageNet-16-120, with the test accuracy surpassing that of the state-of-the-art network architectures. [ABSTRACT FROM AUTHOR]