We demonstrated a watt-level continuous-wave intracavity diamond Raman laser operating at 1634 nm in the eye-safe spectral region. By employing a dual Nd:YVO4 crystal pumping configuration in combination with a Z-shaped cavity design, the thermal lensing effect in the laser crystal was significantly suppressed. Consequently, the maximum injected pump power was increased from 40 W in a single-crystal scheme to 130 W, yielding 18.4 W of fundamental laser at 1342 nm. With the fundamental beam tightly focused to a 65 mu m waist radius on the diamond, continuous-wave Raman output at 1634 nm was achieved with a power of 1.46 W, representing the highest power reported. The device also exhibits good beam quality (M-2 < 1.3) and power stability (RMS < 2% over 1 hour). Numerical simulations further indicate that by increasing the local pump density within the Raman medium and optimizing the output coupling ratio, the output power can potentially be scaled to the 10 W level. (c) 2026 Optica Publishing Group. All rights, including for text and data mining (TDM), Artificial Intelligence (AI) training, and similar technologies, are reserved.

This paper investigates the $H_\infty$ control of microbial fuel cells (MFCs) within the Takagi–Sugeno (T–S) fuzzy framework. A nonlinear MFC model is first recast into a T–S fuzzy representation via the sector–nonlinearity approach, together with a control-oriented lumped disturbance channel to represent short-term external perturbations. On this basis, a non-matched affine membership-function-dependent PDC-type state-feedback controller is developed. Different from standard matched-PDC schemes, the controller employs affine membership functions that are not required to coincide with the plant memberships and explicitly incorporates membership-derivative-dependent feedback terms. A switched fuzzy Lyapunov function aligned with the controller, together with vertex-decomposition bounds for $\dot{h}$ (and $\dot{e}$, when present), leads to convex LMI stability and $H_\infty$ performance conditions that can be solved directly by off-the-shelf SDP solvers. A numerical example together with closed-loop simulations on the MFC model demonstrate a larger feasibility region and reduced conservatism than representative fuzzy designs. Moreover, relative to backstepping, sliding mode, fuzzy sliding mode, and classical $H_\infty$ controllers, the proposed method achieves faster convergence, shorter settling time, reduced residual regulation error under the considered disturbance scenario, and stronger disturbance rejection, yielding improved voltage regulation and practical applicability for bioelectrochemical energy systems.

To effectively deal with the problem of high noise and low signal-to-noise ratio (SNR) in the ultrasonic echo signal caused by mud pairs when the ultrasonic well logging instrument operates in the underground complex environment, this article proposes a wavelet threshold denoising method based on Newton-Raphson-based optimizer (NRBO)-improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN). This method employs the NRBO algorithm to optimize the parameters of ICEEMDAN, targeting the optimal combination of white noise amplitude weights (Nstd) and the number of noise additions (NE). Then with the help of correlation coefficient method to filter out the effective components from the intrinsic mode functions (IMFs) obtained from the decomposition, the improved wavelet threshold function is used to suppress the noise of the signal components, and finally, the denoised components are reconstructed to constitute the denoised signal. The results indicate that the proposed method demonstrates superior performance over conventional signal denoising techniques. Compared with the ICEEMDAN algorithm, it achieves a 23.45% improvement in SNR and a 38.27% reduction in root-mean-square error (RMSE). This approach effectively enhances signal clarity, thereby substantially improving the reliability and measurement accuracy in ultrasonic logging applications.

Edge-enhanced imaging and bright-field imaging reveal different morphological characteristics of an object. Hence, a system capable of realizing these modalities simultaneously is vitally essential for diverse applications. Here, we propose and demonstrate a method to combine deep learning (DL) with Fresnel incoherent correlation holography (FINCH) to achieve edge-enhanced and bright-field 3D imaging synchronously using only a single hologram. An integrated ResNet and U-net DL model is designed to predict the complex holograms with the spiral-FINCH and dual-lens FINCH from one input hologram, thereby obtaining the edge-enhanced and bright-field images at the three spatial dimensions through Fresnel propagation. Imaging experiments with different objects verify its capability to simultaneously perform multifunctional 3D imaging. This multifunctional 3D imaging system holds promising applications in biomedical imaging and defect detection, offering a novel tool for enhanced visualization and analysis.

In this paper, the problem of non-fragile finite-time contractive control of a quadrotor unmanned aerial vehicle with controller perturbations is studied. Based on the Lyapunov functional approach, a sufficient condition is derived for the existence of the desired controller ensuring finite-time contractive stability of the closed-loop system and reducing sensitivity to controller gain perturbations. By designing a time-varying proportional coefficient that works in conjunction with the attitude controller, the approach compensates for the limitation of the Lyapunov function method, which can only design controllers for systems with known matrices, enabling adaptation to an attitude subsystem with parameter uncertainties. Additionally, using the cone complementary linearization technique, the existence conditions for the controller are transformed into a linear matrix inequality form, facilitating the numerical solution of the controller gains. Finally, the numerical results show that the adopted method achieves better convergence performance compared to multiple traditional control methods.

Carbon dots (CDs) are a class of highly efficient, environmentally friendly, wavelength-diverse luminescent materials, exhibiting concentration-dependent emission properties. Understanding the mechanism of concentration-induced aggregation effects on their excited-state emission characteristics is crucial for developing high-performance laser devices based on CDs. In this study, we systematically investigate the spontaneous and stimulated emission actions of CDs with varying degrees of concentration-induced aggregation. Our findings reveal that the degree of aggregation does not inherently alter the type of charge carrier recombination, which remains dominated by excitonic recombination. However, with increasing aggregation, exciton-exciton annihilation becomes more prominent, reducing both total radiative efficiency and radiative decay rates. This ultimately results in an increased threshold for the stimulated emission. The findings in this work may advance the understanding of the excited-state emission mechanisms in CDs and provide valuable guidance for the design of CD-based laser materials and device architectures.

Microbial fuel cells (MFCs) play a vital role in water quality monitoring, where stable power generation is essential for ensuring the accuracy of water-quality detection. However, the complex reactions occurring in MFCs make it challenging to maintain a stable output voltage under uncontrolled conditions. Thus, a fractional-order PID (FOPID) controller is proposed. The parameters of this controller are typically determined using the particle swarm optimization (PSO) algorithm. To address the limitations of traditional PSO algorithms, such as low precision and slow convergence, an improved PSO algorithm integrating chaotic mechanisms, reverse learning, golden sine algorithm, and elite Gaussian mutation is proposed. Simulation results demonstrate that faster and more accurate convergence of the improved PSO. The proposed FOPID controller achieves a setting time of 8.2965 s, outperforming others with times of 88.8889 s, 39.0680 s, and so on. The FOPID controller offers advanced technical support for the application of MFCs in water-quality monitoring.

We demonstrated a high repetition rate, narrow spectral linewidth, and multi-watt actively Q-switched diamond Raman laser (DRL) at 1634 nm. A strategy of dual Nd:YVO4 crystal and increasing the Q-switch duty cycle was proposed to improve the population inversion and alleviate the thermal effects in the crystal simultaneously, leading to a 1342 nm fundamental laser with both a high output power (11 W) and repetition rate (50 kHz). By further designing a short Raman cavity and optimizing the duty cycle, a 1634 nm ns-pulsed DRL with a repetition rate >= 120 kHz and an output power of 5.2 W was achieved. The DRL also exhibited an excellent single longitudinal mode characteristic with a spectral linewidth of similar to 28 MHz at operational power up to 2 W and a beam quality with M-2 < 1.4.

The collected data of a pumping unit contain environmental noise, which significantly reduces the precision of fault diagnosis. The previous fault detection approach depends on manual feature extraction, which is time-consuming and laborious, and it cannot cope with high-noise conditions. Therefore, we propose a dual-branch time-frequency fusion deep learning model for fault diagnosis of the pumping unit. One branch extracts time-domain information, while the other branch extracts frequency-domain information by employing the fast Fourier transform. The branch information of these two branches is concatenated, and the gate-controlled channel transfer unit module automatically learns the competitive and cooperative relationships between each branch, making the key features more prominent in information fusion. Consequently, an accurate fault diagnosis of the pumping unit can be achieved under high-noise conditions. The results demonstrate that the proposed model outperforms the traditional schemes in terms of noise, with different signal-to-noise ratios.

With advancements in computer vision, computed tomography (CT) has been employed to aid clinicians in clinical diagnosis, thereby enhancing diagnostic efficiency. However, during the medical imaging process, medical images often suffer from issues such as blurring and complex noise as a result of system and equipment limitations. To address these challenges, we propose a novel image enhancement method integrating improved wavelet thresholding with total variation model denoising. Initially, the image is de-composed into high- and low-frequency sub-bands using wavelet decomposition. Subsequently, improved wavelet thresholding is employed to denoise the high-frequency sub-bands, which contain detail and texture information, whereas the total variation model is applied to denoise the low-frequency sub-bands containing the overall structure and rough outline information of an image. Finally, reconstruction is performed using an inverse wavelet transformation. Experimental results demonstrate that the proposed algorithm not only effectively suppresses complex noise in images and enhances the contrast of clinical pulmonary CT images but also preserves the natural appearance of images and enhances texture details and edge features. The proposed method exhibits superior performance compared with existing CT enhancement methods, achieving enhanced visual perception.

The target following problem of mobile robot could be split into two dimensions: direction control and distance control. Therefore, this paper analyzes two aspects separately. The steering and forward and backward movement of the robot are controlled by dual fuzzy controller. First of all, this topic designs and realizes the steering function of mobile robot when following. Based on the existing autonomous robot, the Angle is changed by using the image information of the target, and the corresponding fuzzy rules are designed to realize the direction control to follow the target. Secondly, it also realizes the function of the robot to follow the objective forward and backward, and obtains the target distance through the sensor of the robot, and keeps the two at a certain distance through fuzzy control, so as to realize the range following of the robot. The designed controller is implemented on the Webots simulation platform, and contrasted with a PID regulator, the results are analyzed, and the effectiveness of the algorithm is proved.

A trajectory tracking control method of underactuated unmanned surface vessels (USV) based on backstepping control method is proposed. Aiming at the underactuated characteristics of USV, an effective trajectory tracking controller is designed by using its kinematics model, introducing virtual control quantity and combining Lyapunov function. The asymptotic stability of the controller is proved by Lyapunov direct method, which ensures the stability of the system. The simulation results show that the trajectory tracking performance of the controller in the simulation environment is better than that of the traditional PID controller, especially in the curve path, it has better dynamic response ability and tracking accuracy. However, the test in the actual environment still needs to further verify the reliability and stability of the controller.

The convolutional neural network has significantly enhanced the efficacy of medical image segmentation. However, challenges persist in the deep learning-based method for medical image segmentation, necessitating the resolution of the following issues: (1) Medical images, characterized by a vast spatial scale and complex structure, pose difficulties in accurate edge information extraction; (2) In the decoding process, the assumption of equal importance among different channels contradicts the reality of their varying significance. This study addresses challenges observed in earlier medical image segmentation networks, particularly focusing on the precise extraction of edge information and the inadequate consideration of inter-channel importance during decoding. To address these challenges, we introduce ResTrans-Unet (residual transformer medical image segmentation network), an automatic segmentation model based on Residual-aware transformer. The Transformer is enhanced through the incorporation of ResMLP, resulting in enhanced edge information capture in images and improved network convergence speed. Additionally, Squeeze-and-Excitation Networks, which emphasize channel relationships, are integrated into the decoder to precisely highlight important features and suppress irrelevant ones. Experimental validations on two public datasets were carried out to assess the proposed model, comparing its performance with that of advanced models. The experimental results unequivocally demonstrate the superior performance of ResTrans-Unet in medical image segmentation tasks.

The 1.4-1.8 mu m eye-safe lasers have been widely used in the fields of laser medicine and laser detection and ranging. The diamond Raman lasers are capable of delivering excellent characteristics, such as good beam quality concomitantly with high output power. The intra-cavity diamond Raman lasers have the advantages of compactness and low Raman thresholds compared to the external-cavity Raman lasers. However, to date, the intra-cavity diamond cascaded Raman lasers in the spectral region of the eye-safe laser have an output power of only a few hundred milliwatts. A 1485 nm Nd:YVO4/diamond intra-cavity cascaded Raman laser is reported in this paper. The mode matching and stability of the cavity were optimally designed by a V-shaped folded cavity, which yielded an average output power of up to 2.2 W at a pulse repetition frequency of 50 kHz with a diode to second-Stokes conversion efficiency of 8.1%. Meanwhile, the pulse width of the second-Stokes laser was drastically reduced from 60 ns of the fundamental laser to 1.1 ns, which resulted in a high peak power of 40 kW. The device also exhibited single longitudinal mode with a narrow spectral width of < 0.02 nm.

In many production fields such as chemical, food, fertilizers, cement, etc., products are usually shipped in bags and stacked on pallets in warehouses for sale and transportation. Currently, long-distance transportation mainly relies on semitrailer trucks or box trucks, while the loading and unloading process of bagged materials relies heavily on manpower and material resources, significantly reducing the overall efficiency of product logistics transportation. This paper proposes a 3D object localization method based on EPNP and dual-view images for addressing the stacking issues of goods in industries such as factory production and logistics transportation. The method employs two cameras and a gantry to form a goods stacking positioning detection system. The Zhang Zhengyou chessboard calibration method is utilized to calibrate the cameras and determine their intrinsic parameters. The three-dimensional coordinates of the goods are calculated through methods such as the EPNP algorithm, perspective projection principles, and coordinate transformations. Experimental results have shown that this method can accurately measure the position of goods under the camera, making it convenient for palletizing machines to accurately stack and improve palletizing efficiency.

With the rapid development of internet technology, security protection of information has become more and more prominent, especially information encryption. Considering the great advantages of chaotic encryption, we propose a 2D-lag complex logistic map with complex parameters (2D-LCLMCP) and corresponding encryption schemes. Firstly, we present the model of the 2D-LCLMCP and analyze its chaotic properties and system stability through fixed points, Lyapunov exponent, bifurcation diagram, phase diagram, etc. Secondly, a block cipher algorithm based on the 2D-LCLMCP is proposed, the plaintext data is preprocessed using a pseudorandom sequence generated by the 2D-LCLMCP. Based on the generalized Feistel cipher structure, a round function F is constructed using dynamic S-box and DNA encoding rules as the core of the block cipher algorithm. The generalized Feistel cipher structure consists of two F functions, four XOR operations, and one permutation operation per round. The symmetric dynamic round keys that change with the plaintext are generated by the 2D-LCLMCP. Finally, experimental simulation and performance analysis tests are conducted. The results show that the block cipher algorithm has low complexit, good diffusion and a large key space. When the block length is 64 bits, only six rounds of encryption are required to provide sufficient security and robustness against cryptographic attacks.

We proposed a three-dimensional (3D) ranging system based on Fresnel incoherent correlation holography (FINCH). Distinct from the displacement measurement based on coherent digital holography (DH), our system simultaneously achieves a 3D range measurement using incoherent illumination. The observation range is obtained by the holographic reconstruction, while the in-plane range is determined using the two-dimensional digital imaging correlation (2D-DIC) technique. Experimental results on the resolution target demonstrate precise 3D ranging determination and improved measurement accuracy.