Bo Fan

In the multi-motor drive systems, the backlash nonlinearity and the external disturbances should be taken into their modeling, which is essential to design the safer and more efficient load tracking and motor synchronization control strategies. An extended state observer-based optimal control method for the multi-motor drive system is proposed in this paper. The average deviation coupling control structure is introduced to decouple the performance indexes because of the measurable states in the multi-motor drive system. The extended state and disturbance observers are adopted to estimate the unknown state and disturbance of the system. The tracking controller and synchronous controller are designed by the combination of the backward and optimal control methods. The stability of the system is analyzed by the Lyapunov functions. With the comparison of experiments, the results show that the proposed control method has the better performance. The optimal tracking controller can obtain the load position tracking, and the optimal synchronous controller achieves fast convergence to the average value for all motors.

In foggy environments, atmospheric scattering and illumination attenuation severely degrade unmanned aerial vehicle (UAV) imagery, leading to reduced contrast and blurred structures. These degradations suppress discriminative object features and consequently undermine the accuracy and robustness of UAV-based object detection in low-visibility conditions. To address these challenges, we propose MSFP-DSNet, a unified dehazing and detection framework specifically designed for UAV in foggy scenes. Specifically, we first design a dual pooling fusion feature enhancement module that leverages the complementary characteristics of average pooling and max pooling within a lightweight parallel structure. Next, a multi-dimensional haze feature perception module is developed to explicitly model haze-induced feature degradation. Furthermore, we propose an information fusion and reconstruction pyramid that combines Laplacian pyramid decomposition with a wavelet-based sampling strategy. Finally, RT-DETR serves as the detection backbone, and a joint optimization objective is introduced to balance dehazing quality and detection accuracy, allowing the two sub-tasks to mutually reinforce each other during training. Extensive experiments are conducted on the HazyDet, the self-constructed DIOR-haze, and VisDrone-haze datasets. Compared with the baseline, the medium-scale AP of MSFP-DSNet improves by 6.23%, 2.74%, and 5.21%, respectively. These results validate the effectiveness of the proposed framework and highlight its strong potential as a practical and efficient solution for robust UAV visual perception in low-visibility and adverse weather scenarios.

Gear wear, backlash, and parameter uncertainties in dual-motor drive systems cause dead-zone nonlinearity and model mismatches, leading to torque discontinuities and degraded tracking. To address these issues, a novel prescribed-performance adaptive synchronization control strategy is proposed. The dead-zone nonlinearity is handled by a continuous function approximation to simplify controller design, while an improved prescribed performance function is introduced to transform constrained dynamics into an unconstrained form, which guaranteeing transient performance. The backstepping framework, incorporating a novel differentiator, circumvents the differential explosion problem. Furthermore, adaptive laws are developed to online estimate uncertain transmission parameters, effectively compensating for bias torque disturbances caused by backlash. Lyapunov analysis proves the asymptotic stability of the closed-loop system. Experimental results demonstrate that the proposed method significantly enhances position tracking performance and suppresses torque oscillations.

In recent years, real-time object detection technologies have shown promising potential for application on resource-constrained platforms such as unmanned aerial vehicles. However, maintaining detection accuracy while satisfying real-time and lightweight constraints remains a significant challenge. To address the limitations of traditional detectors in multi-scale feature extraction, cross-scale feature fusion, and decoder-level information transmission, we propose a real-time object detection method, termed HEFNet-DETR. First, we propose a novel parallel intra-scale feature interaction (PIFI) encoder, which replaces the conventional stacked architecture with a dual-branch, single-layer encoding framework to efficiently learn global information within each hierarchical level. Second, we develop a lightweight cross-scale repeated selective connection (CRSC) module that integrates multi-scale features with extremely low computational overhead, substantially enhancing the model's capacity to represent objects of varying scales. Finally, we design a selective hybrid query decoder (SHQD) that employs a dynamic cross-layer query selection mechanism, effectively avoiding computational redundancy and error accumulation caused by traditional layer-by-layer query transmission, thereby enhancing overall decoder performance. Experiments on the VisDrone2019 and Pascal VOC 2012 datasets demonstrate that, compared with the RT-DETR baseline, HEFNet-DETR reduces GFLOPs by 22.58%. On VisDrone2019, it achieves a 27.88% improvement in inference speed and a 4.30% improvement in mAP50; on Pascal VOC 2012, it achieves a 26.51% improvement in inference speed and a 2.46% improvement in mAP50, thereby confirming the efficiency and robustness of the proposed method for real-time object detection.

The conventional deadbeat predictive current control method for permanent magnet synchronous motor depends on the accuracy of motor parameters, which leads to large current deviation and fluctuation. An improved deadbeat predictive current control method based on the PID search algorithm (PSA) is proposed. With analysis on the mathematical model of the permanent magnet synchronous motor, a current prediction algorithm is introduced to eliminate the one-beat delay in the current command calculation. The weighting coefficients are combined to reduce the current fluctuation and static error. The PSA is utilized to accurately identify the resistance, magnetic chain and inductance parameters. It is to eliminate the current static error caused by parameter mismatch and improve the system robustness. Experiments verify the effectiveness of the proposed method under rated speed and load conditions. The performance of the proposed deadbeat predictive current control method with improved weighting factors is better than that of the conventional method.

Maneuvering extended object tracking has garnered significant attention owing to the continuous advancements in the resolution capabilities of modern high-precision radar sensors. The efficacy of tracking algorithms for such objects is heavily contingent upon the design of the model set. However, existing methodologies for model set design often yield suboptimal performance when confronted with highly maneuvering extended objects. The expected model augmentation (EMA) algorithm offers a data-driven mechanism for updating the model set in real time. Despite its advantages, the EMA algorithm is constrained by the fixed parameters of its basic models and static transition probabilities between models, thereby limiting its adaptability to extended objects exhibiting complex and dynamic maneuvering behaviors. To address these limitations, this paper proposes a modified variable structure multiple model (VSMM) framework for maneuvering extended object tracking, referred to as the adaptive grid expected model augmentation based on the golden section (GSAG-EMA) algorithm. The approach adaptively adjusts both the model structure and parameters in a grid-based format to accommodate the varying maneuvering patterns. It incorporates both local and global weighting schemes, with two models within the grid based on the golden section. Furthermore, the transition probability matrix is dynamically updated following specific rules, and the execution strategy for each module is determined according to the filtering results. Simulation results under both weak and strong maneuvering scenarios demonstrate that the proposed GSAG-EMA algorithm consistently outperforms the IMM-based, EMA, and AG-BMA algorithms in terms of root mean square error (RMSE) and Hausdorff distance, thereby substantiating its superior tracking performance.

The closed-loop control methods of bridge crane systems adopt the real-time signal feedback to improve the performance of the systems. The energy-based closed-loop control method of bridge crane systems can be used to design the controller by constructing the energy function of the system, which avoids the direct analysis of the complex motion state of the systems and has the clear physical significance. However, the conventional energy-based control methods have some problems, including slow response, poor performance of eliminating pendulum, and less parameters reflected by the control law. A novel load energy coupling-based underactuated control method for bridge cranes is proposed. The relationship between the energy of the bridge crane system and the energy of the load system is analyzed with the dynamic model of the two-dimensional bridge crane. The coupling function based on the load displacement and swing angle is constructed. The driving force of the system is obtained by Lyapunov method, and the control law is designed. The boundedness and convergence of the closed-loop control system are explained by the principle of LaSalle invariance. With the comparison experiments and results analysis, it is shown that the proposed method can obtain the better performance of eliminating swing and realize the accurate positioning of the trolley. It also reflects that this proposed method effectively reduces the dependence on model parameters and simplifies the control law.

The operational performances of the conventional controlled multi-motor drive systems are affected by the parameters of the motor, sudden load changes, and external disturbances. These factors may cause issues such as motor speed loss and torque imbalance, with severe cases potentially resulting in motor shaft fractures. In this paper, an enhanced deviation coupling-based synchronization and coordination control method is proposed for multi-motor rigid connection systems. The monomer and gear structure of the induction motor are modeled, respectively, and the transfer functions are given to analyze the factors that affect speed unsynchrony and torque uncoordination for the multi-motor systems. The inherent shortcomings of simple deviation coupling control are clarified, which makes it difficult to achieve the speed synchronization and torque balance of the multi-motor systems. An enhanced deviation coupling control method that simultaneously compensates for speed and torque is designed, and its stability is analyzed. Compared with the deviation coupling control method, the experimental results show that the proposed control method effectively reduces the speed synchronization errors and suppresses the torque output fluctuation. The control system has better performance of real-time and robustness.