To address the problems that most convolutional neural network-based image defogging algorithm models capture incomplete global feature information and incomplete defogging, this paper proposes an end-to-end convolutional neural network and vision transformer hybrid image defogging algorithm. First, the shallow features of the haze image were extracted by a preprocessing module. Then, a symmetric network structure including a convolutional neural network (CNN) branch and a vision transformer branch was used to capture the local features and global features of the haze image, respectively. The mixed features were fused using convolutional layers to cover the global representation while retaining the local features. Finally, the features obtained by the encoder and decoder were fused to obtain richer feature information. The experimental results show that the proposed defogging algorithm achieved better defogging results in both the uniform and non-uniform haze datasets, solves the problems of dark and distorted colors after image defogging, and the recovered images are more natural for detail processing.

Surface defect detection of strip steel is an important guarantee for improving the production quality of strip steel. However, due to the diverse types, scales, and texture structures of surface defects on strip steel, as well as the irregular distribution of defects, it is difficult to achieve rapid and accurate detection of strip steel surface defects with existing methods. This article proposes a real-time and high-precision surface defect detection algorithm for strip steel based on YOLOv7. Firstly, Partial Conv is used to replace the conventional convolution blocks of the backbone network to reduce the size of the network model and improve the speed of detection; Secondly, The CA attention mechanism module is added to the ELAN module to enhance the ability of the network to extract detect features and improve the effectiveness of detect detection in complex environments; Finally, The SPD convolution module is introduced at the output end to improve the detection performance of small targets with surface defects on steel. The experimental results on the NEU-DET dataset indicate that the mean average accuracy (mAP@IoU = 0.5) is 80.4%, which is 4.0% higher than the baseline network. The number of parameters is reduced by 8.9%, and the computational load is reduced by 21.9% (GFLOPs). The detection speed reaches 90.9 FPS, which can well meet the requirements of real-time detection.

Continuous robots have attracted more and more attention from the robotics community due to their high degree of flexibility and pliability, and have shown great potential for application in a variety of fields. With the continuous progress of material science, control technology, and artificial intelligence, the performance and application range of soft robotics have been further expanded, in which the cable drive has the advantages of large workspace, high flexibility, etc. The cable-driven soft robotic arm serves as an ultra-redundant robot that can operate in cramped and confined environments. In this paper, a cable-driven soft robot based on soft continuums and a cross gimbal is presented. The kinematics of the cable-driven soft robot is modeled and the mapping relations of the kinematics are solved by the D–H method and piecewise constant curvature, and the relations between the cable length, joint angle, and pose are further derived. Finally, the motion space of the cable-driven soft robot in the three-dimensional coordinate system is obtained by MATLAB2021b, and the single-segment soft body is simulated and analyzed using ADAMS to compare the theoretical data with the actual data and verify the reliability of this structure and method.

The nonlinear characteristics of the pneumatic servo system are the main factors limiting its control accuracy. A new mathematical model of the nonlinear system of the valve control cylinder is proposed in order to improve the control accuracy of the pneumatic servo system. Firstly, the mass flow equation of the gas flowing through each port is established by analyzing the physical structure of the proportional directional control valve. Then, the dynamic equation of the system is set up by applying the Stribeck friction model for the friction model of the valve control cylinder and building a pneumatic circuit experiment to identify the friction model parameters. Finally, the correctness of the mathematical model is verified by the inflation and deflation experiment of the fixed capacitive chamber and the servo controls experiment based on PID position. The Simulink simulation of the mathematical model better reflects the characteristics of the pneumatic position servo system.

This paper presents the design of a symmetric variable stiffness joint that employs worm gear and sliding helical transmissions to adjust the effective length of the leaf springs. Firstly, the design concept and working principle of the variable stiffness joint are presented, along with two different assembly methods. Secondly, the stiffness equations and characteristics of the variable stiffness joint are then derived and analyzed. Next, the dynamics of the variable stiffness joint are modeled and simulated visually using Simulink. Finally, a prototype of the variable stiffness joint is constructed and its stiffness characteristics are experimentally verified. The experimental results demonstrate that both assembly methods are capable of adjusting the stiffness and position of the joint within a certain range. This study contributes to the understanding and development of symmetric variable stiffness joints by presenting a comprehensive design, analysis, simulation, and experimental evaluation. The proposed joint has potential applications in various fields that require adaptability, adjustability, and safety.

As the core technology for realizing autonomous navigation of unmanned ground vehicles, the path planning algorithm directly determines the reliability and stability of navigation tasks in complex dynamic environments. With the expanding range of application scenarios, traditional path planning approaches have become increasingly inadequate in terms of real-time performance, dynamic obstacle avoidance, and multi-objective optimization. The recent rise in AI-based methods has provided new opportunities for this field. This paper systematically analyzes the latest research progress in this area. By reviewing and analyzing the highly recognized literature in recent years, we classify mainstream path planning and related algorithms into six types: graph-search-based, sampling-based, local optimization-based, meta-heuristic optimization, AI-based, and optimal control methods. The core improvement trends, advantages, and inherent limitations of each algorithm type are deeply analyzed. Through bibliometric analysis, we identify major gaps in current research, including over-reliance on simulation methods, overly restrictive environmental assumptions, and insufficient handling of multiple objectives. Finally, we point out the critical gap between simulation environments and real-world deployment and advocate the use of hybrid algorithms to address the deficiencies of single algorithms, along with effective validation in real environments. This direction is crucial for promoting the broader practical application of unmanned ground vehicle technology.