Mechanism parameters of bipedal robots are crucial for achieving efficient locomotion in complex environments. Inspired by the human energy-efficient walking style, this paper proposes a novel concept of full-range walking energy efficiency and explores the optimal linkage mechanism within certain ranges of step length and walking speed for bipedal robots. First, a bipedal model incorporating an upper body is established for dynamic analysis. Next, an optimal walking gait subject to walking constraints is solved by considering the full-range energy efficiency. Further, an optimal linkage mechanism is investigated, and the influence of dynamic parameters on energy efficiency is analyzed. Finally, the push-off impulse, minimum ground support force, and walking torque features are discussed. It shows that the full-range walking energy efficiency can be lowered by reducing the ratio of leg mass, concentrating mass at the hip joint, decreasing the length of the upper body, or increasing the center of mass of the leg. In addition, efficient walking motion can be achieved by designing the coordination of positive hip joint torque and push-off impulse at the ankle. This paper can be used to guide the mechanism parameter optimization and efficient walking gait design of bipedal robots.

Image classification and segmentation are important tasks in computer vision. ResNet and U-Net are representative networks for image classification and image segmentation, respectively. Although many scholars used to fuse these two networks, most integration focuses on image segmentation with U-Net, overlooking the capabilities of ResNet for image classification. In this paper, we propose a novel U-ResNet structure by combining U-Net’s convolution–deconvolution structure (UBlock) with ResNet’s residual structure (ResBlock) in a parallel manner. This novel parallel structure achieves rapid convergence and high accuracy in image classification and segmentation while also efficiently alleviating the vanishing gradient problem. Specifically, in the UBlock, the pixel-level features of both high- and low-resolution images are extracted and processed. In the ResBlock, a Selected Upsampling (SU) module was introduced to enhance performance on low-resolution datasets, and an improved Efficient Upsampling Convolutional Block (EUCB*) with a Channel Shuffle mechanism was added before the output of the ResBlock to enhance network convergence. Features from both the ResBlock and UBlock were merged for better decision making. This architecture outperformed the state-of-the-art (SOTA) models in both image classification and segmentation tasks on open-source and private datasets. Functions of individual modules were further verified via ablation studies. The superiority of the proposed U-ResNet displays strong feasibility and potential for advanced cross-paradigm tasks in computer vision.

Collaborative robots face barriers to widespread adoption due to the complexity of programming them to achieve human-like movement. Learning from demonstration (LfD) has emerged as a crucial solution, allowing robots to learn tasks directly from expert demonstrations, offering versatility and an intuitive programming approach. However, many existing LfD methods encounter issues such as convergence failure and lack of generalization ability. In this paper, we propose: (1) a generative adversarial network (GAN)-based model with multilayer perceptron (MLP) architecture, coupled with a novel loss function designed to mitigate convergence issues; (2) an affine transformation-based generalization method aimed at enhancing LfD tasks by improving their generalization performance; (3) a data preprocessing method tailored to facilitate deployment on robotics platforms. We conduct experiments on a UR5 robotic platform tasked with handwritten digit recognition. Our results demonstrate that our proposed method significantly accelerates generation speed, achieving a remarkable processing time of 23 ms, which is five times faster than movement primitives (MPs), while preserving key features from demonstrations. This leads to outstanding convergence and generalization performance.

High-density 3D warehousing is the future cornerstone of modern logistics storage. It aims to efficiently manage and store various types of goods by utilizing 3D space and automation technology. The accurate and fast docking algorithm of the four-way shuttles, as a core method of carrying the goods in high-density 3D warehousing, has become an increasingly important technical challenge. Inaccurate docking will result in a failure to change direction, while long docking time will affect the operation efficiency of the four-way shuttle. To overcome these obstacles, this paper presents an accurate and rapid docking algorithm for four-way shuttles. Firstly, the deceleration and jerk of a four-way shuttle in the braking stage are initialized according to the motion parameter set calculated by the docking motion model. Then, the four-way shuttle starts to move until the stop signal has been detected. If the system fails to detect the stop signal, the distance compensation is needed to ensure that the four-way shuttle can arrive at the stop point. Finally, the four-way shuttle stops immediately by way of amplifying the deceleration when reaching the stop point. The result shows that the accuracy of rapid docking is superior to that of direct docking. Compared with crawling docking, rapid docking reduces the braking time by 3.99s and stops at a speed lower than the crawling speed. The rapid docking algorithm not only enhances the throughput of high-density 3D warehousing but also improves the service life of four-way shuttles.

6D object pose estimation is crucial for applications such as robotic manipulation and augmented reality, yet it remains highly challenging when dealing with objects of significantly different aspect ratios or the drastic appearance variations of a single object caused by pose changes. Most existing methods focus on designing more complex backend fusion modules, while largely overlooking a fundamental problem at the feature extraction frontend: the geometric mismatch between the fixed, square receptive fields of standard convolutions and the varied projected morphologies of objects. This mismatch, along with noise in fused features and ambiguity in regression, limits the performance ceiling of current methods. To this end, this paper proposes a novel Pose-Perceptive Convolution (PPC) and constructs a new Pose-Perceptive Fusion Network (PPF-Net). Its core component, the Pose-Perceptive Convolution, fundamentally resolves the aforementioned geometric mismatch by dynamically adapting the shape and sampling density of its receptive field. Experiments on four benchmarks show that PPF-Net improves the VSD score by 19.4% over FFB6D on MP6D, and achieves 96.7% ADD-S on YCB-Video, approaching state-of-the-art accuracy. Crucially, these gains are realized with minimal computational overhead, avoiding the heavy latency of backend-intensive approaches. This validates that frontend feature extraction is an efficient strategy for robust 6D pose estimation.