Recent years have seen significant advancements in category-level object pose estimation, largely driven by multimodal (RGB-D) approaches. Despite their success, depth-only methods remain widely adopted in practical applications due to their superior computational efficiency and ease of deployment. However, these methods typically suffer from a noticeable performance gap compared to multimodal methods. To bridge this gap, we propose a novel framework, Cross-Modal Uncertainty Distillation for Pose Estimation (CMUD-Pose), which transfers discriminative knowledge from an RGB-D teacher to a depth-only student network. Furthermore, to mitigate overfitting induced by the modality gap, we propose Cross-Modal Uncertainty Distillation (CMUD), which utilizes a learned uncertainty-aware weighting mechanism for adaptively assigning importance to training samples. By incorporating uncertainty into the distillation process, CMUD allows the student model to focus selectively on reliable and transferable cross-modal features. Extensive experiments on the REAL275 and CAMERA25 benchmarks show that our method significantly improves the performance of depth-only pose estimation models.

Estimating the pose of objects from images is a crucial task of 3D scene understanding, and recent approaches have shown promising results on very large benchmarks. However, these methods experience a significant performance drop when dealing with unseen objects. We believe that it results from the limited generalizability of image features. To address this problem, we have an in-depth analysis on the features of diffusion models, e.g. Stable Diffusion, which hold substantial potential for modeling unseen objects. Based on this analysis, we then innovatively introduce these diffusion features for object pose estimation. To achieve this, we propose three distinct architectures that can effectively capture and aggregate diffusion features of different granularity, greatly improving the generalizability of object pose estimation. Our approach outperforms the state-of-the-art methods by a considerable margin on three popular benchmark datasets, LM, O-LM, and T-LESS. In particular, our method achieves higher accuracy than the previous best arts on unseen objects: 98.2% vs. 93.5% on Unseen LM, 85.9% vs. 76.3% on Unseen O-LM, showing the strong generalizability of our method. Our code is released at https://github.com/Tianfu18/diff-feats-pose.

Bridge inspections are essential for maintaining key infrastructure and preventing structural and functional failures. Nevertheless, traditional manual inspection techniques are plagued by laboriousness, high risk, and low efficiency. Although numerous automation inspection methods have been studied, inspection performance remains challenging. The main difficulties are redundant mechanisms, complex control, high energy consumption, and limited autonomy and safety. To address these problems, we are developing a small, lightweight, electrically-driven robotic manipulator for bridge inspection named the BIRM. Here, we propose a kinetic-energy-optimal and safety-guaranteed trajectory planning for BIRM. Compared with existing methods, it simultaneously addresses energy consumption and safety. The approach formulates a quadratic programming (QP) problem by considering the robot’s kinetic energy as the objective function, and the augmented Lagrange multiplier (ALM) is applied to find the solution of the QP. The proposed method completely satisfies the joint position, velocity, and acceleration limits at the speed level while considering collision avoidance. In this paper, the collision detection strategy can achieve low-complexity computation through several structural parameters of the bridge, thereby quickly adapting to environmental changes. Through simulation experiments, we validate the effectiveness and superiority of the proposed method. Through physical experiments, we demonstrate the sustainability and safety of bridge inspections in the field.

This paper presents a wall-climbing robot for reconnaissance in anti-hijacking application. A novel biped-wheel hybrid locomotion mechanism is proposed, which is composed of a planetary gear train, a vacuum adhesion module and a negative pressure adhesion module. The bipedal, wheeled and hybrid locomotion modes are analyzed respectively. A prototype of the wall-climbing robot with compact size and low power consumption has been developed and a lot of performance tests have been conducted. The experimental results demonstrate that the wall-climbing robot has such characteristics as fast moving speed, excellent surface adaptability and obstacle negotiation capability.

The processing of blood vessels is an indispensable part in complicated surgeries of livers and hearts as the development of medical image technologies, which requires an automatic segmentation system over CT images of organs. However, the vascular pattern of livers in CT images suffers from low contrast to background so that the existing segmentation technologies are not able to extract the blood vessels completely. In the paper, we propose a new algorithm to extract the blood vessels of livers based on the adaptive multi-scale segmentation. First, we prove that the background histogram of normal scale blood vessels obeys the Gaussian distribution in CT images, and obtain the vascular distribution function from the vascular signal segmented from the background with a local optimal threshold. Second, Hessian matrix is employed to enhance the thin blood vessels before the extraction, and a complete and clear segmentation system for blood vessels is constructed by combining the major and thin blood vessels via filtering. Experimental results show the effectiveness of the proposed method, which is able to extract more complete blood vessels for 3D system, and assist the clinical liver surgeries efficiently.

High-generation FPD (Flat Panel Display) glass substrates need to be carried by large clean transfer robot. Due to the space of the deposited glass box is limited, the distance between the two substrates is very small, this requires the sag amount of the robot fork must be controlled within a certain range. The traditional design is completed only in robot manufacturing stage by adding a shim to compensate for this distortion, but this method can only make the fork be compensated at a determined location, deformation at other locations may not be considered. A new solution to the dynamic compensation for the robot fork is proposed. By establishing the robot FEA model and mechanism accuracy model respectively, the synthetic dynamic error is then obtained. One can just change the mechanism accuracy through rational allocation of design tolerances of the robot links to minimize the dynamic error.

The developments in mobile robotics have increasingly played an important role in the inspection and maintenance work of Power Transmission Lines (PTLs). This paper presents the research and development of the inspection robots for 500kV PTLs in Shenyang Institute of Automation, Chinese Academy of Sciences (SIACAS). An overview of the research work and the development of an inspection robot AApe-B, are introduced, respectively. The AApe-B can run on the Overhead Ground Wires (OGWs) and navigate different types of obstacles in remote and locally autonomous control mode. A novel dual-wheel-arm hybrid mechanism with the excellent locomotion and obstacle-navigation performance is designed, and obstacle navigation process is analyzed. An embedded control system which can withstand the strong electromagnetic interference has been developed. A lot of field experiments have been carried out, and the experimental results have shown that the inspection robot can reliably work during the 500kV Extra High Voltage (EHV) electromagnetic environment and possess the primary ability to implement the inspection task of the 500kV PTLs.

The purpose of inspection tasks for power transmission lines is to check running state and find damages of extra-high voltage (EHV) power transmission lines equipment. Obviously, inspection tasks are very important for normal operation of power transmission lines. So far, there are two methods for checking power transmission lines. One method is that power transmission line equipment has been checked manually by workers with a telescope on the ground. Another method is that power transmission line equipment has been inspected by helicopter with the checking devices which are thermal infrared imager, visible light camera and so on. After inspection robot configuration is described, the centroid adjustment principle of inspection robot is introduced. Then the control strategy of inspection robot is investigated. From the experiments, we find that the centroid adjustment method proposed and the control strategy presented is feasible.

The paper analysis a kinematic model for a wheeled mobile robot (WMR) traversing uneven terrain. A new form of the kinematics for wheeled mobile robot is deduced, through analyzing Jacobian matrices of individual wheel and rearranging the variables. The performance and characters of the kinematic formulation are explained using physical conception. A new method is proposed to set up the kinematics formulation for wheeled mobile robots. After analyzing the actuation kinematics, simulation results are provided to validate the motion of wheeled mobile robot over a special terrain

This paper presents a multiple-model and Boolean logic reasoning (MMBLR) approach to detect and identify faults in the suction foot control of a climbing robot. For this control system, some fault models are easily given by kinematics equations. Moreover, the logic relations of the system states have been known in advance. Based on the combination of the multiple-model adaptive estimation (MMAE) algorithm and the Boolean logic reasoning, the MMBLR approach is properly fit for the fault detection and identification (FDI) application to the climbing robot. In the MMBLR architecture, the MMAE algorithm is used to reliably detect and identify the model-known faults. Then based on the robot's states and the results of the MMAE, other faults are detected and identified using the Boolean logic reasoning. Experimental results validated that the faults of the sensors and actuators in the suction foot control of the robot can be readily detected and identified by the MMBLR approach

The static measuring model and the deadweight compensation of a Stewart platform based force/torque sensor have been studied in this paper. Firstly, some preliminary kinematics of this sensor is presented. Coordinates attached to every links are established, and the transforming relations between link frames and other frames are derived. Then, the static measuring model and the deadweight compensation of the sensor are produced. In the modeling, forces on links are composed based on an analytic method, and the resultant forces and torques obtained can be used as the measuring model. The weight need to be compensated of the whole mechanism is divided into two parts: links weights and top platform weight. The compensation of the links weights is involved in static modeling. And the top platform weight is compensated, subsequently. Finally, some experiments are carried out, and the results verify the validity of the static measuring model and the deadweight compensation

Inspection robot with dual wheels encounters upslope and down slope, when it runs on the overhead ground wires (OGWs) which displays catenary shape. The tilt of the body of inspection robot, which is caused by the angle variation of catenary in different position, will occur during running procedure, and disenable to check vertically the transmission lines by camera. An adjustment method, which can keep the body of inspection robot in horizontal pose, is presented by the obliquitous sensor changing the length of the manipulators in this paper. In order to implement the pose adjustment, the dynamic model is established by Lagrange method. Finally, the simulation and experiment results show that the adjustment method proposed is feasible

This paper presents an obstacle-navigation control strategy for a mobile robot suspended on overhead ground wires of power transmission lines based on a novel movement mechanism. The kinematics of the mobile robot and the process of obstacle-navigation control are analyzed in detail. This paper presents to develop robot expert system using hybrid programming of C language integrated production system (CLIPS) and C language, and the robot expert system has powerful logical reasoning ability and data processing ability. The robot control modes have layered planning and direct control. This paper describes the robot expert system with hybrid architecture and analyzed decision-making of obstacle-navigation in layered planning mode in detail. The robot can learn online or offline, and when it encounter the same obstacles, it can navigate them quickly. Obstacle-navigation control is the difficult point of the type of wire-suspended robot and the expert system can instruct the mobile robot to navigate obstacles autonomously on overhead ground wires.