Safety is an essential requirement as well as a major bottleneck for legged robots in the real world. Particularly for learning-based methods, their trial-and-error nature and unexplainable policy have raised widespread concerns. Existing methods usually treat this challenge as a trade-off between safety assurance and task performance. One reason for this drawback stems from the inaccurate inference for the robot’s safety. In this paper, we re-examine the segmentation of the robot’s state space in terms of safety. According to the current state and the prediction of the state transition trajectory, the states of legged robots are classified into safe, recoverable, unsafe, and failure, and a safety verification method is introduced to online infer the robot’s safety. Then, task, recovery, and fall protection policies are trained to ensure the robot’s safety in different states, forming a safety supervision framework independently from the learning algorithm. To validate the proposed method and framework, experiment results are conducted both in the simulation and on the real-world robot, indicating improvements in terms of safety and efficiency.

This article addresses the localization problem in robotic autonomous luggage trolley collection at airports and provides a systematic evaluation of different methods to solve it. The robotic autonomous luggage trolley collection is a complex system that involves object detection, localization, motion planning and control, manipulation, etc. Among these components, effective localization is essential for the robot to employ subsequent motion planning and end-effector manipulation because it can provide a correct goal position. This article explores four popular and representative localization methods for object localization in luggage trolley collection: radio frequency identification (RFID), Keypoints, ultrawideband (UWB), and Reflectors. A qualitative evaluation framework is constructed to assess performance, encompassing Localization Accuracy, Mobile Power Supplies, Coverage Area, Cost, and Scalability. Furthermore, a series of quantitative experiments concerning Localization Accuracy and Success Rate have been conducted on a real-world robotic autonomous luggage trolley collection system. The performance of various localization methods is further analyzed based on experimental results, indicating that the Keypoints method is optimally suited for indoor environments to facilitate luggage trolley collection. Significantly, these experiment results provide a valuable reference point, extending the application of indoor localization methods across diverse scenarios. A website about this work is available at https://sites.google.com/view/localization-evaluation/.

A method is proposed for identifying robot gravity and friction torques based on joint currents. The minimum gravity term parameters are obtained using the Modified Denavit–Hartenberg (MDH) parameters, and the dynamic equations are linearized. The robot’s friction torque is identified using the Stribeck friction model. Additionally, a zero-force drag algorithm is designed to address the issue of excessive start-up torque during dragging. A sinusoidal compensation algorithm is proposed to perform periodic friction compensation for each stationary joint, utilizing the identified maximum static friction torque. Experimental results show that when the robot operates at a uniform low speed, the theoretical current calculated based on the identified gravity and friction fits the actual current well, with a maximum root mean square error within 50 mA, confirming the accuracy of the identification results. The start-up torque compensation algorithm reduces the robot’s start-up torque by an average of $ 60.58\mathrm{\%}$, improving the compliance of the dragging process and demonstrating the effectiveness of the compensation algorithm.

Wheel-leg composite robots exhibit robust mobility and exceptional obstacle-crossing capabilities in complex environments. This paper proposes a novel transformable wheel-leg composite structure and presents the design of a wheel-leg composite obstacle-crossing robot, fundamentally configured as a two-wheeled quadruped. The research encompasses a comprehensive analysis of the robot’s overall mechanical structure, a detailed kinematic investigation of its body and obstacle-crossing gait planning, virtual prototype dynamics simulation, and field experimentation. Utilizing advanced modeling software, a 3D model of the robot was established. The kinematic characteristics of the robot in both wheeled and legged modes were thoroughly examined. Specifically, for the legged mode, the Denavit-Hartenberg coordinate system was established, and a detailed kinematic model was analyzed. The obstacle-crossing gait was planned based on the robot’s leg action mechanism. Furthermore, the Lagrangian method was employed to develop a mathematical model for the dynamics of the robot in both wheel-foot modes, allowing for a comprehensive force analysis. To validate the feasibility and rationality of the robot’s obstacle-crossing capabilities under various conditions, extensive simulations and prototype tests were conducted across diverse terrains. The results provide valuable insights and practical guidance for the structural design of wheel-leg composite obstacle-crossing robots, contributing to advancements in this promising field.

Cartographer is an algorithm that was open sourced by Google in 2016 and adapted to multiple sensors. To address issues of the original algorithm, such as the negative impact of outlier point cloud on the scan matching, and low accuracy of position fusion. This paper preprocesses the sensor data and presents HT-Carto, an improved hybrid point-cloud filtering system, and a tightly coupled LiDAR/IMU framework based on Cartographer’s front-end. The inertial measurement unit (IMU) provides initial values for the point cloud, and the IMU pre-integration combines the scan-matched pose to construct the factors, which are added as constraints to the factor graph. The result is used to update the current pose and work as odometer residuals at the back-end. The optimization of the selected strategy during point cloud preprocessing, PassThrough, and RadiusOutlierRemoval are combined to ensure quality. An actual vehicle is used in complex indoor environment to verify the stability and robustness of HT-Carto. Compared to the Cartographer, Karto, Hector, and GMapping, HT-Carto demonstrates better localization and mapping, it can obtain a more precise trajectory.

This paper introduces a distributed online learning coverage control algorithm based on sparse Gaussian process regression for addressing the problem of multi-robot area coverage and source localization in unknown environments. Considering the limitations of traditional Gaussian process regression in handling large datasets, this study employs multiple robots to explore the task area to gather environmental information and approximate the posterior distribution of the model using variational free energy methods, which serves as the input for the centroid Voronoi tessellation algorithm. Additionally, taking into consideration the localization errors, and the impact of obstacles, buffer factors and centroid Voronoi tessellation algorithms with separating hyperplanes are introduced for dynamic robot task area planning, ultimately achieving autonomous online decision-making and optimal coverage. Simulation results demonstrate that the proposed algorithm ensures the safety of multi-robot formations, exhibits higher iteration speed, and improves source localization accuracy, highlighting the effectiveness of model enhancements.

To meet the high-precision positioning requirements for hybrid machining units, this article presents a geometric error modeling and source error identification methodology for a serial–parallel hybrid kinematic machining unit (HKMU) with five axis. A minimal kinematic error modeling of the serial–parallel HKMU is established with screw-based method after elimination of redundant errors. A set of composite error indices is formulated to describe the terminal accuracy distribution characteristics in a quantitative manner. A modified projection method is proposed to determine the actual compensable and noncompensable source errors of the HKMU by identifying such transformable source errors. Based on this, the error compensation and comparison analysis are carried out on the exemplary HKMU to numerically verify the effectiveness of the proposed modified projection method. The geometric error evaluations reveal that the parallel module has a larger impacts on the terminal accuracy of the platform of the HKMU than the serial module. The error compensation results manifest that the modified projection method can find additional compensable source errors and significantly reduce the average and maximum values of geometric errors of the HKMU. Hence, the proposed methodology can be applied to improve the accuracy of kinematic calibration of the compensable source errors and can reduce the difficulty and workload of tolerance design for noncompensable source errors of such serial–parallel hybrid mechanism.

Motion primitives play an important role in motion planning for autonomous vehicles, as they effectively address the sampling challenges inherent in nonholonomic motion planning. Employing motion primitives (MPs) is a widely accepted approach in nonholonomic motion planning based on sampling. This study specifically addresses the problem of learning from human-driving data to create human-like trajectories from predefined start-to-end states, which then serve as MP within the sampling-based nonholonomic motion planning framework. In this paper, we propose a deep learning-based method for generating MP that capture human-driving trajectory data features. By processing human-driving trajectory data, we create a Motion Primitive dataset that uniformly covers typical urban driving scenarios. Based on this dataset, a vehicle model long short-term memory neural network model is constructed to learn the features of the human-driving trajectory data. Finally, a framework for the generation of MP for practical applications is given based on this neural network. Our experiments, which focus on the dataset, the MMP generation network, and the generation process, demonstrate that our method significantly improves the training efficacy of the MP generation network. Additionally, the MP generated by our method exhibit higher accuracy compared to traditional methods.

Target tracking technology is a key research area in the field of mobile robots, with wide applications in logistics, security, autonomous driving, and more. It generally involves two main components: target recognition and target following. However, the limited computational power of the mobile robot’s controller makes achieving high precision and fast target recognition and tracking a challenge. To address the challenges posed by limited computing power, this paper proposes a target-tracking control algorithm based on lightweight neural networks. First, a depthwise separable convolution-based backbone is introduced for feature extraction. Then, an efficient channel attention module is incorporated into the target recognition algorithm to minimize the impact of redundant features and emphasize important channels, thereby reducing model complexity and enhancing network efficiency. Finally, based on the data collected from visual and ultrasonic sensors, a model predictive control strategy is used to achieve target tracking. Validation of the proposed algorithm is conducted using a mobile robot equipped with Raspberry Pi 4B. Experimental results demonstrate that the proposed algorithm achieves rapid target tracking.

Descriptions of various subsets of $\mathbb{SO}(3)$ are encountered frequently in robotics, for example, in the context of specifying the orientation workspaces of manipulators. Often, the Cartesian concept of a cuboid is extended into the domain of Euler angles, notwithstanding the fact that the physical implications of this practice are not documented. Motivated by this lacuna in the existing literature, this article focuses on studying sets of rotations described by such cuboids by mapping them to the space of Rodrigues parameters, where a physically meaningful measure of distance from the origin is available and the spherical geometry is intrinsically pertinent. It is established that the planar faces of the said cuboid transform into hyperboloids of one sheet and hence, the cuboid itself maps into a solid of complicated non-convex shape. To quantify the extents of these solids, the largest spheres contained within them are computed analytically. It is expected that this study would help in the process of design and path planning of spatial robots, especially those of parallel architecture, due to a better and quantitative understanding of their orientation workspaces.

Connecting individual robots to form an inter-reconfigurable system with a flexible base size enhances the ability to access and cover areas for cleaning and maintenance tasks. Given that increased configuration complexity expands the search space dimension, an optimal routing solution ensuring efficiency is essential. In this paper, we present an inter-reconfigurable multi-robot system capable of adjusting the bases of its two units, along with an optimal path planning approach for confined spaces based on a modified informed rapidly-exploring random tree algorithm by a greedy set (RIRRT*). We validate the navigation of the proposed inter-reconfigurable platform using RIRRT* for four informed dimensional search spaces as a case study in both simulated and real-world environments. The proposed path planning method for the inter-reconfigurable system outperformed conventional strategies, achieving significant reduction in both execution time and energy utilization.

Soft robots show an advantage when conducting tasks in complex environments due to their enormous flexibility and adaptability. However, soft robots suffer interactions and nonlinear deformation when interacting with soft and fluid materials. The reason behind is the free boundary interactions, which refers to undetermined contact between soft materials, specifically containing nonlinear deformation in air and nonlinear interactions in fluid for soft robot simulation. Therefore, we propose a new approach using material point method (MPM), which can solve the free boundary interactions problem, to simulate soft robots under such environments. The proposed approach can autonomously predict the flexible and versatile behaviors of soft robots. Our approach entails incorporating automatic differentiation into the algorithm of MPM to simplify the computation and implement an efficient implicit time integration algorithm. We perform two groups of experiments with an ordinary pneumatic soft finger in different free boundary interactions. The results indicate that it is possible to simulate soft robots with nonlinear interactions and deformation, and such environmental effects on soft robots can be restored.

In this article, the performance analysis and multiobjective structure optimization of 4RRR parallel mechanism are carried out. Firstly, the 4RRR pure rotation parallel mechanism and its design route are introduced. Secondly, the Jacobian matrices in 2DoF pure rotation and 3DoF pure rotation modes are derived using the motion equations of the mechanism. Next, the singularity analysis, kinematic dexterity analysis, dynamic dexterity analysis, and stiffness analysis of the mechanism are carried out, respectively, and it is proved that there is no singularity in the mechanism in its workspace. Since the dexterity performance expression is a nonlinear piecewise function, the kinematic local comprehensive dexterity index and the dynamic local comprehensive dexterity index are proposed as the objects of analysis. Furthermore, the kinematic global comprehensive dexterity index, the dynamic global comprehensive dexterity index, and the global comprehensive stiffness index are proposed to carry out the multiobjective structural optimization. Finally, NSGA3 was used to complete the optimization, and the comprehensive optimal solution of the structure size was obtained.

The two main sources of difficulty for a group of mobile robots employing sensors to find a source are robot collisions and wireless ambient noise, such as light, sound, and other sounds. This paper introduces a novel approach to multi-robot system cooperation and collision avoidance: the new modified source-seeking control with noise cancelation technology. The robot team works together on an incline of a light source field; the team’s mobility is dependent upon following the upward gradient’s direction and forming a particular movement pattern. The proposed program also takes into account each robot’s size, speed limit, obstacles, and noise. The noise cancelation technique has been used to avoid the delay and false decisions to find the target point of the source. When the noise is canceled, all control inputs to the algorithm are accurate, and the feedback decision will be true. In this study, we use the MATLAB simulation tools to test the velocity, position, time delay, and performance of each robot in the used group of robots. The simulation and practical results of the robots in searching for a light source showed very satisfactory performance compared with the results in the literature.

In this paper, a novel special structure derived from the double-slider mechanism is presented and apply it to the design of a parallel mechanism, endowing the designed parallel mechanism with reconfigurable characteristics. First, a comprehensive analysis of the motion modes of the double-slider mechanism is carried out. By ingeniously varying the slider structure of the double-slider mechanism, a special structure capable of three distinct motion modes is obtained. This special structure is then integrated into the 3UPU parallel mechanism. As a result, the redesigned 3UPU parallel mechanism exhibits reconfigurability and can seamlessly switch among the three motion modes. Subsequently, the inverse kinematics, workspace, and singular pose of the parallel mechanism in these three modes are meticulously analyzed. Moreover, the Jacobian matrix is utilized to evaluate its flexibility and load-bearing performance. The analysis reveals that in different motion modes, all performance indicators of the mechanism are remarkable, indicating a promising application prospect. Finally, a prototype is fabricated using 3D printing technology to further validate the effectiveness of the proposed special structure. Additionally, its versatility is further explored and analyzed in-depth.

Detecting cracks in underwater dams is crucial for ensuring the quality and safety of the dam. However, underwater dam cracks are easily obscured by aquatic plants. Traditional single-view visual inspection methods cannot effectively extract the feature information of the occluded cracks, while multi-view crack images can extract the occluded target features through feature fusion. At the same time, underwater turbulence leads to nonuniform diffusion of suspended sediments, resulting in nonuniform flooding of image feature noise from multiple viewpoints affecting the fusion effect. To address these issues, this paper proposes a multi-view fusion network (MVFD-Net) for crack detection in occluded underwater dams. First, we propose a feature reconstruction interaction encoder (FRI-Encoder), which interacts the multi-scale local features extracted by the convolutional neural network with the global features extracted by the transformer encoder and performs the feature reconstruction at the end of the encoder to enhance the feature extraction capability and at the same time in order to suppress the interference of the nonuniform scattering noise. Subsequently, a multi-scale gated adaptive fusion module is introduced between the encoder and the decoder for feature gated fusion, which further complements and recovers the noise flooding detail information. Additionally, this paper designs a multi-view feature fusion module to fuse multi-view image features to restore the occluded crack features and achieve the detection of occluded cracks. Through extensive experimental evaluations, the MVFD-Net algorithm achieves excellent performance when compared with current mainstream algorithms.

The multi-robot path planning problem is an NP-hard problem. The coati optimization algorithm (COA) is a novel metaheuristic algorithm and has been successfully applied in many fields. To solve multi-robot path planning optimization problems, we embed two differential evolution (DE) strategies into COA, a self-adaptive differential evolution-based coati optimization algorithm (SDECOA) is proposed. Among these strategies, the proposed algorithm adaptively selects more suitable strategies for different problems, effectively balancing global and local search capabilities. To validate the algorithm’s effectiveness, we tested it on CEC2020 benchmark functions and 48 CEC2020 real-world constrained optimization problems. In the latter’s experiments, the algorithm proposed in this paper achieved the best overall results compared to the top five algorithms that won in the CEC2020 competition. Finally, we applied SDECOA to optimization multi-robot online path planning problem. Facing extreme environments with multiple static and dynamic obstacles of varying sizes, the SDECOA algorithm consistently outperformed some classical and state-of-the-art algorithms. Compared to DE and COA, the proposed algorithm achieved an average improvement of 46% and 50%, respectively. Through extensive experimental testing, it was confirmed that our proposed algorithm is highly competitive. The source code of the algorithm is accessible at: https://ww2.mathworks.cn/matlabcentral/fileexchange/164876-HDECOA.

This research proposes an adaptive human-robot interaction (HRI) that combines voice recognition, emotional context detection, decision-making, and self-learning. The aim is to overcome challenges in dynamic and noisy environments while achieving real-time and scalable performance. The architecture is based on a three-stage HRI system: voice input acquisition, feature extraction, and adaptive decision-making. For voice recognition, modern pre-processing techniques and mel-frequency cepstral coefficients are used to robustly implement the commands. Emotional context detection is governed by neural network classification on pitch, energy, and jitter features. Decision-making uses reinforcement learning where actions are taken and then the user is prompted to provide feedback that serves as a basis for re-evaluation. Iterative self-learning mechanisms are included, thereby increasing the adaptability as stored patterns and policies are updated dynamically. The experimental results show substantial improvements in recognition accuracy along with task success rates and emotional detection. The proposed system achieved 95% accuracy and a task success rate of 96%, even against challenging noise conditions. It is apparent that emotional detection achieves a high F1-score of 92%. Real-world validation showed the system’s ability to dynamically adapt, thus mitigating 15% latency through self-learning. The proposed system has potential applications in assistive robotics, interactive learning systems, and smart environments, addressing scalability and adaptability for real-world deployment. Novel contributions to adaptive HRI arise from the integration of voice recognition, emotional context detection, and self-learning mechanisms. The findings act as a bridge between the theoretical advancements and the practical utility of further system improvements in human-robot collaboration.

Multi-loop coupling mechanisms (MCMs) are extensively utilized in the aerospace and aviation industries. This paper analyzes the mobility, singularity, and optimal actuation selection of a 3RR-3RRR MCM on the basis of geometric algebra (GA), where R denotes revolute joint. First, the principle of the shortest path is employed to identify the basic limbs and ascertain the type of coupling limbs. The analytical expression for the twist space and mobility characteristics of the mechanism is obtained by calculating the intersection of the limb’s twist space. The blade of limb constraint is subsequently employed to construct the singular polynomials of the mechanism. The singular configurations of the 3RR-3RRR MCM are analyzed in accordance with the properties of the outer product, resulting in the identification of two distinct types of boundary singularities. Next, the local transmission index is employed to evaluate the motion/force transmission performance of the two actuation schemes and finalize the selection of the superior actuation scheme for the mechanism. Finally, a prototype is developed to evaluate the energy loss resulting from the two actuation schemes, which verifies the correctness of the actuation selection scheme.

This paper proposes a generalized method for designing tendon-driven serial-chained manipulators with an arbitrary number of tendon redundancy. First, a special class of tendon-driven structures is defined by introducing the controllable block triangular form (CBTF) of a null space matrix and its complementary CBTF of a structure matrix, satisfying physical constraints related to the minimal connection of tendons and to the placement of actuators. Then it is shown that any general design of tendon-driven serial manipulators can be reduced to the design of such a special class of tendon-driven structures. Two associated design problems are derived and solved. The first design problem is about finding a complementary CBTF structure matrix for a given CBTF null space matrix using algebraic relations, whereas the second one seeks the both matrices that optimize the wanted structural characteristics based on the result of the first design problem. Numerical design examples are provided to show the validity of the proposed method.