International Journal of Intelligent Robotics and Applications

Perceiving distributed tactile information during robotic dexterous grasping process is essential to improve its intelligence and automation. This paper presents a tactile sensing system for a multi-fingered robotic hand, and used to detect distributed contact forces andtactile information during robotic hand dexterous multimodal grasping applications. The tactile sensing system relies on the design of tactile sensors with multiple sensing units and is integrated onto robotic thumb, index and middle fingers, respectively. For robotic dexterous grasping, six types of grasping modes are selected and performed objects’ grasping experiments. Using the developed robotic hand tactile sensing system, the generated contact forces during grasping processes are recorded. Through analyzing the robotic hand grasp actions in each grasping mode, the characteristics of the detected tactile forces can be studied and compared, which can be used as the factor to further distinguish the different grasping modes. Therefore, our developed robotic hand tactile sensing system can provide the possibility to accurately measure the distributed contact forces during robotic hand dexterous manipulation applications, and be used for analyzing the relationship between tactile information characteristics and grasping mode movements.

This study presents a comparative analysis to be a resource for those who want to work in the wearable hand robot field. The fact that robotic rehabilitation is more accessible than classical methods and shows promising results increases interest in it. When the last 20 years are analyzed, the number of studies in this field has increased continuously, without exception. According to Google Scholar data, 17,200 studies were conducted only in 2022 related to hand rehabilitation robots. When the literature is reviewed, thousands of studies are faced with design and control systems that seem to have been designed into many different types and structures. The number of studies reviewing and surveying some of these is also insufficient to solve the confusion in the minds of the researchers in a shorter way. Rather than reporting what other researchers have done, our work is based on classifying and putting concepts together instead of creating dozens of sub-categories. We aim to inform researchers who are interested in the field but are hesitant, to create a template in their minds, and to suggest the most beneficial design alternatives in terms of mechanics and hardware-software implementations. The most cited and recent studies were selected, and a systematic categorization was made based on criteria such as kinematic design, control system, and actuator type. The kinematic categorization it employs is an original classification specific to this study and has been applied to all models examined.

This paper presents a walking pattern detection method for a smart rollator. The method detects the rollator user’s lower extremities from the depth data of an RGB-D camera. It then segments the 3D point data of the lower extremities into the leg and foot data points, from which a skeletal system with 6 skeletal points and 4 rods is extracted and used to represent a walking gait. A gait feature, comprising the parameters of the gait shape and gait motion, is then constructed to describe a walking state. K-means clustering is employed to cluster all gait features obtained from a number of walking videos into 6 key gait features. Using these key gait features, a walking video sequence is modeled as a Markov chain. The stationary distribution of the Markov chain represents the walking pattern. Three Support Vector Machines (SVMs) are trained for walking pattern detection. Each SVM detects one of the three walking patterns. Experimental results demonstrate that the proposed method has a better performance in detecting walking patterns than seven existing methods.

This paper presents a non-singular terminal sliding mode control (NTSMC) design based on an improved extended state observer (IESO) with application to an omnidirectional mobile manipulator (OMM) for trajectory tracking control. Firstly, a unified dynamic model is derived based on Lagrange method for an OMM prototype. An IESO that can reduce the initial peaking phenomenon is applied to estimate the model uncertainties and external disturbances. Then a non-singular terminal sliding mode controller is applied for trajectory tracking control. Stability of the closed-loop system is analyzed using Lyapunov theory. Finally, both simulations and experimental tests verify the effectiveness of the proposed control scheme.

This paper proposes an architecture that explores a gap in the spectrum of existing strategies for robot control mode switching in adjustable autonomy. In situations where the environment is reasonably known and/or predictable, pre-planning these control changes could relieve robot operators of the additional task of deciding when and how to switch. Such a strategy provides a clear division of labour between the automation and the human operator(s) before the job even begins, allowing for individual responsibilities to be known ahead of time, limiting confusion and allowing rest breaks to be planned. Assigned Responsibility is a new form of adjustable autonomy-based teleoperation that allows the selective inclusion of automated control elements at key stages of a robot operation plan’s execution. Progression through these stages is controlled by automatic goal accomplishment tracking. An implementation is evaluated through engineering tests and a usability study, demonstrating the viability of this approach and offering insight into its potential applications.

The ability to precisely and efficiently manipulate nano- and micro-scale objects is a key step in enabling applications in areas such as nano-medicine and nano/micro-device assembly. In this paper, we propose and demonstrate real-time motion-planning algorithms for effectively steering multiple nanowires simultaneously in fluid suspension using electric fields. The proposed motion planners $$\mathtt{{SRRT}}^*$$ and learning- $$\mathtt{{SRRT}}^*$$ ( $$\mathtt{{LSRRT}}^*$$ ) are built on and extended by sparse-structure, Rapidly-exploring Random Tree (RRT)-based motion-planning algorithms. The algorithms also use heuristics to effectively guide the search process to quickly generate initial solutions. In order to reduce the online computational cost, the $$\mathtt{{LSRRT}}^*$$ algorithm shifts the intensive computation of optimal additive cost metric to offline training using a supervised learning method. Compared with the previously developed network-flow and RRT-based algorithms, the $$\mathtt{{SRRT}}^*$$ and $$\mathtt{{LSRRT}}^*$$ guarantee the asymptotically near-optimal trajectories of multiple nanowires. Moreover, the algorithms do not restrict in theory the maximum number of nanowires that can be simultaneously controlled. We present simulation and experimental results to demonstrate the minimum-time performance of the motion-planning and control design. The results demonstrate that the proposed algorithms significantly reduce the computational cost and increase the efficiency of manipulating multiple nanowires simultaneously when compared with the $${\mathtt{RRT}}^*$$ and other $${\mathtt{RRT}}^*$$ enhanced algorithms.

Recent works in the person re-identification task mainly focus on the model accuracy while ignoring factors related to efficiency, e.g., model size and latency, which are critical for practical application. In this paper, we propose a novel Hierarchical and Efficient Network (HENet) that learns hierarchical global, partial, and recovery features ensemble under the supervision of multiple loss combinations. To further improve the robustness against the irregular occlusion, we propose a new dataset augmentation approach, dubbed random polygon erasing, to random erase the input image’s irregular area imitating the body part missing. We also propose an Efficiency Score (ES) metric to evaluate the model efficiency. Extensive experiments on Market1501, DukeMTMC-ReID, and CUHK03 datasets show the efficiency and superiority of our approach compared with epoch-making methods. We further deploy HENet on a robotic car, and the experimental result demonstrates the effectiveness of our method for robotic navigation.