Artificial-Hand Technology—Current State of Knowledge in Designing and Forecasting Changes

sEMG-Based Robust Recognition of Grasping Postures with a Machine Learning Approach for Low-Cost Hand Control

The design and control of artificial hands remains a challenge in engineering. Popular prostheses are bio-mechanically simple with restricted manipulation capabilities, as advanced devices are pricy or abandoned due to their difficult communication with the hand. For social robots, the interpretation of human intention is key for their integration in daily life. This can be achieved with machine learning (ML) algorithms, which are barely used for grasping posture recognition. This work proposes an ML approach to recognize nine hand postures, representing 90% of the activities of daily living in real time using an sEMG human-robot interface (HRI). Data from 20 subjects wearing a Myo armband (8 sEMG signals) were gathered from the NinaPro DS5 and from experimental tests with the YCB Object Set, and they were used jointly in the development of a simple multi-layer perceptron in MATLAB, with a global percentage success of 73% using only two features. GPU-based implementations were run to select the best architecture, with generalization capabilities, robustness-versus-electrode shift, low memory expense, and real-time performance. This architecture enables the implementation of grasping posture recognition in low-cost devices, aimed at the development of affordable functional prostheses and HRI for social robots.

Do robots outperform humans in human-centered domains?

The incessant progress of robotic technology and rationalization of human manpower induces high expectations in society, but also resentment and even fear. In this paper, we present a quantitative normalized comparison of performance, to shine a light onto the pressing question, "How close is the current state of humanoid robotics to outperforming humans in their typical functions (e.g., locomotion, manipulation), and their underlying structures (e.g., actuators/muscles) in human-centered domains?" This is the most comprehensive comparison of the literature so far. Most state-of-the-art robotic structures required for visual, tactile, or vestibular perception outperform human structures at the cost of slightly higher mass and volume. Electromagnetic and fluidic actuation outperform human muscles w.r.t. speed, endurance, force density, and power density, excluding components for energy storage and conversion. Artificial joints and links can compete with the human skeleton. In contrast, the comparison of locomotion functions shows that robots are trailing behind in energy efficiency, operational time, and transportation costs. Robots are capable of obstacle negotiation, object manipulation, swimming, playing soccer, or vehicle operation. Despite the impressive advances of humanoid robots in the last two decades, current robots are not yet reaching the dexterity and versatility to cope with more complex manipulation and locomotion tasks (e.g., in confined spaces). We conclude that state-of-the-art humanoid robotics is far from matching the dexterity and versatility of human beings. Despite the outperforming technical structures, robot functions are inferior to human ones, even with tethered robots that could place heavy auxiliary components off-board. The persistent advances in robotics let us anticipate the diminishing of the gap.

Recent advances in the 3D printing of ionic electroactive polymers and core ionomeric materials

The recent advances in the 3D printing, or additive manufacturing, of ionic electroactive polymers (EAP) and their future applications.

Manipulation Planning for Large Objects through Pivoting, Tumbling, and Regrasping

Robotic manipulation of a bulky object is challenging due to the limited kinematics and payload of the manipulator. In this study, a robot realizes the manipulation of general-shaped bulky objects utilizing the contact with the environment. We propose a hierarchical manipulation planner that effectively combined three manipulation styles, namely, pivoting, tumbling, and regrasping. In our proposed method, we first generate a set of superimposed planar segments on the object surface to obtain an object pose in stable contact with the table, and a set of points on the object surface for the end-effectors (EEFs) of a dual-arm manipulator to stably grasp the object. Object manipulation can be realized by solving a graph, considering the kinematic constraints of pivoting and tumbling. For pivoting, we consider two supporting styles: stable support (SP) and unstable support (USP). Our proposed method manipulates large and heavy objects by selectively using the two different support styles of pivoting and tumbling according to the conditions on the table area. In addition, it can effectively avoid the limitation arising due to the arm kinematics by regrasping the object. We experimentally demonstrate that a dual-arm manipulator can move an object from the initial to goal position within a limited area on the table, avoiding obstacles placed on the table.