Advanced intelligent systems (Weinheim an der Bergstrasse, Germany)最新文献

Mole-Inspired Robot for Planetary Exploration

This cover image (referring to article 2300392 by Tao Zhang, Kunyang Wang, Kun Xu, and co-workers) shows a mole-inspired robot burrowing inside the lunar regolith to create a small cave for in-situ detection or instrument placement in extraterrestrial subterranean exploration. The mole-inspired robot (child robot) connects to a rover (mother robot) with a cable for communication, power supply, and retraction. The rover first creates a cave entrance to fix the mole-inspired robot, and then the robot can burrow into the cave with low power consumption and high energy efficiency.

Cable-Crawling Robots

In article number 2300828, Bo Li, Guimin Chen, and co-workers report a centimeter-scaled cable-crawling robot for thin cables by utilizing multilayered dielectric elastomer actuators. The miniaturized robot (2.1 g, 43 mm long) can quickly crawl on extremely thin cables with different orientations, transport objects, and locomote across the water–air interface. Benefit from the excellent locomotion performance, the robot can be applied to perform inspection and maintenance tasks along pre-built cables inside some equipment with narrow, confined internal spaces, such as airplane wings. This novel centimeter-scaled robot offers a potential platform solution for operating in hazardous or difficult-to-access spaces.

Soft Robotic Morphing Wing for Unmanned Underwater Vehicles

Soft robots for underwater exploration are a powerful tool to assess ecological health and survey deep-sea environments autonomously. This cover illustrates an unmanned underwater vehicle featuring soft, hydraulically actuated wings deployed in the subsurface ocean of Enceladus. The incorporated soft wings have been designed to be pressure agnostic, impact resilient, and highly efficient, to enable autonomous surveying of large volumes. For further details, see article number 2300702 by Fabian Wiesemüller, Mirko Kovač, and co-workers.

Neuromorphic computing implemented with spiking neural networks (SNNs) based on volatile threshold switching is an energy-efficient computing paradigm that may overcome future limitations of the von Neumann architecture. Herein, threshold switching in oxyvanite (V₃O₅) memristors and their application as a leaky integrate-and-fire (LIF) neuron are explored. The spiking response of individual neurons is examined as a function of circuit parameters, input pulse train, and temperature and reveals a pulse height-dependent spike rate in which devices exhibit excitatory spiking behavior under low input voltages and protective inhibition spiking under high voltages. Resistively coupled LIF neurons are shown to exhibit additional neural functionalities (i.e., phasic, regular and adaptation, etc.) depending on the input voltage and circuit parameters. The behavior of both individual and coupled neurons is shown to be described by a physics-based lumped element circuit model, which therefore provides a solid foundation for exploring more complex systems. Finally, the performance of a perceptron SNN employing these LIF neurons is assessed by simulating the classification of image recognition algorithm. These results advance the development of robust solid-state neurons with low power consumption for neuromorphic computing.

Studying the behavior of electroactive cells, such as firing dynamics and chemical secretion, is crucial for developing human disease models and therapeutics. Following the recent advances in cell culture technology, traditional monolayers are optimized to resemble more 3D, organ-like structures. The biological and electrochemical complexity of these structures requires devices with adaptive shapes and novel features, such as precise electrophysiological mapping and stimulation in the case of brain- and heart-derived tissues. However, conventional organ-on-chip platforms often fall short, as they do not recreate the native environment of the cells and lack the functional interfaces necessary for long-term monitoring. Origami-on-a-chip platforms offer a solution for this problem, as they can flexibly adapt to the structure of the desired biological sample and can be integrated with functional components enabled by chosen materials. In this review, the evolution of origami-on-a-chip biointerfaces is discussed, emphasizing folding stimuli, materials, and critical findings. In the prospects, microfluidic integration, functional tissue engineering scaffolds, and multi-organoid networks are included, allowing patient-specific diagnoses and therapies through computational and in vitro disease modeling.

Surveillance cameras produce vast amounts of video data, posing a challenge for analysts due to the infrequent occurrence of unusual events. To address this, intelligent surveillance systems leverage AI and computer vision to automatically detect anomalies. This study proposes an innovative method combining 3D convolutions and long short-term memory (LSTM) modules to capture spatiotemporal features in video data. Notably, a structured coarse-level feature fusion mechanism enhances generalization and mitigates the issue of vanishing gradients. Unlike traditional convolutional neural networks, the approach employs depth-wise feature stacking, reducing computational complexity and enhancing the architecture. Additionally, it integrates microautoencoder blocks for downsampling, eliminates the computational load of ConvLSTM2D layers, and employs frequent feature concatenation blocks during upsampling to preserve temporal information. Integrating a Conv-LSTM module at the down- and upsampling stages enhances the model's ability to capture short- and long-term temporal features, resulting in a 42-layer network while maintaining robust performance. Experimental results demonstrate significant reductions in false alarms and improved accuracy compared to contemporary methods, with enhancements of 2.7%, 0.6%, and 3.4% on the UCSDPed1, UCSDPed2, and Avenue datasets, respectively.

Image preprocessing models are usually employed as the preceding operations of high-level vision tasks to improve the performance. The adversarial attack technology makes both these models face severe challenges. Prior research is focused solely on attacking single object detection models, without considering the impact of the preprocessing models (multifocus image fusion) on adversarial perturbations within the object detection system. Multifocus image fusion models work in conjunction with the object detection models to enhance the quality of the images and improve the capability of object detection system. Herein, the problem of attacking object detection system that utilizes multifocus image fusion as its preprocessing models is addressed. To retain the attack capabilities of adversarial samples against as many perturbations as possible, new attack method called adaptive retention attack (ARA) is proposed. Additionally, adversarial perturbations concentration mechanism and image selection mechanism, which, respectively, enhance the transferability and attack capability of ARA-generated adversarial samples. Extensive experiments have demonstrated the feasibility of the ARA. The results confirm that the ARA method can successfully bypass multifocus image fusion models to attack the object detection model.

This study introduces a flexure-based supernumerary robotic finger (FBSF) inspired by the proportions of the human thumb, aiming to overcome existing limitations in robotic finger design. In pursuit of seamless cooperation with the user's hand, human finger proportions are replicated. Finite element analysis of all five fingers indicates that the thumb-mimicking configuration offers the largest workspace and bending angle. The FBSF, featuring a polycarbonate paired crossed flexural hinge structure and high impact polystyrene links, closely mirrors the human thumb. Weighing 59 g (main body) and 170 g (control box), the FBSF enables user-driven control and decoupled actuation based on user intent, utilizing the electromyographic signal of the extensor carpi ulnaris via isometric contractions. An experimental protocol, including task blocks (releasing, clenching), confirms the FBSF's responsiveness to user intentions. When utilizing FBSF, it has been verified using a motion capture camera system that it is possible to extend the existing hand workspace by approximately 29.72%. Performance tests demonstrate the FBSF's capability to grasp various objects and assist in tasks, with a maximum load-bearing capacity of 2.6 kg experimentally verified. This study demonstrates the potential of the developed FBSF to augment hand functionality in diverse applications.

Fish have excellent swimming performance, and one key factor is their ability to autonomously adjust body stiffness, which can help them efficiently swim at different speeds and complex environments. At present, the variable-stiffness design of robotic fish still suffers from structural complexity, severe deformation, and small variation range, which limits the application of variable-stiffness theory in robotic fish. In this article, a variable-stiffness tail is designed based on layer-jamming for robotic fish, which can conveniently achieve online stiffness adjustment while maintaining the optimal stiffness distribution and the shape is unaffected. A modeling method for the tail is proposed by combining the mechanical characteristics of the layer-jamming structure with the pseudo-rigid body model. To validate the performance of the tail, a series of experiments are conducted, which show that the stiffness variation range of the tail is around 10 times, and the accuracy of the model in predicting the kinematics of the tail is also verified. Moreover, the thrust tests demonstrate that stiffness adjustment is beneficial for fish swimming at different frequencies. The proposed variable-stiffness tail will promote the development of efficient underwater biomimetic robots.

Modular robots have unique advantages in handling faults and improving their robustness due to their self-reconfiguration capacity and homogeneous interchangeability. When locked joint failure occurs in modular robots, the distribution of the failed modules will affect the manipulation capacity of the robot. In this article, a novel self-reconfiguration method that utilizes only the remaining resources available to mitigate the damage caused by module joint failures is proposed. The key node configuration is searched by the particle swarm optimization (PSO) algorithm, and then the collision-free reconfiguration path is planned by the rapidly exploring random trees (RRT) algorithm. The proposed method not only handles single-module lockup failure but can also be expanded to multi-module lockup failures, fully improving the fault tolerance of the modular robot. The method is deployed on the hardware, and the feasibility of the algorithm is verified by self-reconfiguration experiments containing faulty modules.