Search

Photothermal hydrogenation of carbon dioxide (CO2) into value-added products is an ideal solution for addressing the energy crisis and mitigating CO2emissions. However, achieving high product selectivity remains challenging due to the simultaneous occurrence of numerous competing intermediate reactions during CO2hydrogenation. We present a novel approach featuring isolated single-atom nickel (Ni) anchored onto indium oxide (In2O3) nanocrystals, serving as an effective photothermal catalyst for CO2hydrogenation into methane (CH4) with a remarkable near-unity (∼99%) selectivity. Experiments and theoretical simulations have confirmed that isolated Ni sites on the In2O3surface can effectively stabilize the intermediate products of the CO2hydrogenation reaction and reduce the transition state energy barrier, thereby changing the reaction path to achieve ultrahigh selective methanation. This study provides comprehensive insights into the design of single-atom catalysts for the highly selective photothermal catalytic hydrogenation of CO2to methane.

Different from the previous neutral reaction solvent system, this work explores the synthesis of Al-oxo rings in ionic environments. Deep eutectic solvents (DESs) formed by quaternary ammonium salts hydrogen bond acceptor (HBA) and phenols hydrogen bond donor (HBD) further reduce the melting point of the reaction system and provide an ionic environment. Further, the quaternary ammonium salt was chosen as the HBA because it contains a halogen anion that matches the size of the central cavity of the molecular ring. Based on this thought, five Al8ion pair cocrystals were synthesized via “DES thermal”. The general formula is Q+⊂ {Cl@[Al8(BD)8(μ2-OH)4L12]} (AlOC-180–AlOC-185, Q+= tetrabutylammonium, tetrapropylammonium, 1-butyl-3-methylimidazole; HBD = phenol, p-chlorophenol, p-fluorophenol; HL = benzoic acid, 1-naphthoic acid, 1-pyrenecarboxylic acid, anthracene-9-carboxylic acid). Structural studies reveal that the phenol-coordinated Al molecular ring and the quaternary ammonium ion pair form the cocrystal compounds. The halogen anions in the DES component are confined in the center of the molecular ring, and the quaternary ammonium cations are located in the organic shell. Such an adaptive cocrystal binding pattern is particularly evident in the structures coordinated with low-symmetry ligands such as naphthoic acid and pyrene acid. Finally, the optical behavior of these cocrystal compounds is understood from the analysis of crystal structure and theoretical calculation.

Purpose: The purpose of this study is realizing human-like motions and performance through musculoskeletal robots and brain-inspired controllers. Human-inspired robotic systems, owing to their potential advantages in terms of flexibility, robustness and generality, have been widely recognized as a promising direction of next-generation robots. Design/methodology/approach: In this paper, a deep forward neural network (DFNN) controller was proposed inspired by the neural mechanisms of equilibrium-point hypothesis (EPH) and musculoskeletal dynamics. Findings: First, the neural mechanism of EPH in human was analyzed, providing the basis for the control scheme of the proposed method. Second, the effectiveness of proposed method was verified by demonstrating that equilibrium states can be reached under the constant activation signals. Finally, the performance was quantified according to the experimental results. Originality/value: Based on the neural mechanism of EPH, a DFNN was crafted to simulate the process of activation signal generation in human motion control. Subsequently, a bio-inspired musculoskeletal robotic system was designed, and the high-precision target-reaching tasks were realized in human manner. The proposed methods provide a direction to realize the human-like motion in musculoskeletal robots.

Purpose: Limited by the types of sensors, the state information available for musculoskeletal robots with highly redundant, nonlinear muscles is often incomplete, which makes the control face a bottleneck problem. The aim of this paper is to design a method to improve the motion performance of musculoskeletal robots in partially observable scenarios, and to leverage the ontology knowledge to enhance the algorithm’s adaptability to musculoskeletal robots that have undergone changes. Design/methodology/approach: A memory and attention-based reinforcement learning method is proposed for musculoskeletal robots with prior knowledge of muscle synergies. First, to deal with partially observed states available to musculoskeletal robots, a memory and attention-based network architecture is proposed for inferring more sufficient and intrinsic states. Second, inspired by muscle synergy hypothesis in neuroscience, prior knowledge of a musculoskeletal robot’s muscle synergies is embedded in network structure and reward shaping. Findings: Based on systematic validation, it is found that the proposed method demonstrates superiority over the traditional twin delayed deep deterministic policy gradients (TD3) algorithm. A musculoskeletal robot with highly redundant, nonlinear muscles is adopted to implement goal-directed tasks. In the case of 21-dimensional states, the learning efficiency and accuracy are significantly improved compared with the traditional TD3 algorithm; in the case of 13-dimensional states without velocities and information from the end effector, the traditional TD3 is unable to complete the reaching tasks, while the proposed method breaks through this bottleneck problem. Originality/value: In this paper, a novel memory and attention-based reinforcement learning method with prior knowledge of muscle synergies is proposed for musculoskeletal robots to deal with partially observable scenarios. Compared with the existing methods, the proposed method effectively improves the performance. Furthermore, this paper promotes the fusion of neuroscience and robotics.

Organisms naturally have extraordinary motor ability. They can autonomously control the highly redundant and nonlinear musculoskeletal system to complete fast and flexible movements. Motor primitive theory proposed that a complex movement can be realized by modular organization of simple movement patterns. This theory provides a convincing theoretical basis for explaining the extraordinary motor ability of organisms, and has been supported by plenty of experimental evidences in neurosciences. As a kind of typical motor primitive related to the modular regulation of the spinal motor system, convergent force field was thought to be significant for reducing the complexity of controlling numerous muscles and joints simultaneously. Inspired by the primitive property of convergent force field, in this paper, we proposed a new algorithm to efficiently construct constraint force field on musculoskeletal system with highly redundant actuators by taking optimized parameters of constraint force field as motor primitives. Compared with previous methods of constructing constraint force field, the proposed algorithm is able to reduce the dimension of the solution space so as to effectively improve the computational efficiency of constructing constraint force field in musculoskeletal system with redundant muscles. Validation experiments were carried out in a musculoskeletal system with 10 redundant muscles. The number of optimization iterations and computational time required for constructing constraint force field at a new position were significantly reduced. The system can accurately reach the target position using constant activations. This work may bring in new inspiration for realizing efficient and accurate motion control of musculoskeletal robots with limited feedback accuracy.

As the global production of medical waste continues to rise, effective management of medical waste has become essential. In the era of Industry 5.0, new intelligent technologies such as the Internet of Things (IoT) and blockchain have been applied in Medical waste management (MWM) systems. Despite their advantages, these technologies face challenges related to information security, influenced by various factors. Therefore, the information security factors identification and prioritization are essential for safety improvement. This paper aims to develop an FMEA-based framework to analyze the risks of information security factors of intelligent technologies applied to MWM systems. First, Probabilistic linguistic term sets (PLTSs) are employed to capture the uncertainty of assessment information. Then, the Criteria Importance through Intercriteria Correlation (CRITIC) method is utilized to handle the interactive relationship among risk criteria. Finally, the Preference Ranking Organization Method for Enrichment Evaluations-II (PROMETHEE-II) method prioritizes the risk of information security factors. An example of the MWM system in Istanbul, Turkey, is selected to illustrate the validity of the proposed framework. Third-party risk management is the most critical information security factor. The findings of this research can provide valuable insights and practical recommendations for decision-makers in MWM systems.

It is a significant research direction for highly complex musculoskeletal robots that how to develop the ability of motion learning and generalization. The cooperations of multiple brain regions are crucial to improving motion performance. Inspired by the neural mechanisms of structures, functions, and interconnections of basal ganglia and cerebellum, a biologically inspired integration model for motor learning of musculoskeletal robots is proposed. Based on the neural characteristics of the basal ganglia, the basal ganglia actor network, which mainly simulates the dorsal striatum, outputs motion commands, and the basal ganglia critic network, which simulates the ventral striatum, estimates action-state values. Their network parameters are updated using the soft actor-critic method. Based on the sensorimotor prediction mechanism of the cerebellum, the cerebellum network evaluates the state feature extraction quality of the basal ganglia actor network and then updates the weights of its feature layer. This learning method is proven to converge to the optimal policy. Furthermore, drawing on the mechanism of dopaminergic dynamic regulation in the basal ganglia, the adaptive adjustment of target entropy and the dopaminergic experience replay are proposed to further improve the integration model, which contributes to the exploration-exploitation trade-off of motor learning. The bio-inspired integration model is validated on a musculoskeletal system. Experimental results indicate that this model can effectively control the musculoskeletal robot to accomplish the motion task from random starting locations to random target positions with high precision and robustness.

The continual learning and development are significant for robots to learn multiple tasks sequentially. The difficulty lies in balancing the efficient learning of new tasks and overcoming catastrophic forgetting of old tasks. Although many continual learning methods have been proposed for pattern recognition, continual reinforcement learning methods for redundant musculoskeletal and robotic systems are few and have limitations. Therefore, inspired by the developmental mechanisms in motor cortex, this article proposes a neural manifold modulated continual reinforcement learning method for musculoskeletal and robotic systems. First, a recurrent neural network (RNN) with an expected neural manifold is designed and conditions of weights are derived. Second, the ability of projectors for characterizing the neural manifold within RNN is analyzed. Furthermore, a continual reinforcement learning method of RNN is proposed with the modulation of a neural manifold. The method is validated by redundant musculoskeletal and robotic systems in simulation. The results suggest that it can realize continual reinforcement learning of multiple tasks in different movements and environments. Furthermore, compared with related works, the proposed method achieves better performance.

Tracking-and-inserting a moving peg using a robot manipulator is a challenging task in manufacturing. In the past decades, various visual-based methods have been proposed for robotic manipulating static targets, which usually ignore the time delay in robot command transmission and image processing. However, for tracking and inserting a moving peg, time delays cannot be overlooked because they can reduce the tracking performance and even cause manipulations to fail. In this article, a robot visual-based control with a time-delay compensator is presented to solve the problem of inserting a moving peg. The time-delay compensator was designed using radial basis function neural network, a feedback compensator aimed at eliminating the tracking errors caused by the time delays. Thus, we could manipulate a moving object using a commercial industry robot, even with the time-variant delays in the control loop. Furthermore, the visual-based controller with the pseudoinverse image Jacobian matrix was designed using a linearization model. Thus, the matrix could be efficiently updated using the model. In the experiment, we inserted a peg into a moving hole using an eye-in-hand robot with precision.

The spiking neural networks (SNNs), as brain-inspired neural networks, have received noteworthy attention due to their advantages of low power consumption, high parallelism, and high fault tolerance. While SNNs have shown promising results in uni-modal data tasks, their deployment in multimodal audiovisual classification remains limited, and the effectiveness of capturing correlations between visual and audio modalities in SNNs needs improvement. To address these challenges, we propose a novel model called spiking multimodel transformer (SMMT) that combines SNNs and Transformers for multimodal audiovisual classification. The SMMT model integrates uni-modal subnetworks for visual and auditory modalities with a novel spiking cross-attention module for fusion, enhancing the correlation between visual and audio modalities. This approach leads to competitive accuracy in multimodal classification tasks with low energy consumption, making it an effective and energy-efficient solution. Extensive experiments on a public event-based data set (N-TIDIGIT&MNIST-DVS) and two self-made audiovisual data sets of real-world objects (CIFAR10-AV and UrbanSound8K-AV) demonstrate the effectiveness and energy efficiency of the proposed SMMT model in multimodal audiovisual classification tasks. Our constructed multimodal audiovisual data sets can be accessed at https://github.com/Guo-Lingyue/SMMT.

The pivotal role of clusters and aromaticity in chemistry is undeniable, but there remains a gap in systematically understanding the aromaticity of metal–organic clusters. Herein, this article presents a novel metal–organic π-cluster, melding both metal–organic chemistry and aromaticity, to guide the construction of structurally stable Os-organic π-clusters. An in-depth analysis of these clusters reveals their bonding attributes, π-electronic composition, and origins of aromaticity, thereby confirming their unique metal–organic π-cluster properties. Furthermore, the Os5cluster exhibits a promising third-order nonlinear optical (NLO) response, attributable to its narrow band gap and uniform electron/hole distribution, suggesting its potential as an optical switching material. This research introduces a fresh perspective on clusters, centered on delocalization, and broadens the domain of aromaticity studies. It also presents a novel method for designing efficient third-order NLO materials through consideration of the structure–activity relationship.

Manifold learning now plays an important role in machine learning and many relevant applications. In spite of the superior performance of manifold learning techniques in dealing with nonlinear data distribution, their performance would drop when facing the problem of data sparsity. It is hard to obtain satisfactory embeddings when sparsely sampled high-dimensional data are mapped into the observation space. To address this issue, in this article, we propose hierarchical neighbors embedding (HNE), which enhances the local connections through hierarchical combination of neighbors. And three different HNE-based implementations are derived by further analyzing the topological connection and reconstruction performance. The experimental results on both the synthetic and real-world datasets illustrate that our HNE-based methods could obtain more faithful embeddings with better topological and geometrical properties. From the view of embedding quality, HNE develops the outstanding advantages in dealing with data of general distributions. Furthermore, comparing with other state-of-the-art manifold learning methods, HNE shows its superiority in dealing with sparsely sampled data and weak-connected manifolds.

Phenolic compounds are considered the main bioactive components in kiwifruit. In order to well understand the accumulation pattern of health-beneficial phenolics, and to obtain high-quality kiwifruits with health-promoting characteristics, the changes of physicochemical properties, phenolic compounds, antioxidant capacities, and inhibitory effects on digestive enzymes of “Donghong” kiwifruits during maturation were systematically investigated. Noticeable variations in physicochemical properties and phenolic profiles were observed throughout the maturation stages of “Donghong” kiwifruit. According to the HPLC analysis, phenolic compounds, including chlorogenic acid, (−)-epicatechin, procyanidin B2, quercetin-3-O-rhamnoside, gallic acid, neochlorogenic acid, and caffeic acid, were found in “Donghong” kiwifruit, while chlorogenic acid was the most abundant phenolic compound. Considering the accumulation patterns of both total phenolics and soluble solids content, the optimum harvest time of kiwifruit with relatively high level of health-beneficial components was determined. Furthermore, “Donghong” kiwifruit exerted remarkable antioxidant capacities, and inhibitory effects on pancreatic lipase and α-glucosidase. Indeed, the Pearson’s correlation showed that chlorogenic acid, caffeic acid, (−)-epicatechin, and quercetin-3-O-rhamnoside could be major contributors to the antioxidant activity and inhibitory effects on digestive enzymes. Results are beneficial for better understanding of the accumulation patterns of health-beneficial phenolic compounds, which can aid in the targeting of specific developmental stages with an optimal phenolic profile for the production of health-beneficial products.

This article targets the task of novel category discovery (NCD), which aims to discover unknown categories when a certain number of classes are already known. The NCD task is challenging due to its closeness to real-world scenarios, where we have only encountered some partial classes and corresponding images. Unlike previous approaches to NCD, we propose a novel adaptive prototype learning method that leverages prototypes to emphasize category discrimination and alleviate the issue of missing annotations for novel classes. Concretely, the proposed method consists of two main stages: prototypical representation learning and prototypical self-training. In the first stage, we develop a robust feature extractor that could effectively handle images from both base and novel categories. This ability of instance and category discrimination of the feature extractor is boosted by self-supervised learning and adaptive prototypes. In the second stage, we utilize the prototypes again to rectify offline pseudo labels and train a final parametric classifier for category clustering. We conduct extensive experiments on four benchmark datasets, demonstrating our method’s effectiveness and robustness with state-of-the-art performance.

Compared to general joint-link robotic systems, bioinspired musculoskeletal robotic systems provide more potential advantages in terms of robustness, flexibility, and operation accuracy. Research on bioinspired control and structure has been a hot topic in recent years, indicating that musculoskeletal robots are a promising option for next-generation robots. However, standard modeling and analysis of the system are scarce; therefore, controller design and stability proof are still open questions. Based on previous studies, we built the standardized state-space equations for musculoskeletal robotic dynamics. The robustness was proved through muscle contractile dynamics and the Lyapunov stability theorem. Furthermore, a bioinspired equilibrium-point (EP) controller was designed to realize high-precision target-reaching tasks that required low control frequency and computation cost. Simulations and experiments were conducted to demonstrate the conclusions of the theoretical analysis. The effectiveness and anti-interference abilities of the EP controller were initially verified. This study provides a promising direction for motion control in bioinspired musculoskeletal robotic systems. A relatively complete system model was established, and a preliminary controller design and proof framework was proposed, which offers a reference for research and applications of musculoskeletal robotic systems.

Abstract: At high temperature, the chemical reaction mechanism at the interface of magnetite and the influence mechanism of gangue element Al on the oxidation performance of magnetite are not clear. In addition, due to the limitation of existing experimental equipment, it is difficult to clarify the interface reaction mechanism in the oxidation process and the influence mechanism of Al on the surface oxidation reaction of Fe3O4at an atomic scale at high temperature. The surface oxidation reaction mechanism of magnetite and the influence mechanism of Al on the oxidation of magnetite were studied by experiments and AIMD (ab initio molecular dynamics). The experimental results show that the existence of Al2O3will reduce the initial oxidation temperature and comprehensive oxidation performance of magnetite. The AIMD results show that aluminum can accelerate the interface reaction rate of Fe3O4in the high-temperature oxidation atmosphere, but the strong Al–O binding ability is not conducive to the transfer of O atoms in the oxidation process.

For RGB-based temporal action segmentation (TAS), excellent methods that capture frame-level features have achieved remarkable performance. However, for motion-centered TAS, it is still challenging for existing methods that ignore the extraction of spatial features of joints. In addition, inaccurate action boundaries caused by the frames of similar motion destroy the integrity of the action segments. To alleviate the issues, an end-to-end Involving Distinguished Temporal Graph Convolutional Networks called IDT-GCN is proposed. First, we construct an enhanced spatial graph structure that adaptively captures the similar and differential dependencies between joints in a single topology through learning two independent correlation modeling functions. Then, the proposed Involving Distinguished Graph Convolutional (ID-GC) models the spatial correlations of different actions in a video by using multiple enhanced topologies on the corresponding channels. Furthermore, we design a generic modeling temporal action regression network, termed Temporal Segment Regression (TSR), to extract segmented encoding features and action boundary representations by modeling action sequences. Combining them with label smoothing modules, we develop powerful spatial-temporal graph convolutional networks (IDT-GCN) for fine-grained TAS, which notably outperforms state-of-the-art methods on the MCFS-22 and MCFS-130 datasets. Adding TSR to TCN-based baseline methods achieves competitive performance compared with the state-of-the-art transformer-based methods on RGB-based datasets, i.e., Breakfast and 50Salads. Further experimental results on the action recognition task verify the superiority of the enhanced spatial graph structure over the previous graph convolutional networks.

Post-assembly modification strategy is used as an effective approach to improve the functional applications of metal-organic cages. The active vertex of the racemic tetrahedral Ti4L6(L = embonate) cages traps four chiral diamine chelated dinuclear silver ions, obtaining homochiral cages which exhibit helical perovskite stacks. Such a surface modification engineering enables the structural precise cages presenting homochirality and circularly polarized luminescence (CPL) response, further achieving enantioselective recognition toward chiral Camphorquinone molecules and displaying enhanced temperature-dependent effects of CPL output.

It is an important issue that how to regulate the existing control models of musculoskeletal robots to improve the ability of motion learning and generalization. In this article, based on the motion modulation function of the cerebellum, a cerebellum-inspired prediction and correction model is proposed to carry out feedforward regulation of the original controller. First, drawing on the reservoir computing mechanism of the cerebellar granular layer, the cerebellum prediction model is established by using the echo state network. Incremental learning for the network is achieved using the replay method, which is able to process control signals with different distributions. The cerebellum prediction network can accurately predict the motion results of the robot under the action of the time-series control signals. Second, referring to the neural pathways of the cerebellum, the cerebellum correction model is constructed. The network learning rules are designed by drawing on the long-term potentiation and depression processes of cerebellar synaptic plasticity. The characteristics of the parameters in the network weight update equations are further analyzed. And the hyperparameter update rules of the correction network weights are proposed. To simulate the function of the cerebellum involved in the limb ballistic movement, the adaptive adjustment method for cerebellum correction duration is proposed. The cerebellum correction network can accurately modulate the original control signals by using the motion prediction results. In experiments, a musculoskeletal robot is used to verify the movement effects under the control of the cerebellum model. The results show that the cerebellum-inspired model can effectively improve the motion accuracy and learning efficiency of the musculoskeletal robot, and enhance the motion generalization ability and system robustness of the robot.

In this paper, the optimal tracking control for robotic manipulators with state constraints and uncertain dynamics is investigated, and a sliding mode-based adaptive tube model predictive control method is proposed. First, utilizing the high-order fully actuated system approach, the nominal model of the robotic manipulator is constructed as the predictive model. Based on the nominal model, a nominal model predictive controller with the sliding mode is designed, which relaxes the terminal constraints, and realizes the accurate and stable tracking of the desired trajectory by the nominal system. Then, an auxiliary controller based on the node-adaptive neural networks is constructed to dynamically compensate nonlinear uncertain dynamics of the robotic manipulator. Furthermore, the estimation deviation between the nominal and actual states is limited to the tube invariant sets. At the same time, the recursive feasibility of nominal model predictive control is verified, and the ultimately uniformly boundedness of all variables is proved according to the Lyapunov theorem. Finally, experiments show that the robotic manipulator can achieve fast and efficient trajectory tracking under the action of the proposed method.

In order to determine the polarizability and hyperpolarizability of a molecule, several key parameters need to be known, including the excitation energy of the ground and excited states, the transition dipole moment, and the difference of dipole moment between the ground and excited states. In this study, a machine-learning model was developed and trained to predict the molecular polarizability and second-order hyperpolarizability on a subset of QM9 data set. The density of states was employed as input to the model. The results demonstrated that the machine-learning model effectively estimated both polarizability and the order of magnitude of second-order hyperpolarizability. However, the model was unable to predict the dipole moment and first-order hyperpolarizability, suggesting limitations in its ability to predict the difference of dipole moment between the ground and excited states. The computational efficiency of machine-learning models compared to traditional quantum mechanical calculations enables the possibility of large-scale screening of molecules that satisfy specific requirements using existing databases. This work presents a potential solution for the efficient exploration and analysis of molecules on a larger scale.

Traditional joint-link robots have been widely used in production lines because of their high precision for single tasks. With the development of the manufacturing and service industries, the requirement for the comprehensive performance of robotics is growing. Numerous types of bio-inspired robotics have been investigated to realize human-like motion control and manipulation. A study route from inner mechanisms to external structures is proposed to imitate humans and animals better. With this idea, a brain-inspired intelligent robotic system is constructed that contains visual cognition, decision-making, motion control, and musculoskeletal structures. This paper reviews cutting-edge research in brain-inspired visual cognition, decision-making, motion control, and musculoskeletal systems. Two software systems and a corresponding hardware system are established, aiming at the verification and applications of next-generation brain-inspired musculoskeletal robots.

Revolutionary advances have occurred in robot learning research, where a resurgence in reinforcement learning (RL) algorithms has fueled breakthroughs in acquiring complicated robotic skills without human intervention. Unfortunately, one limitation that has hampered RL methods is that RL may be quite inefficient, that is, it may cost unrealistic learning time and prohibitively large numbers of trajectories to provide implausible models for achieving multitask learning. In a broader perspective, the realization of robotics control by RL relies heavily on the availability of compact and expressive representations of the state spaces. In this article, we propose to learn vivid general structural representations by utilizing structural prior knowledge of robots. Particularly, a novel framework called structural unsupervised representations for robot learning (SURRL) is presented to enable multitask learning. The task-agnostic sample trajectories are leveraged to learn the structural representations that are constructed by graph autoencoder (GAE) in an unsupervised fashion. When learning a new task, the learned structural representations can be directly adopted for subsequent policy learning without training from scratch. Extensive experiments on continuous robotic environments, including dexterous manipulation tasks, characterize our method’s effectiveness to learn optimal policies for handling multiple tasks learning, demonstrating significant improvements over other competitive methods.

Feature correspondence is an essential issue in computer science, which could be well formulated by graph matching (GM). However, traditional GM is susceptible to outliers, thus limiting the applications. To address the issue, we present a biologically inspired feature descriptor (BIFD) corresponding to the simple and complex cell layers in primary visual cortex, which shows robust performance against deformations. Furthermore, we propose a novel biologically inspired structural model (BISM) for feature correspondence by fusing the graph structural information and appearance information described by BIFD in the images. The proposed BIFD imitates the cortical pooling operation across multiscale and multiangle cell layers, which makes BISM robust to outliers and distortions. To demonstrate the validity of the proposed method, we evaluate it in feature correspondence tasks on the public databases. The experimental results on synthetic data prove the validity of the proposed method.

The four-pin gripper is an ideal end-effector to achieve form-closure grasps, which is especially useful in scenarios where a stable grasp is necessary. In this article, we propose an improved parallel four-pin gripper design. Compared with the previous designs, the new design achieves more planar configurations while introducing only one more actuator, thus enlarging the possibility of the gripper to stably grasp. Based on the improved design, we extend the environment constraint-based form-closure grasping point search algorithm to find the optimal parameters for the improved four-pin gripper. Furthermore, we apply an inverse kinematics-based planning method to drive the gripper from any initial state to a form-closure state. The simulation and experiment demonstrate that, compared with the previous designs, the improved design can achieve stable grasping for a wider range of planar objects.

Robust object detection methods in traffic surveillance scenarios often encounters challenges due to large-scale deformations and illumination variations in outdoor scenes. To enhance the tolerance of such methods against these variations, we design a cross-scale and illumination-invariant detection model (CSIM) based on the You Only Look Once (YOLO) architecture. A main cause of false detection in large-scale detection tasks is the inconsistency between various feature scales. To address this issue, we introduce an adaptive cross-scale feature fusion model to ensure the consistency of the constructed feature pyramid. To overcome the influence of uneven light, we build an illumination-invariant chromaticity space on the CSIM model, which is independent of the correlated color temperature. In addition, we adopt spatial attention modules, K-means clustering and the Mish activation function for further model optimization. The obtained experimental results show that the proposed CSIM produces excellent detection results for addressing the challenges derived from large-scale deformations and the illumination changes encountered during traffic surveillance. Compared with state-of-the-art object detection methods on public datasets, our proposed model has achieved competitive results in robust object detection tasks in traffic surveillance scenarios.

Imbalanced classification tasks are widespread in many real-world applications. For such classification tasks, in comparison with the accuracy rate (AR), it is usually much more appropriate to use nondecomposable performance measures such as the area under the receiver operating characteristic curve (AUC) and the $F_\beta $ measure as the classification criterion since the label class is imbalanced. On the other hand, the minimax probability machine is a popular method for binary classification problems and aims at learning a linear classifier by maximizing the AR, which makes it unsuitable to deal with imbalanced classification tasks. The purpose of this article is to develop a new minimax probability machine for the $F_\beta $ measure, called minimax probability machine for the $F_\beta $ -measures (MPMF), which can be used to deal with imbalanced classification tasks. A brief discussion is also given on how to extend the MPMF model for several other nondecomposable performance measures listed in the article. To solve the MPMF model effectively, we derive its equivalent form which can then be solved by an alternating descent method to learn a linear classifier. Further, the kernel trick is employed to derive a nonlinear MPMF model to learn a nonlinear classifier. Several experiments on real-world benchmark datasets demonstrate the effectiveness of our new model.

The DELLAgenes, also known as ‘Green Revolution’ genes, encode conserved master growth regulators that control plant development in response to internal and environmental cues. Functioning as nuclear-localized transcription regulators, DELLAs modulate expression of target genes via direct protein–protein interaction of their carboxy-terminal GRAS domain with hundreds of transcription factors (TFs) and epigenetic regulators. However, the molecular mechanism of DELLA-mediated transcription reprogramming remains unclear. Here by characterizing new missense alleles of an Arabidopsis DELLA, repressor of ga1-3(RGA), and co-immunoprecipitation assays, we show that RGA binds histone H2A via the PFYRE subdomain within its GRAS domain to form a TF–RGA–H2A complex at the target chromatin. Chromatin immunoprecipitation followed by sequencing analysis further shows that this activity is essential for RGA association with its target chromatin globally. Our results indicate that, although DELLAs are recruited to target promoters by binding to TFs via the LHR1 subdomain, DELLA–H2A interaction via the PFYRE subdomain is necessary to stabilize the TF–DELLA–H2A complex at the target chromatin. This study provides insights into the two distinct key modular functions in DELLA for its genome-wide transcription regulation in plants.

Musculoskeletal robot with high precision and robustness is a promising direction for the next generation of robots. However, motion learning and rapid generalization of complex musculoskeletal systems are still challenging. Therefore, inspired by the movement preparation mechanism of the motor cortex, this article proposes a motion learning framework based on the recurrent neural network (RNN) modulated by initial states. First, two RNNs are introduced as a preparation network and an execution network to generate initial states of the execution network and time-varying motor commands of movement, respectively. The preparation network is trained by a reward-modulated learning rule, and the execution network is fixed. With the modulation of initial states, initial states can be explicitly expressed as knowledge of movements. By dividing the preparation and execution of movements into two RNNs, the motion learning is accelerated to converge under the application of the node-perturbation method. Second, with the utilization of learned initial states, a rapid generalization method for new movement targets is proposed. Initial states of unlearned movements can be computed by searching for low-dimensional ones in latent space constructed by learned initial states and then transforming them into the whole neural space. The proposed framework is verified in simulation with a musculoskeletal model. The results indicate that the proposed motion learning framework can realize goal-oriented movements of the musculoskeletal system with high precision and significantly improve the generalization efficiency for new movements.