Aiming to improve quality indicator prediction for the tobacco drying process in a cigarette factory, this study proposes an adaptive-weighted multi-level feature evidence and redundancy-penalty framework (F-XMF). To alleviate the limitations of isolated feature evaluation in complex industrial data, F-XMF jointly considers the direct contribution of individual features, the contribution of feature interactions, and the redundant dependency among candidate features. First, eXtreme Gradient Boosting (XGBoost) is used to estimate the importance of individual features and obtain an initial main-effect feature subset. Meanwhile, an Attentional Factorization Machine (AFM) is employed to identify high-contribution second-order interactions and derive interaction-related evidence. On this basis, conditional mutual information (CMI) is introduced to characterize redundant dependency among candidate features, and a Differential Redundancy Penalty Mechanism is further developed to impose differentiated penalties on highly overlapping features while preserving informative variables with meaningful interaction utility. Finally, a validation-based adaptive weight fusion strategy is used to integrate the main-effect subset and the interaction-related subset, yielding a compact and informative final feature subset. Experimental results on real tobacco-processing data show that F-XMF consistently reduces Mean Absolute Error (MAE) in ablation comparisons and achieves superior performance against multiple representative baseline methods. Mechanism validation further demonstrates that the proposed framework can effectively refine the candidate subset through interaction enhancement and redundancy suppression, showing good interpretability and practical value for outlet moisture content prediction in the tobacco drying process.

Robot skill reconfiguration often disrupts system continuity in dynamic flexible production systems, particularly in mixed-model assembly. To address this challenge, this article proposes behavior tree and action primitive (BTAP) enabled low-code/no-code program method for robot task execution, overcoming the limitations of traditional static programming for robot operation. First, the proposed method establishes a genes-inspired modular recombination mechanism, in which action primitives are structured like DNA-sequence components. Ontology Web Language is utilized to formally characterize the relational logic among action primitives, which helps form the primitive library. This library supports genes-inspired operations, such as crossover and mutation, for dynamic skill generation. Next, planning domain definition language-to-behavior tree (BT) conversion algorithm is developed to automatically transform high-level instructions into executable BTs, achieving automated mapping from task logic to robotic sequential behaviors. Then, parametric meta-learning approach is proposed, which integrates adaptive parameter space contraction with Bayesian optimization. This method narrows the search space through historical task feature retrieval and optimizes action parameters using Gaussian process surrogate models coupled with Expected Quantile Improvement function. Finally, the proposed BTAP method is verified on a multirobot motor assembly line. Results demonstrate rapid convergence to a 95.46% success rate within 15 training cycles for the TwoArmPegInsertion, while consistently maintaining contact forces and torque fluctuations well below the safety thresholds.

ABSTRACT Robotic manipulators have become essential in demolition tasks involving hazardous, confined, or structurally unstable environments, where precise and responsive motion control is critical for safety and efficiency. Solving the inverse kinematics (IK) of demolition robot manipulators poses considerable challenges due to the inherent strong nonlinearity and coupling in their kinematic equations, along with potential singularities that can undermine real‐time computational efficiency and system robustness. To address these issues, this study introduces an enhanced Newton–Raphson (NR) approach, specifically optimized for inverse kinematic analysis of 6‐DOF manipulators equipped with a spherical wrist architecture frequently adopted in demolition robotic arms. The proposed method strategically partitions the manipulator into two three‐DOF segments and constructs NR iterative equations by utilizing the reconnecting constraints between these kinematic substructures. After solving a subset of joint variables iteratively, the remaining two joint angles are obtained analytically. The experimental results show that: Compared with the traditional Newton–Raphson method, the improved method has a faster convergence speed, higher accuracy, and better robustness, especially when dealing with singular points at the wrist. This makes the improved method applicable to the inverse kinematics control of dismantling robotic arms in complex environments.

With the development of new-generated IT technologies and the launch of a series of industrial Internet of things platforms, service-oriented manufacturing is an inevitable trend of manufacturing industry. Therefore, the platform-based manufacturing service collaboration (MSC) becomes a recognized answer to the complex and personalized manufacturing demands. However, the changes in both supply and demand of the platform in its operation process are usually unpredictable. To cope with the scheduling problem on the platform-based MSC with the dynamic uncertainties of both supply and demand, an adaptive scheduling mechanism is explored in this article. In which, the real-time system state evaluation method considering supply and demand are designed, and a supply-demand aware rescheduling trigger judgement is proposed. Experimental results show the effectiveness and adaptiveness of the proposed mechanism, which also provides a reference for other MSC scheduling problems towards different dynamic situations.

End-face defect control (EF-DC) of the hot-rolled coil is crucial to the quality management. The core of EF-DC is to identify defects accurately, predict defects in advance, and improve production in time to prevent similar defects. The current research works on EF-DC are mainly at the defect recognition stage, which merely rely on the operating data in physical space. However, different types of defects exist in coils with different reasons, therefore it is difficult to accurately control defects in time only through the hot rolling operating data. In order to solve the above problems, first an EF-DC framework based on digital twin with cloud-edge collaboration is designed in this article. The computing tasks are reasonably allocated through cloud-edge collaboration to realize the timeliness of control. Second, the virtual model of hot-rolled coil is constructed from four aspects: geometry, physics, behavior, and rules. Through the analysis of the defect mechanism, the abnormal behavior events are obtained, and the mapping relationship between the defect and the behavior event is established to realize the defect traceability and control. Finally, the feasibility of the proposed method is verified by taking the edge scratch defect as an example.

The Industrial Internet platform-based manufacturing service collaboration has made it possible for decentralized manufacturing enterprises to cooperate on a broad scale. The selection conflict problem may arise when many providers choose the same task at the same time, especially when there are more providers than consumers. At this moment, how to choose the appropriate manufacturing services, which should both satisfy the users’ requirements and enhance the participation of the providers, is of utmost importance. To increase the number of collaboration chances, the functional and quantitative manufacturing service collaboration is carried out simultaneously in this article. The variable utility models are used to represent the satisfaction levels of users while taking into account the bilateral coupling between providers and consumers and the unilateral irrationality of the provider. Finally, it is advised to choose providers with a short-term preference in situations where there is a strict time limit based on the findings and analysis of the manufacturing service collaboration optimization. At the same time, we can find that the providers, who get lower utilities when they focus on the current, are more likely to get greater utilities than others when they concentrate on long-term gains.

The recent decade has witnessed an upsurge in the demands of intelligent and simplified Internet of Things (IoT) networks that provide ultra-low-power communication for numerous miniaturized devices. Although the research community has paid great attention to wireless protocol designs for these networks, researchers are handicapped by the lack of an energy-efficient software-defined radio (SDR) platform for fast implementation and experimental evaluation. Current SDRs perform well in battery-equipped systems, but fail to support miniaturized IoT devices with stringent hardware and power constraints. This paper takes the first step toward designing an ultra-low-power SDR that satisfies the ultra-low-power or even battery-free requirements of intelligent and simplified IoT networks. To achieve this goal, the core technique is the effective integration of µW-level backscatter in our SDR to sidestep power-hungry active radio frequency chains. We carefully develop a novel circuit design for efficient energy harvesting and power control, and devise a competent solution for eliminating the harmonic and mirror frequencies caused by backscatter hardware. We evaluate the proposed SDR using different modulation schemes, and it achieves a high data rate of 100 kb/s with power consumption less than 200 µW in the active mode and as low as 10 µW in the sleep mode. We also conduct a case study of railway inspection using our platform, achieving 1 kb/s battery-free data delivery to the monitoring unmanned aerial vehicle at a distance of 50 m in a real-world environment, and provide two case studies on smart factories and logistic distribution to explore the application of our platform.

In this paper, a multimodal magnetic resonance imaging (MRI) and heterogeneous metadata (including age, gender) dataset containing263 patients was established. Based on this dataset, a new multimodal deep neural network (KiNet) was proposed, aiming to effectively predict the Ki67 index in gliomas in a non-invasive way by fusing multimodal MRI features and metadata. We adopted a five-fold crossvalidation approach to verify the performance of the network. KiNet achieved results with an AUC of 0.79 and a kappa coefficient of 0.47. The proposed approach’s outperformance indicated the feasibility of predicting the Ki67 index in gliomas in a non-invasive way.

The different orbit design and launching conditions of Soil Moisture and Ocean Salinity (SMOS, ESA) and Soil Moisture Active Passive (SMAP, NASA) result in different passing time over any point on the ground. The time lag between the two satellites is thought to be one of the reasons to induce uncertainties in soil moisture data comparison and validation. This letter calculates the temporal difference between SMOS and SMAP at first; it is found that their mismatch mainly concentrates within a period of 30–90 min. During such time lag, the change in surface soil moisture (5 cm) and other meteorological variables is analyzed on the basis of the U.S. Climate Reference Network (USCRN) high-frequency (5-min) field observations and Murrumbidgee Soil Moisture Monitoring Network (MSMMN)in situmeasurements (20-min). This letter found that in most cases, air temperature, wind, and relative humidity present a moderate change of about 10%–20%, while solar radiation shows very strong variation from tens to hundreds (%). Soil moisture and soil temperature are always stable, the value of soil moisture at the two time points when SMOS and SMAP pass overhead are almost the same, and the averaged minimum and maximum fluctuations of soil moisture are only 0.004/0.003 and 0.007/0.01$\text{m}^{3}/\text{m}^{3}$, respectively, which are far less than the nominal accuracy of satellites (0.04$\text{m}^{3}/\text{m}^{3})$and probably unrecognizable. Soil moisture experiences a natural fading of very small magnitude during the time intervals of satellites, the temporal mismatch may not induce external uncertainties in soil moisture data comparison and validation, and it is safe to conclude that the impact is negligible.

Abstract Under the background of economic globalization, enterprises face with more severe and uncertain environmental dynamism, and its lean production system redesign strategy is more critical. Firstly, given the dynamic environment that enterprises are facing, the redesign of the lean production system based on the environmental dynamism is proposed. Secondly, environmental dynamism is divided into two dimensions: market dynamics and technology dynamics, which is calculated by the objective method. Thirdly, by establishing the redesigned model of the lean production system based on environmental dynamism, the relationship between environmental dynamism and lean production level is analyzed. Fourth, the data of 251 listed companies from different industries in 2014 to 2017 were analyzed to verify the specific impact of environmental dynamism on the enterprise's lean production level. It was found that the relationship between environmental dynamism and the enterprise lean production level is presented as “U,” “S,” and other more complex relationships. At last, some suggestions are put forward to optimal the redesign of lean production systems under different environmental dynamism.