Search

Pickup vehicle scheduling in steel logistics parks is an important problem for determining the outbound efficiency of steel products. In a steel logistics park, each yard contains different types of steel products, which provides flexible yard selection for each pickup operation. In this case, the yard allocation and the loading sequence for each vehicle must be considered simultaneously in pickup vehicle scheduling, which greatly increases the scheduling complexity. To overcome this challenge, in this paper, we propose a pickup vehicle scheduling problem with mixed steel storage (PVSP-MSS) to optimize the makespan of pickup vehicles and the makespan of steel logistics parks simultaneously. The optimization problem is formulated as a multi-objective mixed-integer linear programming model, and an enhanced algorithm based on SPEA2 (ESPEA) is proposed to solve the problem with a high efficiency. In the ESPEA, a cooperative initialization strategy is firstly proposed to initialize the vehicle pickup sequence for each yard. Then, an insertion decoding method is designed to improve the scheduling efficiency, utilizing the idle time of a yard. Furthermore, local search technology based on critical paths is proposed for the ESPEA to improve the solution quality. Experiments are executed based on data collected from a real steel logistics park. The results confirm that the ESPEA can significantly reduce both the makespan of each pickup vehicle and the makespan of the steel logistics park.

Wood surface broken defects seriously damage the structure of wooden products, these defects have to be detected and eliminated. However, current defect detection methods based on machine vision have difficulty distinguishing the interference, similar to the broken defects, such as stains and mineral lines, and can result in frequent false detections. To address this issue, a multi-source data fusion network based on U-Net is proposed for wood broken defect detection, combining image and depth data, to suppress the interference and achieve complete segmentation of the defects. To efficiently extract various semantic information of defects, an improved ResNet34 is designed to, respectively, generate multi-level features of the image and depth data, in which the depthwise separable convolution (DSC) and dilated convolution (DC) are introduced to decrease the computational expense and feature redundancy. To take full advantages of two types of data, an adaptive interacting fusion module (AIF) is designed to adaptively integrate them, thereby generating accurate feature representation of the broken defects. The experiments demonstrate that the multi-source data fusion network can effectively improve the detection accuracy of wood broken defects and reduce the false detections of interference, such as stains and mineral lines.

In biaxial contouring control applications, the inherent structural flexibility of machines can lead to position discrepancies between the manipulator and actuator, and thus deteriorate the manufacturing performance, especially when the controller is designed without available end-effector side feedback. In this work, we focus on the end-effector contouring control problem for industrial machines with position-dependent flexibility to improve the contouring performance while eliminating the effect of mechanical vibration. A model for the widely used cantilever beam machine is developed to describe the dynamics of the end-effector by capturing the rotation and coupled dynamics between axes. The proposed model is validated through experiment and systematically reduced to switched linear time-invariant models for controller design. By adopting the extended state observer, the proposed control architecture decouples the dynamics between the X and Y-axis and simplifies the controller design process. The model predictive control method is utilised for improving the contouring performance while reducing mechanical vibration. The efficacy of the proposed control framework is demonstrated and validated on the designed high-fidelity model. Performance comparisons between the proposed approach with benchmark controllers are presented.

This study presents two dynamic models, namely recurrent neural network and long short-term memory (LSTM) models, for predicting PM2.5 concentrations in Taiwan by using PM2.5 time series obtained at air quality monitoring stations and weather information obtained at neighboring weather stations. The proposed models can efficiently predict PM2.5 by incorporating a learned memory structure with a forgetting gate. To evaluate the predictive performance of the proposed models, large-scale databases established by Taiwan's Environmental Protection Administration, and Central Weather Bureau were used; these databases include hourly data from 77 air quality monitoring stations and 580 weather stations over a 1-year period. The results demonstrated that the proposed models outperformed three traditional machine learning methods (gradient boosting, support vector machine, and classification and regression tree models) by 27.12% and 33.69% on average in terms of the coefficient of determination and root mean square error, respectively. A geographical divergence analysis was conducted to compare predictive performance in different regions. The results revealed that the most significant improvement in predictive performance was achieved in central Taiwan. The seasonal and pollution effect on predictive performance were reduced by the LSTM and the source distribution of PM2.5 emission in Taiwan was also analyzed.

The mobile industrial human-machine interaction plays an important role in the industrial internet of things, since the engineers can use a mobile device to interact with machines that greatly improves the efficiency and safety. Nevertheless, connecting to a specific machine becomes a non-trivial problem due to the massive machines in the network that make the connection list too long to identify the target machine. Some solutions such as QR code scanning and proximity estimation have been proposed to solve this problem. However, they have limited performance in scalability and accuracy correspondingly, and thus cannot satisfy the requirements in most applications. Observing the fact that the engineers generally interact with the machines in their line-of-sight, we propose the LightCon scheme which adopts proximity estimation to estimate the machines in the line-of-sight, and controls the display module of machines to show different visible symbols (colors or numbers). To connect with a specific machine, the engineers just need to select the corresponding symbol on the mobile device. Therefore, they do not have to remember the trivial address of each machine. Furthermore, the symbol assignment algorithm is designed to reduce the complexity of manual symbol selection, and its performance is analyzed theoretically. The performance of LightCon is evaluated in the testbed, and the experimental results prove that LightCon is a promising solution to simplify line-of-sight connections with low complexity.

Multi-hop data collection in wireless sensor networks (WSNs) is a challenge issue due to the limited energy resource and transmission range of wireless sensors. The hybrid clustering and routing (HCR) strategy has provided an effective solution, which can generate a connected and efficient cluster-based topology for multi-hop data collection in WSNs. However, it suffers from imbalanced energy consumption, which results in the poor performance of the network lifetime. In this paper, we evaluate the energy consumption of HCR and discover an important result: the imbalanced energy consumption generally appears in gradient k = 1, i.e., the nodes that can communicate with the sink directly. Based on this observation, we propose a new protocol called HCR-1, which includes the adaptive relay selection and tunable cost functions to balance the energy consumption. The guideline of setting the parameters in HCR-1 is provided based on simulations. The analytical and numerical results prove that, with minor modification of the topology in gradient k = 1, the HCR-1 protocol effectively balances the energy consumption and prolongs the network lifetime.

Color classification of wooden boards is helpful to improve the appearance of wooden furniture that is spliced from multiple wooden boards. Due to the similarity of colors among wooden boards, manual color classification is inaccurate and unstable. Thus, supervised learning algorithms can hardly be used in this scenario. Moreover, wooden boards are long, and their images have a high resolution, which may lead to the growth of computational complexity. To overcome these challenges, in this paper, we propose a new mechanism for color classification of wooden boards based on machine vision. The image of the wooden board is preprocessed to subtract irrelevant colors, and the feature vector is extracted based on 3D color histogram to reduce the computational complexity. In the offline clustering, the feature vector sets are partitioned into different clusters through the K-means algorithm. Then, the clustering result can be used in the online classification to classify the new wood image. Furthermore, to process the abnormal images of wooden boards, we propose an improved algorithm with centroid improvement and image filtering. The experimental results verify the effectiveness of the proposed mechanism.

Due to the limited energy of wireless sensor nodes, the energy efficiency of data collection is a key issue in wireless sensor network (WSN). Dynamic clustering is a scalable and energy efficient solution for data collection in WSN. However, it suffers from the imbalanced energy consumption in the intercluster communication. To solve this problem, in this paper, we propose a distributed and adaptive routing protocol for cluster-based wireless sensor networks (DARC). In DARC, an adaptive energy threshold is proposed for the cluster head to control its intercluster routing mode, and a tunable cost function is designed for relay selection. Simulation results prove that DARC effectively balances the energy consumption in the intercluster communication and hence improves the energy efficiency and prolongs the network lifetime.

In wireless sensor and actor networks, when an event is detected, the sensor node needs to transmit an event report to inform the actor. Since the actor moves in the network to execute missions, its location is always unavailable to the sensor nodes. A popular solution is the search strategy that can forward the data to a node without its location information. However, most existing works have not considered the mobility of the node, and thus generate significant energy consumption or transmission delay. In this paper, we propose the trail-based search (TS) strategy that takes advantage of actor’s mobility to improve the search efficiency. The main idea of TS is that, when the actor moves in the network, it can leave its trail composed of continuous footprints. The search packet with the event report is transmitted in the network to search the actor or its footprints. Once an effective footprint is discovered, the packet will be forwarded along the trail until it is received by the actor. Moreover, we derive the condition to guarantee the trail connectivity, and propose the redundancy reduction scheme based on TS (TS-R) to reduce nontrivial transmission redundancy that is generated by the trail. The theoretical and numerical analysis is provided to prove the efficiency of TS. Compared with the well-known expanding ring search (ERS), TS significantly reduces the energy consumption and search delay.