Search

The working conditions of loaders alternate between stages of full or empty loads, loading or unloading, and moving forward or backward, which complicates the vehicle's characteristic response. Based on the K-nearest neighbor (KNN) algorithm and a principal component analysis (PCA) method, stages recognition algorithm under the V-type working conditions of a loader was studied. First, the collected transmission signals were noise-reduced and filtered. Second, the PCA was used to reduce the dimensions of the data. Finally, the working condition samples were established from the data obtained, which were later trained and classified using the KNN algorithm. Compared with the neural network algorithm, the accuracy of the optimized KNN algorithm reaches 99.4%, and its running time is 2.1s. The algorithm described in this paper guarantees a high accuracy under recognition of the loader conditions in a short running time. It can be extended to an intelligent identification using big data and artificial intelligence control of the construction machinery.

In recent years, image deblurring has been widely investigated. In order to solve this ill-posed problem, a variety of prior models have been proposed successively. The two-tone prior has been successfully applied to text images deblurring and achieved significant results. However, the natural image with rich color clusters does not meet the two-tone prior, which requires only two color clusters in the image. In this paper, a local two-tone prior is proposed for images with complex color clusters, which decomposes the image with complex color clusters into patches with simple color clusters. We also find that the process of image blurring is a weighted average of pixel values, which will lead to an increase of intermediate pixel values. Therefore, a new measure of the dynamic range of the image is proposed, which indicates the difference of the color cluster using the average value of the color cluster, and it is rigorously proved in mathematics. Besides, we analyze the effectiveness of the proposed prior in image deblurring in detail. Experimental results on the widely used datasets show that the proposed method performs favorably against the state-of-the-art algorithms, both qualitatively and quantitatively.

In this paper, we rethink the problem of scene reconstruction from an embodied agent's perspective: While the classic view focuses on the reconstruction accuracy, our new perspective emphasizes the underlying functions and constraints such that the reconstructed scenes provide \em{actionable} information for simulating \em{interactions} with agents. Here, we address this challenging problem by reconstructing an interactive scene using RGB-D data stream, which captures (i) the semantics and geometry of objects and layouts by a 3D volumetric panoptic mapping module, and (ii) object affordance and contextual relations by reasoning over physical common sense among objects, organized by a graph-based scene representation. Crucially, this reconstructed scene replaces the object meshes in the dense panoptic map with part-based articulated CAD models for finer-grained robot interactions. In the experiments, we demonstrate that (i) our panoptic mapping module outperforms previous state-of-the-art methods, (ii) a high-performant physical reasoning procedure that matches, aligns, and replaces objects' meshes with best-fitted CAD models, and (iii) reconstructed scenes are physically plausible and naturally afford actionable interactions; without any manual labeling, they are seamlessly imported to ROS-based simulators and virtual environments for complex robot task executions., Comment: ICRA 2021 paper. Project: https://sites.google.com/view/icra2021-reconstruction

In recent years, many researchers have started to construct Graph Neural Networks (GNNs) to deal with graph classification task. Those GNNs can fit into a framework named Message Passing Neural Networks (MPNNs), which consists of two phases: a Message Passing phase used for updating node embeddings and a Readout phase. In Readout phase, node embeddings are aggregated to extract graph feature used for classification. However, the above operation may obscure the effect of the node embedding of each node on graph classification. Therefore, a node voting based graph classification model is proposed in this paper, called Node Voting net (NVnet). Similar to the MPNNs, NVnet also contains the Message Passing phase. The main differences between NVnet and MPNNs are: 1, A decoder for graph reconstruction is added to NVnet to make node embeddings contain graph structure information as much as possible; 2, In NVnet, the Readout phase is replaced by a new phase called Node Voting phase. In this new phase, an attention layer based on the gate mechanism is constructed to help each node to observe the node embeddings of other nodes in the graph, and each node predicts the class of the graph from its own perspective. The above process is called node voting. After voting, the results of all nodes are aggregated to get the final graph classification result. In addition, considering that aggregation operation may also obscure the differences between node voting results, a regularization term is added to drive node voting results to reach group consensus. We evaluate the performance of NVnet on 4 benchmark datasets. The experimental results show that NVnet performs well on graph classification task.

Based on the combination of the conventional rotor field-oriented control and the mechanical energy conversion theory, an optimal current torque angle compensation control strategy for interior permanent-magnet (IPM) brushless dc motors (BLDCMs) is proposed to eliminate the machine losses and suppress the commutation torque ripple simultaneously. First, with the proper assignment of the machine energy, the torque component of the stator current can be controlled directly so that the optimal motor current can be injected into the motor windings. Then, considering that some vectors related to the output torque of BLDCMs cannot be decoupled, the validity of the proposed strategy is proved and analyzed in detail. Moreover, the non-linear variation of the motor parameters during operation cannot be neglected; thus, some simplified pre-experiments are implemented to construct the complete curves of the parameters. Experimental results verify the feasibility of the proposed strategy.

Recent advances achieved in the field of reinforcement learning have led AI algorithms capable of beating world champions in some perfect information games like Chess and Go. However, the AI approach to imperfect information games (such as Poker) is much more difficult because the complexities in estimating hidden information and behaviors of opponents may become extremely challenging. Since Markov Decision Process (MDP) is the underlying mathematical model of reinforcement learning with perfect information games, Partially Observable Markov Decision Process (POMDP) deserves research attention for studying the games with imperfect information. In this paper, we study a 16-cards Rhode Island Hold’em poker game and present a POMDP model to formulate this imperfect information extensive game. Based on the POMDP model, we use Bayesian approach to estimate the opponent’s hand and transform the original problem to several perfect information games. Furthermore, to handle the challenge of explosively huge storage space and computation burdens, we develop a Monte Carlo optimization algorithm to estimate the action values of the POMDP model. Finally, we conduct numerical experiments in the Rhode Island Hold’em poker game to demonstrate the effectiveness of our approach.

This paper proposes a novel hybrid PWM mode for braking control of the brushless dc motor in the electric vehicle drive applications. To simplify the hardware circuit topology and reconstruct the real-time motor currents, a single dc-link current sensor sampling method is employed. However, in small duty ratio applications, the current signals cannot be sampled accurately under the conventional unipolar PWM mode due to the limitation of amplifier bandwidth and analog to digital converter sampling time. Therefore, a hybrid PWM mode based on unipolar-bipolar switching is used to obtain the current signals in a wide duty ratio range. The method proposed in this paper combines the advantages of unipolar and bipolar PWM, and realizes the real-time current control of brushless dc motor in the braking process. Simulation and experimental results verify the feasibility of the proposed control mode.

We present the summary of experimental results on multiple generations of multiwavelength selective crossbar switches designed and fabricated in silicon photonics platform as a part of AIM Photonics. We report CW measurements of 8 × 4 and 4 × 4 switches. The latest generation of the switch improved worst case path losses by 75 % as compared with the first. We also demonstrate error free transmission in the presense of incoherenct crosstalk with multiple crosstalk sources with 40 Gbps non return to zero signal and 111 Gbps four-level pulse amplitude modulation (PAM-4) transmission with hard decision forward error correction limit.

Design for trust approaches can protect an MPSoC system from hardware Trojan attack due to the high penetration of third-party intellectual property. However, this incurs significant design cost by purchasing IP cores from various IP vendors, and the IP vendors providing particular IP are always limited, making these approaches unable to be performed in practice. This paper treats IP vendor as constraint, and tasks are scheduled with a minimized security constraint violations, furthermore, the area of MPSoC is also optimized during scheduling. Experimental results demonstrate the effectiveness of our proposed algorithm, by reducing 0.37% security constraint violations.

We design and develop a new shared Augmented Reality (AR) workspace for Human-Robot Interaction (HRI), which establishes a bi-directional communication between human agents and robots. In a prototype system, the shared AR workspace enables a shared perception, so that a physical robot not only perceives the virtual elements in its own view but also infers the utility of the human agent--the cost needed to perceive and interact in AR--by sensing the human agent's gaze and pose. Such a new HRI design also affords a shared manipulation, wherein the physical robot can control and alter virtual objects in AR as an active agent; crucially, a robot can proactively interact with human agents, instead of purely passively executing received commands. In experiments, we design a resource collection game that qualitatively demonstrates how a robot perceives, processes, and manipulates in AR and quantitatively evaluates the efficacy of HRI using the shared AR workspace. We further discuss how the system can potentially benefit future HRI studies that are otherwise challenging., Comment: Accepted to IROS 2020

Most Graph Convolutional Networks (GCNs) used for graph classification task can fit into the Message Passing Neural Networks (MPNNs) framework. However, traditional MPNNs don't consider global information in the message passing phase in which the node embeddings are updated. In this paper, we propose a new model called Cross Message Passing Graph Neural Network (CMPGNN). The new model consists of two message passing phases, of which one is the local message passing phase used for node embeddings updating and another is the global message passing phase used for graph feature updating. Several convolutional layers are stacked together in the local message passing phase, in which different convolutional layers are used to update the node embeddings at different time steps. Each convolutional layer updates node embeddings not only according to the outputs calculated at the previous convolutional layer but also to the graph feature calculated at the previous time step. A readout layer shared by all time steps is used in the global message passing phase. At each time step, after the node embeddings are updated, the readout layer aggregates the embeddings of all nodes by using a global gated network, and feeds the aggregation result into a Gated Recurrent Unit (GRU) to update the graph feature.The above two message passing phases are executed alternatively. After all time steps, the graph feature obtained by the readout layer is fed into a MultiLayer perception for graph classification. We evaluate the performance of CMPGNN on 6 graph classification datasets. Experimental results show that compared with other 10 baselines, CMPGNN achieves the highest accuracy on 4 of the 6 benchmark datasets.

Federated learning (FL), as a privacy-preserving machine learning framework, draws growing attention in both industry and academia. It obtains a jointly accurate model by distributing training tasks into data owners and aggregating their model updates. However, FL faces new security problems, as it losses direct control to training processes. One fundamental demand is to ensure whether participants execute training tasks as intended.In this paper, we propose TrustFL, a practical scheme that leverages Trusted Execution Environments (TEEs) to build assurance of participants’ training executions with high confidence. Specifically, we use TEE to randomly check a small fraction of all training processes for tunable levels of assurance, while all computations are executed on the co-located faster yet insecure processor (e.g., GPU) for efficiency. To prevent various cheating behaviors like only processing TEE-requested computations or uploading old results, we devise a commitment-based method with specific data selection. We prototype TrustFL using GPU and SGX and evaluate its performance. The results show that TrustFL achieves one/two orders of magnitude speedups compared with naive training with SGX, when assuring correct training with a confidence level of 99%.

Aiming at the problem that the signal data of mining machinery is often subject to noise interference, this study selects a threshold wavelet denoising method based on the actual working conditions of mining machinery. Taking the front axle signal of the iron ore powder operation by the loader as an example, the threshold wavelet denoising method is used to perform Db9-4 decomposition. After comparison with Butterworth denoising and wavelet transform denoising methods, it is concluded that Noise will cause signal drift. The signal-to-noise ratio of threshold wavelet denoising is 14.32 and the mean square error is 51.91, which is similar to the value of wavelet transform method, but it is more suitable for mining machinery. The research results can also be used in other fields such as construction machinery and agricultural machinery.

We present a congestion-aware routing solution for indoor evacuation, which produces real-time individual-customized evacuation routes among multiple destinations while keeping tracks of all evacuees' locations. A population density map, obtained on-the-fly by aggregating locations of evacuees from user-end Augmented Reality (AR) devices, is used to model the congestion distribution inside a building. To efficiently search the evacuation route among all destinations, a variant of A* algorithm is devised to obtain the optimal solution in a single pass. In a series of simulated studies, we show that the proposed algorithm is more computationally optimized compared to classic path planning algorithms; it generates a more time-efficient evacuation route for each individual that minimizes the overall congestion. A complete system using AR devices is implemented for a pilot study in real-world environments, demonstrating the efficacy of the proposed approach., Accepted to ICRA 2020

In this work, we investigate the data retention of programmable linear random-access memory (PLRAM), where the geometry and structure of memory cells have been modified to decouple the program/erase operation and gate oxide. Since PLRAM utilizes the bidirectional Fowler-Nordheim tunneling during program and erase operations through the sidewall tunneling oxide, the gate oxide does not receive the stress from the excessive program/erase cycles, resulting in a significant improvement of the data retention. Furthermore, it is confirmed that PLRAM can guarantee the stable 7-bit accuracy under high temperatures (>200°C), while a standard multi-level flash memory cell shows a significant degradation of multi-bit functionality due to the data loss and it ends up with a 4-bit accuracy in this study.

Visualization tools like t-SNE and UMAP give insight into the high-dimensional structure of datasets. When there are related datasets (such as the high-dimensional representations of image data created by two different Deep Learning architectures), roughly aligning those visualizations helps to highlight both the similarities and differences. In this paper we propose a method to align multiple low dimensional UMAP visualizations by adding an alignment term to the UMAP loss function. We provide an automated procedure to find a weight for this term that encourages the alignment but only minimally changes the fidelity of the underlying embedding.

In recent years, scene text recognition has achieved significant improvement and various state-of-the-art recognition approaches have been proposed. This paper focused on recognizing text in natural photos of equipment nameplates, which has wide applications in industrial automations. This task only receives little attentions in previous works. The challenge of this problem comes from multi-orientation, curved, noisy and blurry text patches in equipment nameplates. To address this problem, we propose a deep model for text recognition in multi-oriented nameplates, namely, Orientation Robust Scene Text Recognition (ORSTR). Specifically, our model employs a rectification module to transform curved, distorted or multi-orientation text to near-horizontal text with a carefully designed rectification module. Once the near-horizontal text has been generated, recognition network will output the predictions of text patches. Our scene text recognition model achieves 90 . 8% recognition accuracy on equipment nameplate dataset which outperforms previous scene text recognition model (CRNN) about 0 . 8%. Several extensive experiments have been conducted to verify the effectiveness of our model.

Intelligent devices based on Internet of Things technology has more and more applications in the power grid. However, the development of the integration of primary equipment and secondary equipment causes the secondary intelligent equipment to be vulnerable to electromagnetic interference generated by the primary equipment. In this paper, electromagnetic interference generated during the distribution switch closing the capacitive load was measured and the frequency band distribution of electric field interference and magnetic field interference was studied by wavelet analysis. By comparing the frequency bands of common communication signals used in smart grid and the interference signal, it found that the interference signal may interfere with the normal communication work in the smart grid.

De-identification enables to protect privacy of data from different attacks. At present, the specific de-identification standards and privacy quantification methods are given in many models, such as K-anonymity model and differential privacy model. But, the K-anonymity model does not provide an effective method to prove its degree of privacy protection, and when the model parameters change, the degree of privacy protection cannot be quantified. And due to rigorous calculation method of the differential privacy model, which quantifies the degree of privacy protection based on a mathematical basis, it is difficult to be used widely in organizations and institutions. So, this paper proposes a de-identification model, which includes quantification solution of identifying the sensitivity of the personal information, de-identification approach for adaptively matching different standard de-identification methods to personal information with different sensitivities and de-identification effect detection function. The objective is to provide an automatic, efficient and widely used de-identification model, which is built by quantitatively analyzing degree of privacy protection based on conditional entropy. Finally, performance analysis results show that the model makes the trade-off between privacy and secure of data and the widespread use of data.

Cerenkov luminescence (CL) is a novel imaging modality that has recently been used in biomedical imaging field. However, there are plenty of limitations for the clinical transactions of Cerenkov luminescence imaging (CLI), due to the properties of CL. The newly developed europium-doped gadolinium oxide (Gd 2 O 3 :Eu) nanoparticles can interact with the CL and γ-radiation as well, generating a stronger red-shifted light, which is of great benefits for biomedical application and clinical transaction.

The cooling mechanism is crucial for nanoelectronics, which determines the response time limited for the devices. The cooling process after optical pump in graphene, which is a well-known material with ultrahigh carrier mobility, is therefore an urgent research topic. Ultrafast thermalization and relaxation process in a monolayer graphene film grown on a quartz substrate is investigated by femtosecond laser pump, terahertz probe technique. A typical photo induced negative terahertz conductivity is found, and the electron-phonon-impurity supercollision process is determined to about 2.8 picosecond at different pump photon energy and pump fluence. The competition between the impurity supercollision and the interband-intraband scattering in graphene could give new insight into the design of the application of graphene photonic devices.

This paper proposes a new reliability evaluation method for servo systems based on accelerated degradation test. The weaknesses of systems or components are found via Failure Mode and Effect Analysis (FMEA), by which the status parameters reflecting the system degradation could be determined. Then, an accelerated degradation test (ADT) platform design is introduced, where testing, loading, measuring equipment and control method are given in detail. Based on this platform, the ADT is implemented for a permanent magnet synchronous motor (PMSM) servo system. With the degradation data of the status parameters, the Arrhenius model is introduced to verify the MTBF of the system and evaluate its reliability.

This paper presents an incremental learning framework for mobile robots localizing the human sound source using a microphone array in a complex indoor environment consisting of multiple rooms. In contrast to conventional approaches that leverage direction-of-arrival (DOA) estimation, the framework allows a robot to accumulate training data and improve the performance of the prediction model over time using an incremental learning scheme. Specifically, we use implicit acoustic features obtained from an auto-encoder together with the geometry features from the map for training. A self-supervision process is developed such that the model ranks the priority of rooms to explore and assigns the ground truth label to the collected data, updating the learned model on-the-fly. The framework does not require pre-collected data and can be directly applied to real-world scenarios without any human supervisions or interventions. In experiments, we demonstrate that the prediction accuracy reaches 67% using about 20 training samples and eventually achieves 90% accuracy within 120 samples, surpassing prior classification-based methods with explicit GCC-PHAT features.

Cerenkov luminescence imaging (CLI) has been proved to be an effective imaging modality for biomedical application. However, the CLI applications were strongly hindered by the low intensity of the luminescence signals. In this study, the FDA approved sodium fluorescein was applied to enhance the CL signals. Furthermore, a novel endoscopic imaging system for Cerenkov radiation energy transfer imaging (CRET) with a clinical laparoscope was developed, which integrated air-liquid dual cooling solution to improve the imaging sensitivity. Hepatocellular carcinoma (HCC) murine models were used for in vivo experiments of CRET tumor detection and guided-resection. The results indicated the feasibility of the proposed CRET methods and the developed endoscopic imaging system. We believed the endoscopic CRET system is promising to increase the sensitivity of early stage tumor detection with great potential in clinical translation.

We investigate the linewidth enhancement factor of 1.3 μm quantum dot lasers epitaxially grown on silicon. Both the calculation and experiment show small (

We investigate, both experimentally and theoretically, the gain characteristics of modulation p-doped 1.3 μm quantum dot lasers epitaxially grown on silicon. The inhomogeneous broadening is extracted to be 10 meV. A p-doped quantum dot active region has been found to show lower transparency current and higher material gain.

Machine-type communications (MTC) is an important standardized technology in the fifth generation (5G) cellular networks, where a vast number of MTC devices communicate on the limited band. Due to the large number of MTC devices in MTC systems, the traditional OMA method seems very weak to provide enough connection for MTC devices. Thus a joint subcarriers and power allocation problem using power-domain non-orthogonal multiple access (NOMA) is formulated to allow multiple MTC devices to share the same subcarrier in this paper. And the problem is decomposed into two sub-problems so that it can be easy to solve. And two algorithms are proposed for these sub-problems to maximize the number of successfully connected MTC devices. The simulation results show that our proposed scheme can effectively improve the access opportunity of MTC devices compared to the existing schemes.

Cerenkov luminescence imaging (CLI) proved as a promising approach for molecular imaging of cancer. However, most of the conventional optical detection devices fail to acquire CL images since the Cerenkov luminescence intensity is extraordinarily weak. Recently, endoscopic CLI (ECLI) was proposed as a highly attractive approach for tumor detection and surgery. In this study, an ECLI system was constructed with a novel cooling solution, which can effectively reduce the system thermal noise and dramatically improve the imaging sensitivity. The developed ECLI system was further explored for imageguided surgery of hepatocellular carcinoma (HCC) on mice models.

In order to stabilize the engine speed, increase the dynamic performance, save the energy, this paper presents a speed sensing control method for an EFI engine. The best operating points of an engine were discussed according to the static characteristics of an engine. Based on the best operating points and the load, the engine speed could be controlled by a double-closed loop control method, which could obtain the feedback signal from the engine speed and the throttle position. The model of a 161kW diesel engine and the control method were built, and the simulation was calculated based on the best working conditions. The results showed that the method could maintain the actual speed of an engine close to the goal speed, and achieve the requirements of the maximum torque, maximum power and the most economical output using the best operation points.

With the advancement in energy harvesting in terms of wireless charging techniques, it provides a novel way to solve traditional energy constraint problems in Wireless Sensor Networks (WSNs). Renewable energy such as solar, wind, and geo-thermal energy is converted to energy-storage and further use via harvest-then-transmit strategy. This article introduces a two-layer sleep scheduling system in energy-harvesting WSNs with an aim to satisfy sustainable throughput by analysis and optimization of network performance. Evaluation results provide a typical demonstration of how to obtain the appropriate value of key parameters according to specific requirement.

Optical imaging techniques have been developed for localizing lymph nodes before surgical resection due to the non-invasion and high sensitivity. However, its attendant penetrability limitations and auto-fluorescence effect have greatly limited the spatial resolution and imaging precision. In this study, a novel technique radiopharmaceutical-excited fluorescence imaging (REFI) was adopted to image lymph nodes, which use gamma-ray and Cerenkov radiation from radioisotopes to excite lanthanide europium oxide (EO) nanophosphors and boost the light intensity. An effective adaptive-steepest-descent-projection onto convex sets (ASD-POCS) reconstruction algorithm and the anatomic structure information were used to three-dimensionally image lymph nodes in mice models. The results indicate that REFI can greatly boost the light intensity and accuracy of threedimensional imaging of lymph node with location deviation less than 1.03 mm.

It is important to predict consumers' location and motion in a large shopping mall to provide them better service. When a consumer passes regions of a shopping mall, his/her moving trace can be recorded for prediction. Existing approaches cannot be directly used to fulfill such task because handling the ordered region sequences is quite challenging. In this paper, we propose an improved Apriori algorithm called AprioriOS (Apriori for Ordered Sequences) to solve this problem. Using this method, association rules are mined out from ordered region sequences and then used to predict future locations of consumers. We can predict more than one regions that a consumer may pass in future. We also design an association rule querying method and a tree storage structure for location prediction. And we propose a motion prediction method based on 3-axis accelerometer or RFID to predict motions of consumers. Based on the proposed we develop an location and motion prediction system for shopping malls. Our simulation results show that the system is effective in terms of the accuracy of prediction.

Nanophotonic interconnects have been playing an increasingly important role in the datacom regime. Greater integration of silicon photonics demands modeling and simulation support for design validation, optimization and design space exploration. In this work, we develop compact models for a number of key photonic devices, which are extensively validated by the measurement data of a fabricated optical network-on-chip (ONoC). Implemented in SPICE-compatible Verilog-A, the models are used in circuit-level simulations of full optical links. The simulation results match well with the measurement data. Our model library and simulation approach enable the electro-optical (EO) co-simulation, allowing designers to include photonic devices in the whole system design space, and to co-optimize the transmitter, interconnect, and receiver jointly.

In view of the complex structure of large wind turbine system and the characteristics of the operation process variable, this paper puts forward a fault diagnosis method based on multivariate statistics. The mathematical model based on partial least squares (PLS) without the need of complex model. Through the original data PLS algorithm built the relationship between input and output variables and get the monitoring model. Computer simulation results show this method can effectively reduce the dimension of data and realize the fault diagnosis. Compared with PCA, PLS are more fully use the sample space information, improve the accuracy of fault diagnosis effectively.

A Ho:YAG rod amplifier is demonstrated for vortex beams with orbital angular momentum. Simulations show nice correlation with experimental data and potentially high gain. This could be used for complex modes and spatial division multiplexing.

This paper proposed a novel tunable current-mode square-wave converter with current-rectifying function. The converter has a concise and simple structure without using any passive devices, and all the active devices of the converter are grounded, which makes it convenient for IC application. The proposed converter can generate both positive and negative square-wave signals whose amplitude can be adjusted independently by the bias ports. Also it can produce positive half-wave, negative half-wave, positive full-wave, and negative full-wave rectifying current outputs. The PSPICE simulation results of the converter based on 0.25µm CMOS process are presented. The results show that the analysis method is valid and effective