Search

The high-dimensional and imbalanced data classification appears in many actual applications, but there are many problems encountered in practical operation. To overcome the disadvantages of kernel-based multilayer extreme learning machines (ML-KELM), enhanced kernel-based multilayer fuzzy weighted extreme learning machines (EML-KFWELM) has been proposed in this study. First, ML-KELM ignores imbalance learning, so we embed weighted strategy into ML-KELM to enhance the classification performance of the minority class. Meanwhile, we propose fuzzy membership to eliminate classification error caused by outlier and noise samples. Furthermore, we develop an enhanced grey wolf optimization (EGWO) method to perform the parameters optimization and improve the generalization performance of ML-KELM. In addition, the advantage of EML-KFWELM is that representation learning and classification can be integrated into a single learning process. Finally, computational comparisons with other state-of-the-art methods are performed on various real-world and gene expression data. Experimental results demonstrate that the proposed EML-KFWELM has good stability and can efficiently deal with the high-dimensional and imbalanced data.

Multi-label twin support vector machine (MLTSVM), being an effective multi-label classifier based on twin support vector machine (TSVM), has been widely studied and applied due to its excellent classification performance. However, there are some disadvantages in classical MLTSVM: (a) MLTSVM needs to solve a series of quadratic programming problems (QPPs), which makes its learning speed lower. (b) For multi-label learning problems, it is very difficult to obtain all labels of all samples. In fact, the datasets that we can obtain only contain a small amount of labeled samples and a large amount of partially labeled and unlabeled samples. However, MLTSVM can only use expensive labeled samples and ignore cheap unlabeled and partially labeled samples. For the drawbacks, we propose a novel semi-supervised multi-label least squares twin support vector machine, called SS-MLLSTSVM. Firstly, to speed up solving, SS-MLLSTSVM introduces the least squares idea into each sub-classifier of MLTSVM, which makes each sub-classifier only need to solve a system of linear equations, instead of one QPP. Secondly, SS-MLLSTSVM can make full use of the geometric information in unlabeled and partially labeled samples by introducing manifold regularization term into each sub-classifier. The experimental results on the benchmark datasets show that, compared with the existing multi-label classification algorithms, our SS-MLLSTSVM has better classification performance.

There are two problems with fingerprint matching indoor positioning technology based on Bluetooth Low Energy; in the process of collecting Bluetooth signals, due to the shadow effect and multipath effect, the signal is prone to violent fluctuations, abnormal values and special values, and Kalman filtering is commonly used only for Linear systems, for one-dimensional non-linear Bluetooth signals, are prone to data dispersion, causing positioning errors. The commonly used k-means clustering algorithm in the region division stage requires the number of clusters and the value of k to be specified in advance, which is more sensitive to the initial value and cannot determine the number of reference points according to the positioning environment. Based on the above problems, the extended Kalman filter (EKF) optimization database is first proposed. The nonlinear system adopts local Taylor expansion to form a linear system, and the problem of Kalman filter's sensitivity to nonlinear data is solved by the regression iteration method. The multi-dimensional similarity propagation clustering algorithm (MAP) is proposed, and the Pearson correlation coefficient dimension based on RSSI is proposed to redefine the similarity matrix elements. The results show that the fingerprint database of the proposed combination algorithm is optimized, the reference point area is more detailed, the algorithm adaptability is higher, and the average positioning accuracy is improved.

In the past, people control robots by remote-control or keyboard which is quite difficult and not efficient. This method needs to occupy people's hands and is totally unfriendly to the disabled. So we developed a spherical robot remote control system based on silent speech recognition technology. After the surface EMG signals are collected by the equipment, preprocessed and extracted features, the recognition results which are produced by the random forest are sent to the spherical robot to make it move. By this way, the spherical robot can move according to people's thought without hands' movements. In this paper, the reliability of the system is verified by experiments, which prove it has practical application value.

Scratchpad memory (SPM) is widely utilized in many embedded systems as a software- controlled on-chip memory to replace the traditional cache. New non-volatile memory (NVM) has emerged as a promising candidate to replace SRAM in SPM, due to its significant benefits, such as low-power consumption and high performance. In particular, several representative NVMs, such as PCM, ReRAM, and STT-RAM can build multiple-level cells (MLC) to achieve even higher density. Nevertheless, this triggers off higher energy overhead and longer access latency compared with its single-level cell (SLC) counterpart. To address this issue, this paper first proposes a specific SPM with morphable NVM, in which the memory cell can be dynamically programmed to the MLC mode or SLC mode. Considering the benefits of high-density MLC and low-energy SLC, a simple and novel optimization technique, named theory of thermal expansion and contraction, is presented to minimize the energy consumption and access latency in embedded systems. The basic idea is to dynamically adjust the size configure of SLC/MLC in SPM according to the different workloads of program and allocate the optimal storage medium for each data. Therefore, an integer linear programming formulation is first built to produce an optimal SLC/MLC SPM partition and data allocation. In addition, a corresponding approximation algorithm is proposed to achieve near-optimal results in polynomial time. Finally, the experimental results show that the proposed technique can effectively improve the system performance and reduce the energy consumption.

Decoding speaking intention from the cervical and facial muscle activity enables speech recognition independent of acoustic signals, thus allowing silent communication between human and computing devices. This research aims to design and optimize the convolutional network classifier for sEMG mime speech signals. The muscles involved in the vocalization and timbre modulation are found according to the anatomical map. Six-channel signals have been collected from the corresponding position using the sEMG signal acquisition device. The original signals are subjected to pre-processing, including noise reduction, active segment detection and interpolation, to form training and testing sets. Convolutional network models are applied to Figure out the influence of structural parameters, such as convolution kernel size, the number of convolution kernels and the depth of network, on the recognition accuracy. Based on repeated trials, the optimal convolution network is provided, which provide above 80% accuracy rate.

Twin-hypersphere support vector machine (THSVM) is a binary classification method that uses two hyperspheres to depict two classes, which makes the THSVM be more reasonable for many engineering problems. The two hyperspheres can be constructed by solving two smaller-scale quadratic programming problems (QPPs), which makes the THSVM be more efficient. However THSVM treats equally all features of one sample, in real life, the importance of different features of one sample for classification is always different. For the disadvantage, we introduce feature weights into THSVM to avoid classification results being dominated by trivial relevant features, reformulate the mathematical model of THSVM, propose a novel feature weighted twin-hypersphere support vector machine (FWTHSVM) and apply information gain to evaluate the weight of each feature of one sample. The experimental results show, compared with THSVM, FWTHSVM not only ensures training time, but also has better generalization performance.

Along with the occurrence and continuous development of semantic sensor network, more and more sensor data is released on network in the format of RDF. However, SSN ontology is lack of conceptual description about locations. The location information in SSN ontology is only a textual word which is weak at describing location information of sensors. The issue makes sensor data be used inconveniently. In order to solve the issue, we propose a method to strengthen the ability of geographical description of SSN ontology by GeoNames. The method can enrich the geographical information of sensor data. Moreover, users can query sensor data based on geographical information, such as alternate names, coordinates. Additionally, we also design a prototype to help users more conveniently complete the work of querying and using the sensor data.

The research object is the Single-phase photovoltaic (pv) grid inverter based on the household, and adopt the DC/DC/AC two level conversion circuit topology. This paper studies the maximum power point tracking of photovoltaic power generation units and booster function for the Former level conversion circuit, and studies the DC bus voltage stability and interconnection of the same frequency and phase current with the main grid for the DC/AC inverter circuit. The Maximum power point tracking control strategy uses the improved perturbation method of observation, and the inverter circuit control strategy with the improved PID control technology. The feasibility of the control scheme is verified by the simulation, the simulation results show that the grid inverter working performance is stable and reliable, and the control system meets the design requirements.

In order to boost the tourism circular economy development further, in this paper, the renewable energy systems used in tourism circular economy are studied. In this work, a scenic spot is selected as the analyzed object. Here, two emphases are the applications and configurations of renewable energy system by studying the networks of power systems in every functional department of a scenic spot. Finally, a simulation experiment is conducted to analyze the characteristics of renewable energy system operating in tourism circular economy and results show the feasibility and availability of the renewable energy system used in tourism circular economy.

Recovery is a promising approach for building intrusion tolerant systems that remove or minimize the effects of malicious attacks. In this paper, a Virtualization-based Recovery approach for Intrusion Tolerance systems (VRIT) is proposed. We present the architecture of the VRIT, construct the state transition model of the VRIT, and analyze quantitatively the security of the VRIT system. We characterize the intrusion tolerance provided by the system through several measures defined on the model. Finally we have some simulation experiments to evaluate both the availability and the mean time to security failure for the VRIT system.

Several access modes and access methods, combined with new power management schemes and frame structures are defined in the latest IEEE standard 802.15.6. To improve power efficiency of body area network, data compression was studied in this paper. Moreover, an analytical model was developed to evaluate the power efficiency of BAN system. Simulation results show that good power efficiency can be achieved by data compression. It also proved that higher data rate can help to improve energy efficiency. When its compression factor is bigger than 2 and its power consumption does not exceed 40% of the transceiver, better energy efficiency always can be guaranteed by processing unit insertion.

Human body channel, a candidate of PHY layer adopted by the latest IEEE standards (IEEE 802.15.6), has been investigated for applications with on-body communication. The study covers frequencies range from 100Hz to 1GHz and channel lengths from 65mm to 1.6m. After analyzing the measurement results, a suitable frequency range for body area network from 10 kHz to 10MHz was found. Channel model and corresponding modeling methodology are presented. Compared with the traditional models, the presented methodology allows an accurate model to be systematically constructed for either a specific subject or a group of subjects with similar physical forms and characteristics. The accuracy of prediction was proved by experimental validations (prediction error around 0.1 dB in frequency range from 10kHz to 10MHz).

This paper proposes a novel approach of cross-layer resource allocation based on Quality of Experience (QoE) for wireless video transmission. We model and formulate the cross-layer resource allocation problem in accordance with the relationship between transmission bit rate and QoE. Our objective is to ameliorate the system QoE level while guaranteeing the fairness for users. The video quality fluctuation that affects the system QoE is considered as well. The proposed algorithm designs an adaptive resource allocation approach depending on limited network resources. We use a downlink LTE system as a simulation example of a high speed downlink access system. The simulation results demonstrate that the proposed method considerably and agreeably improves system QoE more than other algorithms.

The deterioration of natural environment and exhaustion of conventional energy resources necessitate the use of clean renewable energy. Environmentally-friendly and renewable, wind energy is the world's fastest growing energy. Variable speed constant frequency generators(VSCFGs) are widely used in power systems with wind farms. Regarding valve-point effect of conventional generators, dynamic optimal power flow (DOPF) in VSCFGs wind farms integrated power systems is a typical complex multi-constrained non-convex non-linear programming problem. DOPF mode is established. The use of chaotic sequences in the traditional particle swarm optimization (PSO) algorithm can be helpful to escape more easily from local minima than can be done through PSO. This paper provides chaotic particle swarm optimization (CPSO) approaches for computing the established DOPF model. Simulation result showed that the proposed method is capable of yielding higher quality solutions.

As fabrication process technology continues to advance, mask set costs have become prohibitively expensive. Structured ASICs can offer price and performance between ASICs and FPGAs. They are attractive for mid-volume production and offer good intellectual property security. In this paper, a structured ASIC methodology, where 2 metal- and 1 via-mask are customised, is described. The CAD tools are fully compatible with conventional ASIC design flows and a comparison of area and delay performance with ASICs and FPGAs is given. A prototype structured ASIC implementing an LED-backlit LCD controller was fabricated in a 0.13µm CMOS process. It was verified and power consumption compared with an ASIC design.

For designing a wireless network, field test is mostly preferred for researchers to test and evaluate the system performance. However, it is expensive for field constructions and thus it is not affordable for most researchers. Software simulation is another option. However, the results by simulation may be undesirably biased away the true. This paper proposes a new method to trade off between resource consumption and test veracity, which constructs a hardware-in-loop simulation platform by integrating theoretic computation, computer simulation and hardware-based test. This platform can be used to verify wireless networking technologies, while resource is reduced greatly at the same time.

In this paper, a QoS-guaranteed cross-layer resource allocation algorithm for multiclass services in downlink LTE system is proposed, which takes EXP rule, channel quality variance, real-time services and non-real-time services and minimum transmission rate into account. The key features of the proposed algorithm are that all the users and resource blocks will be allocated step by step respectively. Numerical results demonstrate that the proposed algorithm effectively guarantees the user QoS requirement for multiclass services. At the same time, it mainly maintains the throughput and fairness performances in a high level.

A Structured Application-specific Integrated Circuit (SASIC) is a programmable fabric in which a small set of masks are customized for a particular application, serving to reduce the associated non-recurring engineering cost (NRE). In this paper we describe the implementation of a SASIC logic cell which is programmable via a single metal layer. A SASIC fabric prototype is fabricated and all implemented functions are verified on silicon. Experimental measurement verifies correct operation of our SASIC with a clock frequency of over 250MHz.

An effective heat transfer device such as heat pipe is used to reduce thermal stress and thermal gradients of inner material by transferring heat from one location to another with very large heat fluxes. Radiative transfer is important and complex at elevated temperature of the ceramic material. These temperatures in the two aeroheating regions and working fluid inside heat pipe are unknown and depend on the heat exchange process. The heating conditions for the internal and external surfaces of every aeroheating region can be quite different. In this paper, the control volume method was combined with the Monte Carlo method to calculate the coupled heat transfer in this kind of composite wall. The effects of aeroheating flux to the outer surface and the area of radiative cooling on the heat transfer were analyzed. The results show that the thermal environment of the outer surface and the area of radiating cooling are the key parameters for the heat transfer in the structure.

This paper presents a development platform to generate passive UHF RFID tags compatible with EPCTM C1G2 protocol. The platform is based on one RFID tag IC prototype implemented in a 130nm CMOS technology. Some features of the tag IC include that a voltage multiplier implemented by diodeconnected NMOSFETs, a voltage regulator composed of a selfbiased mutual compensation without large resistors for areasaving, and an energy-aware irregular clock structure with clock gating used in the baseband processor. To further reduce the chip area for low production cost and meet different specifications, the platform design manages to achieve different floorplans according to different user-defined requirements, mainly focusing on the communicating distances. This approach simplifies the design flow, and, most importantly, shortens time-to-market.

Working set selection is an important step in SMO for training support vector machine (SVM). Faced with C-SVM, Fan Rong-En proposed a method, which used second-order approximate information to select working set, and indicated that it had higher rate than the maximal violating pair. Based on this method, faced with weighted support vector machine (W-SVM) this paper proposes a training algorithm, which uses second-order approximate information to select working set. At the same time, two data preprocessing methods are proposed for existing weight knowledge and non-existing weight knowledge. Experiments indicate that the methods not only ensure precision, but also improve training rate highly.