Search

In the actual project of data sharing and data governance, the problems of data heterogeneity and low data quality have not been solved. Data quality detection based on syntax rules can effectively find data quality problems, but it had not improved the data quality, especially in a variety of heterogeneous data source environments. This paper designs a data governance model based on data traceability, and we can get data feedback and revision through this model. The proposed method considers the different ownership of data and may form a closed-loop data service chain including effective data validation, data tracking, and data revision, data release. The analysis of the case shows that the method is effective to improve the data quality and meet the requirements of data security also.

In this paper, an evaluation model based on entropy method, cloud model and evidence theory is established to solve the problem that the evaluation result of the elevator operation status is too subjective. By analyzing the primary features associated with the operation status of the elevator, a layered evaluation index system for elevator condition assessment is built. Firstly, the entropy method is used to acquire the constant weights of each indicator, and the variable weight model based on balance function is introduced, in which the constant weights is modified to get the variable weights according to the deviation of actual data and normal value. Then the cloud theory is applied to achieve the conversion between quantitative index and qualitative concept, in which the membership degree is obtained so that index is mapped to each status. Lastly, based on the evidence theory, the index information is fused, and the evaluation results are obtained by the principle of maximum membership degree and reliability criterion. By applying the model we proposed, the example analysis shows that, the model is correct and effective, which could reflect the operation status of the elevator accurately.

A sliding mode control approach is proposed to synchronize a class of delayed neural networks in a finite time, where the mismatched parameters and neuron activation functions are taken into account. In the controller design, a sliding mode manifold is directly defined by the synchronization error, which greatly reduces the synchronization time. Its concise design process and its ability to synchronize the delayed neural networks in a small finite time are two advantages of the sliding mode controller. Two numerical examples are given to illustrate the effectiveness of the developed approach.

In this paper, we use the recently developed second order sliding mode control approach to design a robust flight controller for a small quadrotor unmanned aerial vehicle (UAV). Due to the underactuated property of a quadrotor UAV, the flight controller can make the quadrotor move three positions and the yaw angle to their desired values and stabilize the pitch and roll angles. The flight controller is designed to provide robust position and attitude tracking control of a quadrotor UAV with respect to model parameter uncertainties. Simulations show that the designed controller robustly stabilizes a quadrotor UAV.

In UPS, Battery mode and electric mode usually use a common boost converter to translate different input voltage into a same DC bus voltage. In this paper, the mathematic model of the boost converter based on PWM three-terminal switch device is derived. To obtain good dynamic performance during the input voltage changes, a directly digital dual-loop controller design method based on Matlab/Simulink is proposed and the universal calculation formulae are derived. Experimental results of a 3.3 kW boost converter used in 380V UPS show that the controller has good dynamic performance.

In this paper, we present an emotion recognition system using the stacked generalization ensemble neural network for special human affective state in the speech signal. 450 short emotional sentences with different contents from 3 speakers were collected as experiment materials. The features relevant with energy, speech rate, pitch and formant are extracted from speech signals. Stacked Generalization Ensemble Neural Networks are used as the classifier for 5 emotions including anger, calmness, happiness, sadness and boredom. First, compared with the traditional BP network or wavelet neural network, the results of experiments show that the Stacked Generalization Ensemble Neural Network has faster convergence speed and higher recognition rate. Second, after discussing the advantage and disadvantage between different ensemble Neural Networks, suitable decision will be made for Robot Pet.