Feng Lv

Mobile robots' modular navigation controller using spiking neural networks

Autonomous navigation plays an important role in mobile robots. Artificial neural networks (ANNs) have been successfully used in nonlinear systems whose models are difficult to build. However, the third generation neural networks - Spiking neural networks (SNNs) - contain features that are more attractive than those of traditional neural networks (NNs). Because SNNs convey both temporal and spatial information, they are more suitable for mobile robots@? controller design. In this paper, a modular navigation controller based on promising spiking neural networks for mobile robots is presented. The proposed behavior-based target-approaching navigation controller, in which the reactive architecture is used, is composed of three sub-controllers: the obstacle-avoidance SNN controller, the wall-following SNN controller and the goal-approaching controller. The proposed modular navigation controller does not require accurate mathematical models of the environment, and is suitable to unknown and unstructured environments. Simulation results show that the proposed transition conditions for sub-controllers are feasible. The navigation controller can control the mobile robot to reach a target successfully while avoiding obstacles and following the wall to get rid of the deadlock caused by local minimum.

Coal mass estimation of the coal mill based on two-step multi-sensor fusion

In the fossil power plant, it is rather difficult to measure the coal mass of the coal mill exactly, in order to make the coal mill work on the optimal active state, multi-sensor are used to fuse multiple signal, and the qualitative estimation of the coal mass is gotten from the algorithm. The neural network has the ability of self-organize, self-learn, and disposing the nonlinear problems, strong fault tolerant and robustness, D-S evidential theory can solve the uncertainty problem, but the evident is hard to get. Two-step fusion method combined the merit of the neural network and the evidential theory, the neural network is on the first step, the second step uses the normalization result as the evident. When this algorithm is simulated on the computer, the result proves that the method can estimate the coal mass qualitatively, according to the historical record of coal mill.

A target-reaching controller for mobile robots using spiking neural networks

Autonomous navigation plays important role in mobile robots. In this paper, a navigation controller based on spiking neural networks (SNNs) for mobile robots is presented. The proposed target-reaching navigation controller, in which the reactive architecture is used, is composed of three sub-controllers: the obstacle-avoidance controller and the wall-following controller using spiking neural networks (SNNs), and the goal-reaching controller. The experimental results show that the navigation controller can control the mobile robot to reach the target successfully while avoiding the obstacle and following the wall to get rid of the deadlock caused by local minimum. The proposed navigation controller does not require accurate mathematical models of the environment, and is suitable to unknown and unstructured environment.

Feedforward Compensation Based the Study of PID Controller

This paper describes the conventional PID controller works, and feedforward control based on analysis of the PID controller works,and establish a feed-forward control of the PID controller the basic model, using MATLAB software and conventional PID control system based on the former the PID feedback control system for the simulation; and further analysis the feed-forward control based on the PID controller simulation waveforms, summed up the feed-forward control based on the superiority of the PID controller to achieve effective tracking of the input signal; this simulation system, the theoretical study of advanced PID, program design and failure analysis are of great help.

A Method of Multi-Classifier Combination Based on Dempster-Shafer Evidence Theory and the Application in the Fault Diagnosis

A problem is aroused in multi-classifier system that normally each of the classifiers is considered equally important in evidences’ combination, which gone against with the knowledge that different classifier has various performance due to diversity of classifiers. Therefore, how to determine the weights of individual classifier in order to get more accurate results becomes a question need to be solved. An optimal weight learning method is presented in this paper. First, the training samples are respectively input into the multi-classifier system based on Dempster-Shafer theory in order to obtain the output vector. Then the error is calculated by means of figuring up the distance between the output vector and class vector of corresponding training sample, and the objective function is defined as mean-square error of all the training samples. The optimal weight vector is obtained by means of minimizing the objective function. Finally, new samples are classified according to the optimal weight vector. The effectiveness of this method is illustrated by the UCI standard data set and electric actuator fault diagnostic experiment.

Electromagnetic Interference and Electromagnetic Compatibility Test Technology

With the extensive application of electronic equipment, issues from electromagnetic compatibility caused by electromagnetic interference will directly affect the normal operation of the system or equipment. In this paper, the electromagnetic interference on the basis of research and analysis the two typical interference source-- the inverter and microwave ovens, to discuss the mechanism and performance of electromagnetic and microwave radiation, conduction, and in according to the electromagnetic characteristics of different structures, electromagnetic interference and electromagnetic compatibility problems, it proposes some different approaches, through experimental testing, data analysis and comparison to verify the validity of the method, provide a reliable technical basis for resolving the interference of the actual products.

The Electric Actuator's Fault Diagnosis Based on Information Fusion

The method based on information fusion fault diagnosis is put forward and the two-step fusion diagnosis model of neural network and D-S evidence theory is given against the complexity and multiplicity when the fault of control system happens; the fault detection laboratory platform is erected based on the information fusion through the analysis of the electric actuators principle and typical faults, and on this basis, the fault of the electric actuator is diagnosed by the experiment and computer simulation. This method overcomes the uncertainty of the neural network fault diagnosis and improves the accuracy of system diagnosis, and the experiment result verified the methods validity.

Method of Fault Diagnosis Based on SVDD-SVM Classifier

Aiming at the problem of incomplete fault data samples, a fault diagnosis method based on Support vector data description and Support vector machine (SVDD-SVM) is presented. First, the data description model is build based on the normal data samples and known fault data samples, and SVM model is built based on known fault data samples. Then the test data samples are tackled by the data description model to reject or accept. The specific categories of accepted samples are diagnosed by the SVM model and the rejected samples are unknown fault types. Tests show that this method can efficiently solve the fault diagnosis problem of incomplete fault samples.

Synthesis and Photocatalytic Activity of Ag-Doped BiVO4

Ag-doped BiVO4 semiconductor photocatalysts were synthesized via the one-step hydrothermal method. The microstructure and morphology of catalysts were characterized by using X-ray diffraction, Scanning electron microscopy, and Energy dispersive X-ray detector (EDS) and photocatalytic activities of BiVO4 catalysts with and without Ag doping were evaluated by degrading methylene blue (MB) under visible-light irradiation. UV-Vis absorption spectra were measured to evaulate the photocatalytic activity of the as-synthesized catalysts. The results suggested that Ag-doped BiVO4 with larger rod-like particle size but better crystallnity has the stronger UV absorption. In comparison with pure BiVO4, degradation rate of MB was increased about 18% in Ag-doped BiVO4 with the Ag+ dopant concentration of 15 mol%.

The Fault Recognition of Motor Based on the Fusion of Neural Network and D-S Evidence Theory

In order to reflect the motor from various aspects and realize the motor system state failure mode automatic identification and accurate diagnosis, neural network combined with the D-S evidence theory to form the motor fault diagnosis system. In data fusion level, fault characteristic is classified; and then the fault feature is extracted by the BP neural network and the local fault of the motor is diagnosed, as a result, the independent evidence is obtained; at last the D-S evidence theory fusion algorithm is used on the evidence to achieve the fault of the motor accurate diagnosis.Broken test proved that the diagnosis system improves the motor of the fault diagnosis of accuracy, and can meet the needs of real-time diagnosis. The diagnostic test proved that the diagnosis system improves the accuracy of motor fault diagnosis, and can satisfy the diagnosis in real-time.