Multi-Fault Diagnosis for Autonomous Underwater Vehicle Based on Fuzzy Weighted Support Vector Domain Description

This paper addresses the multi-fault diagnosis problem of thrusters and sensors for autonomous underwater vehicles (AUVs). Traditional support vector domain description (SVDD) has low classification accuracy in the process of AUV multi-fault pattern classification because of the effect of sample sparse density and the uneven distribution of samples, and so on. Thus, a fuzzy weighted support vector domain description (FWSVDD) method based on positive and negative class samples is proposed. In this method, the negative class sample is introduced during classifier training, and the local density and the class weight are introduced for each sample. To improve the multi-fault pattern classifier training speed and fault diagnosis accuracy of FWSVDD, a multi-fault mode classification method based on a hierarchical strategy is proposed. This method adds fault contain detection surface for each thruster and sensor to isolate fault components during fault diagnosis. By considering the problem of pattern classification for a fuzzy sample, which may be located in the overlapping area of hyper-spheres or may not belong to any hyper-sphere in the process of multi-fault classification based on FWSVDD, a relative distance judgment method is given. The effectiveness of the proposed multi-fault diagnosis approach is demonstrated through water tank experiments with an experimental AUV prototype.

Confidence support vector domain description

To accelerate the training of support vector domain description （SVDD）, confidence support vector domain description （CSVDD） is proposed based on the observation that the description boundary is determined by a small subset of training data called support vectors. Namely, the number of training samples in the userdefined sphere is calculated and taken as the confidence measure, according to which the training samples are ranked in ascending order. Those former ranked ones are selected as the boundary targets for the SVDD training. Simulations on UCI data demonstrate the effectiveness and superiority of CSVDD： the number of training targets and the training time are reduced without any loss of accuracy.

Moving Least Squares Interpolation Based A-Posteriori Error Technique in Finite Element Elastic Analysis

The performance of a-posteriori error methodology based on moving least squares(MLS)interpolation is explored in this paper by varying the finite element error recovery parameters,namely recovery points and field variable derivatives recovery.The MLS interpolation based recovery technique uses the weighted least squares method on top of the finite element method’s field variable derivatives solution to build a continuous field variable derivatives approximation.The boundary of the node support(mesh free patch of influenced nodes within a determined distance)is taken as circular,i.e.,circular support domain constructed using radial weights is considered.The field variable derivatives(stress and strains)are recovered at two kinds of points in the support domain,i.e.,Gauss points(super-convergent stress locations)and nodal points.The errors are computed as the difference between the stress from the finite element results and projected stress from the post-processed energy norm at both elemental and global levels.The benchmark numerical tests using quadrilateral and triangular meshes measure the finite element errors in strain and stress fields.The numerical examples showed the support domain-based recovery technique’s capabilities for effective and efficient error estimation in the finite element analysis of elastic problems.The MLS interpolation based recovery technique performs better for stress extraction at Gauss points with the quadrilateral discretization of the problem domain.It is also shown that the behavior of the MLS interpolation based a-posteriori error technique in stress extraction is comparable to classical Zienkiewicz-Zhu(ZZ)a-posteriori error technique.

Three-Dimensional Meshfree Analysis of Interlocking Concrete Blocks for Step Seawall Structure

This study adapts the flexible characteristic of meshfree method in analyzing three-dimensional(3D)complex geometry structures,which are the interlocking concrete blocks of step seawall.The elastostatic behavior of the block is analysed by solving the Galerkin weak form formulation over local support domain.The 3D moving least square(MLS)approximation is applied to build the interpolation functions of unknowns.The pre-defined number of nodes in an integration domain ranging from 10 to 60 nodes is also investigated for their effect on the studied results.The accuracy and efficiency of the studied method on 3D elastostatic responses are validated through the comparison with the solutions of standard finite element method(FEM)using linear shape functions on tetrahedral elements and the well-known commercial software,ANSYS.The results show that elastostatic responses of studied concrete block obtained by meshfree method converge faster and are more accurate than those of standard FEM.The studied meshfree method is effective in the analysis of static responses of complex geometry structures.The amount of discretised nodes within the integration domain used in building MLS shape functions should be in the range from 30 to 60 nodes and should not be less than 20 nodes.

Progressive transductive learning pattern classification via single sphere

In many machine learning problems, a large amount of data is available but only a few of them can be labeled easily. This provides a research branch to effectively combine unlabeled and labeled data to infer the labels of unlabeled ones, that is, to develop transductive learning. In this article, based on Pattern classification via single sphere （SSPC）, which seeks a hypersphere to separate data with the maximum separation ratio, a progressive transductive pattern classification method via single sphere （PTSSPC） is proposed to construct the classifier using both the labeled and unlabeled data. PTSSPC utilize the additional information of the unlabeled samples and obtain better classification performance than SSPC when insufficient labeled data information is available. Experiment results show the algorithm can yields better performance.