This paper presents a novel step kinematic calibration method for a 3 degree-of-freedom(DOF) planar parallel kinematic machine tool,based on the minimal linear combinations(MLCs) of error parameters.The method using m...This paper presents a novel step kinematic calibration method for a 3 degree-of-freedom(DOF) planar parallel kinematic machine tool,based on the minimal linear combinations(MLCs) of error parameters.The method using mapping of linear combinations of parameters in error transfer multi-parameters coupling system changes the modeling,identification and error compensation of geometric parameters in the general kinematic calibration into those of linear combinations of parameters.By using the four theorems of the MLCs,the sets of the MLCs that are respectively related to the relative precision and absolute precision are determined.All simple and feasible measurement methods in practice are given,and identification analysis of the set of the MLCs for each measurement is carried out.According to the identification analysis results,a step calibration including step measurement,step identification and step error compensation is determined by taking into account both measurement costs and observability.The experiment shows that the proposed method has the following merits:(1) the parameter errors that cannot influence precision are completely avoided;(2) it reflects the mapping of linear combinations of parameters more accurately and enhances the precision of identification;and(3) the method is robust,efficient and effective,so that the errors in position and orientation are kept at the same order of the measurement noise.Due to these merits,the present method is attractive for the 3-DOF planar parallel kinematic machine tool and can be also applied to other parallel kinematic machine tools with weakly nonlinear kinematics.展开更多
In this paper, a multi-view gait based human recognition system using the fusion of two kinds of features is proposed.We use cross wavelet transform to extract dynamic feature and bipartite graph model to extract stat...In this paper, a multi-view gait based human recognition system using the fusion of two kinds of features is proposed.We use cross wavelet transform to extract dynamic feature and bipartite graph model to extract static feature which are coefficients of quadrature mirror filter(QMF)-graph wavelet filter bank. Feature fusion is done after normalization. For normalization of features, min-max rule is used and mean-variance method is used to find weights for normalized features. Euclidean distance between each feature vector and center of the cluster which is obtained by k-means clustering is used as similarity measure in Bayesian framework. Experiments performed on widely used CASIA B gait database show that, the fusion of these two feature sets preserve discriminant information. We report 99.90 % average recognition rate.展开更多
基金the"863"High-Tech Program of China(Grant Nos.2006AA04Z204 and 2006AA04Z227)National Natural Science Foundation of China(Grant Nos.50775118and50605041)+1 种基金the"973"Basic Research Project of China(Grant Nos.2006CB705406 and 2007CB714000)Tsinghua Basic Research Foundation(Grant No.JC200701)
文摘This paper presents a novel step kinematic calibration method for a 3 degree-of-freedom(DOF) planar parallel kinematic machine tool,based on the minimal linear combinations(MLCs) of error parameters.The method using mapping of linear combinations of parameters in error transfer multi-parameters coupling system changes the modeling,identification and error compensation of geometric parameters in the general kinematic calibration into those of linear combinations of parameters.By using the four theorems of the MLCs,the sets of the MLCs that are respectively related to the relative precision and absolute precision are determined.All simple and feasible measurement methods in practice are given,and identification analysis of the set of the MLCs for each measurement is carried out.According to the identification analysis results,a step calibration including step measurement,step identification and step error compensation is determined by taking into account both measurement costs and observability.The experiment shows that the proposed method has the following merits:(1) the parameter errors that cannot influence precision are completely avoided;(2) it reflects the mapping of linear combinations of parameters more accurately and enhances the precision of identification;and(3) the method is robust,efficient and effective,so that the errors in position and orientation are kept at the same order of the measurement noise.Due to these merits,the present method is attractive for the 3-DOF planar parallel kinematic machine tool and can be also applied to other parallel kinematic machine tools with weakly nonlinear kinematics.
文摘In this paper, a multi-view gait based human recognition system using the fusion of two kinds of features is proposed.We use cross wavelet transform to extract dynamic feature and bipartite graph model to extract static feature which are coefficients of quadrature mirror filter(QMF)-graph wavelet filter bank. Feature fusion is done after normalization. For normalization of features, min-max rule is used and mean-variance method is used to find weights for normalized features. Euclidean distance between each feature vector and center of the cluster which is obtained by k-means clustering is used as similarity measure in Bayesian framework. Experiments performed on widely used CASIA B gait database show that, the fusion of these two feature sets preserve discriminant information. We report 99.90 % average recognition rate.
