Estimation of the surface model parameters and analysis of spatial uncertainties

In this paper, we present a method to estimate surface models based on a point cloud taken from the surface of the workobject. The models we generate are used in a robot based workcell for localization of the workobject. The approach to the problem is that we can obtain the point cloud from workobject CAD model or the point cloud can be generated based on actual measurements from the surface of the workobject carried out using a robot and a range sensor. In addition to presenting the different surface forms, we estimate the uncertainties of the surface model parameters and consider the effect of uncertainties in model parameters in workobject localization. The estimation of the surface parameters and workobject localization is carried out using Bayesian-form estimation method and all the noises are considered when modelling the uncertainties of the system. The uncertainty analysis is based on observing the error covariance matrix of the estimated parameters.