Abstract | ||
---|---|---|
The method for range scan segmentation described in a previous paper titled "Improving area center robot navigation using a novel range scan segmentation method", by some authors of present paper in 2011, uses an extended Kalman filter (EKF) for estimating straight segments equations in polar coordinates. That method, equivalent to a non-linear weighted least squares method, is extended in present work for estimating circle segments equations in polar coordinates, but only to the first stages of the full method, including acquisition of initial estimation of circles centers and radii to be used as initial values for EKF an the sequential iterative estimation of circle equations using EKF. The new problem of variance matrices being almost ill-conditioned for circle equations is addressed with a Tikhonov regularization, or ridge regression. Also a correction of initial values for estimations is proposed to solve the added problems found in circle equations estimation. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1007/978-3-319-59773-7_55 | BIOMEDICAL APPLICATIONS BASED ON NATURAL AND ARTIFICIAL COMPUTING, PT II |
Keywords | Field | DocType |
Kalman segmentation,Circular contour,Circle extraction,Robot localization,Tikhonov regularization | Tikhonov regularization,Computer vision,Extended Kalman filter,Regression,Computer science,Segmentation,Matrix (mathematics),Algorithm,Polar coordinate system,Radius,Artificial intelligence,Robot | Conference |
Volume | ISSN | Citations |
10338 | 0302-9743 | 0 |
PageRank | References | Authors |
0.34 | 11 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
José Manuel Cuadra Troncoso | 1 | 53 | 10.58 |
M. A. Muñoz-Bañón | 2 | 0 | 0.34 |
F. de la Paz-López | 3 | 1 | 1.36 |
José Ramón Álvarez Sánchez | 4 | 10 | 4.33 |