Back to the program.
Scan Matching, Covariance Estimation and SLAM: Models and Solutions for the scanSLAM Algorithm
Mauro Brenna
Politecnico di Milano
mauro.brenna@mail.polimi.it

Abstract

Simultaneous Localization and Mapping (SLAM) concerns the problem of correctly estimating the location of the robot (its pose) and the map of the environment given the noisy data acquired by robot sensors. ScanSLAM is a particular SLAM algorithm which uses raw laser sensor scans and scan matching algorithm in an Extended Kalman Filter (EKF) framework. The aim of this work is to extend the basic scanSLAM model using the state of the art scan matching methods, with interest in the themes of multiple lasers and self calibration. Results obtained from Ground truth data will be also shown.

Keywords

Simultaneous Localization and Mapping, mobile robotics, kalman filtering, scan matching, scanSLAM, Extended Kalman Filter

References

  1. Mauro Brenna. (2008) Scan Matching, Covariance Estimation and SLAM: Models and Solutions for the scanSLAM Algorithm. Master Thesis. Politecnico di Milano.
    See http://airlab.elet.polimi.it/index.php/Lurch