My primary focus is in autonomous mapping and localization or commonly called SLAM (simultaneous localization and mapping): a key problem is for a robot to explore its environment and use the information gathered by its sensors to jointly produce a map of its environment, together with an estimate of its position. SLAM is a big interest for the community and it's a propedeutic problem to provide robots with full autonomy. Algorithmics and Probabilistic concept have key role in solving the problem.

A common used method to solve SLAM is to use raw data from range finder sensors (lasers, sonars, infrared, etc) through a method called Scan Matching. Range finder sensors return to the robot a set of points corresponding to obstacle. The points are used to build the map but they can also be used to correct the dead reckoning displacement estimate (odometry) error. To apply a scan matching algorithm it is necessary to have at least two scans taken in different positions and moments. The scans must have sufficient area of overlap for the match to make sense. Scan Matching methods are various and cover many areas: signal processing, image registration and probabilistic methods.

My other interests are connected to general automation, navigation, underwater technologies, probabilistic reasoning and AI.

2D EKF+Scan Matching for robotic localization, a Toolbox (Download) NEW VERSION!

The Scan Matching 2D Toolbox is a collection of 2D ScanMatching functions for range finder sensors. It contains a set of 12 algorithms, published in literature. A robot and sensor simulator is also implemented to model a robotic platform moving and sensing a random generated world (either unstructured and structured). Also, a simple EKF pose based SLAM is implemented to handle the SLAM. Each pose is corrected against the very previous observation. It is also possible not to use EKF, so the correction will just linear and no stochastic processing will be done. The user will be able to work and modify an open source MATLAB code for its own purposes.The code is commented but not extensively and you can refer to the corresponding paper "Scan Matching 2D applied to robot localization: An extensive comparison of the state of the art". A brief manual is included into the package. For any inqiury please contact me.

• Bachelor Degree in Computer Science at the University of Bologna, Italy (2009).

• Master Degree in Computer Science at the University of Bologna, Italy (2009).

S. Zandara P. Ridao A. Mallios.
Scan Matching 2D applied to robot localization: An extensive comparison of the state of the art. Institute of Informatics and Applications, Universitat de Girona.
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). September 25-30, 2011 San Francisco, California.

S. Zandara P. Ridao D. Ribas R. Campos R. Garcia
Kornati Bathymetry Survey Data-Set for Navigation and Mapping. Institute of Informatics and Applications, Universitat de Girona.
The 19th Mediterranean Conference on Control and Automation June 23-25 2011

Zandara, S. and Nicholson, A.
Square Root Unscented Particle Filtering for Grid Mapping.
AI 2009: Advances in Artificial Intelligence