SLAM (simultaneous localization and mapping) in agricultural environments

In the last 25 years, a lot of work was made for improving the localization and mapping abilities of a robot, with the aim of replacing costly GPS antennas. However, in agricultural environments (where data acquired by range sensors are very noisy) SLAM (simultaneous localizaton and mapping) algorithms need yet to be improved, mainly because of the data acquired by the sensors, corresponding to canopies, terrain, fruits, stems, etc. This project is focused on implementing existing SLAM algorithms, able to handle point clouds, and to test their efficiency and performance. To this end, we are going to use existing databases and no field experiments are required.

Expected skills: python/Matlab, statistics, ROS.