A Study of Mobile Robot Algorithms with Sycamore

Abstract

In this thesis we considered a simulation platform for mobile robots algorithms: Sycamore. We implemented several new features for Sycamore and we tested them while studying three different algorithms to achieve gathering by robots with limited visibility: a deterministic well known algorithm, a simple new probabilistic algorithm, and a combination of the two. The deterministic algorithm is known to achieve exact gathering when there are no faults; we tested it for the first time in presence of crashes and observed interesting and unexpected behaviors. We then performed extensive simulations with the probabilistic solution to identify the cause of an unexpected high rate of success, the simulations help us identify the relation between the rate of success and the initial configuration. Finally, we combined the two designing a hybrid solution. This work resulted in improvements of Sycamore, which can now be better employed to study mobile robots algorithms, as well as in empirical observations leading to new theoretical problems to be investigated.