Examenspresentation av Logi Sigurdarson

Abstract

The rise of autonomous lawnmowers equipped with real-time kinematic (RTK) GPS has given access to accurate absolute localization during operation. This has made it possible for mowers to follow predefined systematic paths and avoid relying on inefficient pseudo-random patterns. In this thesis, coverage path planning (CPP) algorithms applicable to RTK-GPS enabled autonomous lawnmowers are investigated. Specifically, exact cell decomposition CPP algorithms with the objective to minimize the coverage time for complex maps are studied.

Further, the split and merge boustrophedon cell decomposition algorithm is implemented and evaluated using two different optimization criteria. For the connecting transport paths between cells, a simple sub-optimal greedy algorithm is proposed. The evaluations are based on an estimated time consumption along with simulated results. The simulations are performed in a hybrid system simulation environment for a differential drive lawnmower using a path following pure pursuit controller adapted to follow the path with the mower cutting deck.

Welcome!

Logi and Martin