An efficient implementation of a particle filter for localization using compass data

This paper presents a localization algorithm for a small mobile platform. Taking advantage from modern technology, Saetta, a low cost mobile robot, has been built from scratch. Due to the limited processing capabilities, some ad hoc solutions have been used: the lack of processing resources has been compensated by an efficient implementation of the estimator and by the use of compass measures which ease the computational load. The results show how a careful design allows the implementation of sophisticated algorithms also on small platforms.