OnTheTipsofOne' sToes: Self-localizatio n ina Dynamic Environment

unige.it Abstract -Thispaperdeals withtheself-localizatio n ofan autonomous mobilerobotwithin a dynamic environment. Theconcept ofdynamic environments is extended along thevertical direction: this paper assumes thataboveourheads exists afree zonewherealsohighly dynamicenvironments becomealmost"static". The localization process, ifmanaged intothefree zone, does notsuffer fromfeature occlusions andotherrelated issues. Thefeature matching process iscarried outusing a typical Split & Mergealgorithm whichsupplies an Extended KalmanFilter withobservations usedto perform position tracking.