Map Building of Unknown Environment Using L1-norm, Point-to-Point Metric and Evolutionary Computation

摘要

"In the present paper, a method for building a map ofan unknown environment (SLAM) derived from the ICPalgorithm using point-to-point metric is proposed. The polarscanmatching technology is used for estimation of the robotlocation change between two scans in sequence estimate thecorrect position of the robot. Since map building is fairly timeconsuming,the algorithm of differential evolution (DE) is used inthe calculation. This efficient optimizer provides very goodresults in different types of small office environment(unstructured and structured). The new type of an algorithm formap building is based purely on simple geometric primitives—vectors and integrates the modern evolutionary algorithm—DE.The presented algorithm falls into the wider group of geometricmap builders and is able to build a map of indoor, mostly office,environment without moving objects."