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

### Authors: Jaroslav Moravec

*Polibits,* Vol. 46, pp. 29-38, 2012.

Abstract: In the presented article the method for map building of unknown environment (SLAM) deduced from ICP algorithm and using point-to-point metric will be proposed. The polar-scan matching technology will be used for estimation of the robot location change between two scans in sequence estimate the correct robot’s pose. Since the map building is still fairly time-consuming, the algorithm of differential evolution (DE) has been integrated into the calculation. This efficient optimizer provides very good results in different types of small office environment (unstructured, structured). The new type of the presented algorithm for map building is purely based on simple geometric primitives – vectors and integrates the modern evolutionary algorithm – DE. The presented algorithm falls into the wider group of geometric map builders and is able to build-up a map of indoor, mostly office environment lacking moving objects.

Keywords: SLAM, robot localization, evolutionary robotics, differential evolution, L1-norm

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

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