The normal probabilistic occupancy grids approach requires huge memory and complex computation of data
association to map in unknown large-scale environments. Therefore, this paper proposes an adaptive size of occupancy grids
method based on quadtree for mapping. By utilizing the theory of quadtree, the size of occupancy grids is automatically
adjusted by the obstacle density of different zones in maps. Simulation experiment results show that, the mapping algorithm
reduces the memory of map data and the computation of data association because of the reduced number of grids, which
improves mobile robot’s performance of mapping in unknown large-scale environments.