Modeling of uncertain environment based on multi-scale grid method

In the research domain of intelligent driving technology, modeling of the dynamic and uncertain environment is very important for path planning. The accuracy and computational efficiency of path planning depend on environment modeling to a great extent. Inspired by selective attention mechanism of human, this paper proposes multi-scale grid method to model the road environment of intelligent vehicle. The road environment is divided into multi-scale grids. The grids representing the regions close to the vehicle are of large size. The grids representing the regions far away from the vehicle are of small size. The multi-scale gird function is presented to calculate the size of grid. Experimental results show that the proposed algorithm provides an efficient solution to the conflict between computational complexity and modeling accuracy.