Multi-strategy based exploration for 3D mapping in unknown environments using a mobile robot

In this paper, a multi-strategy based exploration approach is proposed, enabling a mobile robot to autonomously explore unknown environments and construct 3D maps of the world. The approach combines information-gain based local exploration strategy and global exploration strategy. Firstly, the information-gain based decision model for exploration is formulated and an information gain calculation method which can be applied to evaluate the utility of exploration trajectory over multiple time steps is designed. Then a model-predictive based trajectory generation method is used to generate a set of candidate trajectories for local exploration. The best local exploration trajectory is obtained based on the information-gain evaluation criterion. When this strategy gets to a local minima, a complementary global strategy is executed. Specifically, a set of viewpoints that map to sightings of frontier regions is generated. Then a graph search method is used to search exploration path to the candidate viewpoints. The best global exploration path is also obtained based on the information-gain evaluation criterion. Finally, real-world experiment is conducted to accomplish the exploration and 3D mapping task on a mobile robot in indoor unknown environments. Experimental results show that the proposed multi-strategy based exploration method performs fast exploration in 3D unknown environment, meanwhile, it ensures the completeness of the obtained map.