Efficient Exploration of Mobile Robot Based on DL-RRT and AP-BO

The proliferation in demand for autonomous exploration and map construction of mobile robots has posed an ongoing challenge to the development of an efficient exploration strategy. The unknown map of the path planning environment imposes a challenge for the boundary exploration strategy and boundary point screening mechanism of mobile robots. Motivated by this challenge, an efficient method for boundary exploration using a double-layer rapid-exploration random tree (DL-RRT) is introduced. DL-RRT enables the exploration of boundary points in both global and local manners. Moreover, a boundary point optimization strategy combining affinity propagation and Bayesian optimization (AP-BO) is proposed. The proposed boundary-point exploration strategy is different from traditional target-point exploration. On the one hand, the affinity propagation mechanism is employed in boundary-point optimization, which exhibits superior clustering effects compared with the traditional algorithms. On the other hand, Bayesian optimization can filter the boundary points after clustering. The mobile robot prioritizes these clustered points according to the desired parameters. The comparison results of the simulation are conducted to demonstrate the efficacy of the exploration strategy. The results of the simulation experiment are analyzed through comparison with other methods in terms of running time and path length. It is concluded that the combination of DL-RRT, affinity propagation clustering, and Bayesian optimization enables the robot to efficiently identify boundary points, filter suitable clustering centers, and locate optimal target points. The proposed method exhibits an average reduction in running time by 16.38% compared with using RRT and K-means clustering. In addition, it demonstrates an average decrease in path length by 20.28% compared with using RRT and K-means clustering. At the same time, the standard deviation of the results remains around 1, indicating a stable operation.