Planejamento de trajetórias baseado em espuma probabilística para sistemas robóticos autônomos em ambientes dinâmicos

Abstract

This study presents an enhanced version of the Probabilistic Foam Method (PFM), focused on motion planning for autonomous robots. In its original form, PFM is developed over a static and previously known configuration space. In this context, the free space is partially filled with overlapping convex bubbles that form a structure resembling foam, creating a safe zone for movement and ensuring safety during maneuvers. From this structure, a search tree is constructed over the bubbles, identifying a feasible path between the specified initial and final configurations. This work proposes improvements to the foam propagation strategy of the PFM to adapt the method to dynamic environments. To handle moving obstacles, the bubble expansion process is modified to account for both spatial and temporal requirements. Specifically, from the surface of a parent bubble, coordinates for the center of a new bubble are randomly generated, considering both position and time. The proposed method stands out for its ability to quickly determine efficient solutions for trajectory planning problems in dynamic environments, even in the presence of moving obstacles with varying but uniform speeds. The algorithm’s effectiveness is validated through functional simulations, where parameters such as the minimum bubble radius, the robot’s maximum speed, and the speeds and directions of the moving obstacles are defined. The simulation results demonstrate the feasibility and efficiency of the proposed method, highlighting its capability to generate safe trajectories in dynamic and previously known scenarios. Thus, the enhanced PFM provides a robust and adaptable planning framework for situations subject to changes over time.