Bio-Inspired Complete Coverage Path Planner for Precision Agriculture in Dynamic Environments

摘要

This paper proposes a bio-inspired Complete Coverage Path Planner suitable for several precision agriculture tasks, such as terrain and crop mapping, inspection, and crop spraying. This grid-based method reproduces the dynamics of the neural activity in a biological neural system to represent dynamically varying environments. By providing appropriate inputs to the neurons of the grid, their neural activity can be exploited to guide the robot towards uncovered regions of the area and enforce the desired coverage pattern. Both known and unexpected obstacles can be easily handled, since the sudden discovery of an obstacle simply modifies the local neural activity online. Thus, the need for complete re-planning phases is canceled. A deadlock-escaping mechanism is also proposed to efficiently recover from dead ends. Finally, simulation results are provided to show the flexibility and effectiveness of the method in dynamic environments.