A hybrid strategy for robot navigation in semi-structured environments
Guilherme Carlos R. de Oliveira, Kevin B. de Carvalho, Alexandre Santos Brandao · 2018
Autonomous mobile robots should be able to navigate on different types of environments. Some strategies deals with the navigation problem by taking into consideration only the obstacle avoidance task, adopting a reactive behavior. Others, at cost of increased hardware processing, look for optimal paths that must be replanned once an obstacle appears in the planned path. The proposal of this paper is to integrate the A∗ search and obstacle avoidance tangential escape algorithms, in order to overcome the A∗ limitations in non-static environments. Instead of calculating a new route to avoid collisions with unexpected obstacles, as in most popular solutions in the literature, the robot perform deviations while it tries to keep on the planned route. The strategy also uses the robots sensor data for mapping the environment in an Occupancy Grid, in order to navigate deliberatively. Numerical simulations validate the proposal.