Vision-based mobile robot motion control combining T~2 and ND approaches

摘要

Navigating along a set of programmed points in a completely unknown environment is a challenging task which mostly depends on the way the robot perceives and symbolizes the environment and decisions it takes in order to avoid the obstacles while it intends to reach subsequent goals. Tenacity and Traversability (T~2)~1 -based strategies have demonstrated to be highly effective for reactive navigation, extending the benefits of the artificial Potential Field method to complex situations, such as trapping zones or mazes. This paper presents a new approach for reactive mobile robot behavior control which rules the actions to be performed to avoid unexpected obstacles while the robot executes a mission between several defined sites. This new strategy combines the T~2 principles to escape from trapping zones together with additional criteria based on the Nearness Diagram (ND)~(13) strategy to move in cluttered or densely occupied scenarios. Success in a complete set of experiments, using a mobile robot equipped with a single camera, shows extensive environmental conditions where the strategy can be applied.