Towards an integrated and human-friendly path following and obstacle avoidance behaviour for robots
Camilla Bassani, Antonello Scalmato, Fulvio Mastrogiovanni, Antonio Sgorbissa · 2016
This paper proposes an integrated path following and obstacle avoidance framework for robots in crowded environments. The architecture considers two major requirements: (i) a given performance level for path following must be guaranteed; (ii) human-friendly robots must behave predictably and naturally, in order to avoid dangerous people reactions. In order to generate paths perceived as natural, and inspired by what happens in high traffic highways, we assume a path to consist of a family of curves, which can be switched using sensory information. The resulting behaviour is both predictable (paths are planned and validated beforehand) and natural (the overall qualitative behaviour is maintained when switching curves). The overall error is upper bounded by the curves configuration. The paper presents results in simulation.