ABSTRACT
T HIS CHAPTER focuses on the aptitude of a control architecture to exhibit re-active as well as cognitive skills according to the robot’s situation. The main objective of this kind of HybridRC (Reactive/Cognitive) structure is to enable us to deal on-line and safely with unpredictable or uncertain situations, and to optimize the
overall robot navigation if the environment is well-known/mastered (e.g., accurate localization, large perception field with low uncertainty, etc.). Furthermore, since the previous chapters have emphasized reactive navigation, this chapter will pay more attention to the proposed planning methods, mainly the one based on PELC for a car-like robot. Afterward, a homogeneous (in term of set-points and control law) and HybridRC multi-controller architecture will be presented in detail.
