This paper presents a framework for a visual search behavior of a 3D object in a 3D environment performed by a HRP-2 humanoid robot. The object search falls in the field of sensor planning and is formulated as an optimization problem. The goal is to maximize the target detection proba- bility while minimizing the energy/distance and time to achieve the task. This paper propose some natural constraints based on specificities of the humanoid robot and on the characteristics of the recognition system to reduce the dimension of the problem. It also guides the optimization by restricting the sensor configuration space to meet heavy time constraints on the behavior in order to have a reactive and near human like reaction time.

Toulouse, France