Periodic Activations of Behaviors and Motivational States Ernesto Burattini and Silvia Rossi Dipartimento di Scienze Fisiche Universit` a degli Studi di Napoli ”Federico II” – Napoli, Italy ernb@na.infn.it, silrossi@unina.it Abstract. The possible modulatory influence of motivations and emo- tions is fundamental in designing robotic adaptive systems. In this pa- per, we will try to connect the concept of periodic behavior activations to emotions, in order to link the variability of behaviors to the circum- stances in which they are activated. We will study the impact of emotion, described as timed controlled structures, on simple reactive behaviors. We will show, through this approach, that the emergent behaviors of a simple robot designed with a parallel or hierarchical architecture are comparable. Finally, we will see that conflicts in behaviors may be solved without an explicit action selection mechanism. 1 Introduction In Robotics one of the main issues in designing a control system is to enable an autonomous robot to react and adapt in useful time to environmental changes [1]. This reaction depends on the correct identification of objects and their properties by appropriate sensor devices, with a strong emphasis on the concept of the stimuli-response loop. Moreover, the robotic community, started to pay attention not only to the robot-environment interactions, but also, so to speak, to the interactions that may arise within the robots itself [2] and how these latter (for example its emotional states) may influence the emergent behavior of the robot. In these last years some researchers [2–7] started to pay attention to the role of emotional and motivational states in order to achieve an adaptive emergent behavior of robotics systems. In particular, the role of emotions has been intro- duced for behavior modulations [3, 4], to provide adaptivity to environmental changes. Moreover, cognitive psychology considers thinking, learning and mem- ory activities as a problem of information processing. However, the description of motivational issues and emotional states as a processing problem is not an obvious task [3]. The interest for such “internal mechanisms” comes within the robotic community taking inspiration from ethological, biological and neuro- science studies. In our opinion, in order to model different and new architectures for controlling the robot behavior, both these aspects (the interaction with the surrounding world and the internal states) have to be considered, since they influence each other. For example, the simple perception-action response to an