To Resume or Not to Resume: A Behavior Tree Extension
Wafic El-Ariss, Naseem A. Daher, Imad H. Elhajj · 2021
Behavior Trees (BTs) have become a valuable tool for the development of the decision-making aspect for automated agents, such as the Non-Player Characters (NPCs) in computer games, and more recently, the agents of the highly automated robotic applications. BTs allow for the development of easily reconfigurable behavior-selection algorithms. However, so far, the research has mostly focused on offering reactive BT structures, and less so on structures that remember past executions of BT sub-trees. In this paper, we show that a resuming behavior is an important concept, which cannot be developed in a scalable manner using the current BT tools, and we solve this development problem by adapting a concept used in Hierarchical Finite-State Machines, to propose an extension to the BT tools: the Sequence and Fallback nodes with history memory.