Automatic Decentralized Behavior Tree Synthesis and Execution for Coordination of Intelligent Vehicles

Abstract

This paper proposes a systematic approach for automatic tasking and coordination of a heterogeneous team of cooperative autonomous vehicles forming an intelligent vehicle. Each vehicle is equipped with different resources, operating in a shared dynamic environment, and capable of executing a set of specific tasks. To coordinate such a heterogeneous team of vehicles, we develop a hierarchical modular coordination algorithm for generating local Behavior Trees (BTs) for tasking individual vehicles so that they, as a team, can collectively achieve a global mission. A hierarchical auctioning algorithm is embedded in the proposed framework to effectively assign tasks among the vehicles, so that they can complement each other and complete a mission that may not be achieved individually. Furthermore, by introducing operational and duration cost weight terms, the proposed approach provides the possibility to adjust the total operation cost and duration of the tasks. The details of the developed algorithms are illustrated through different case studies.