Active Inference and Behavior Trees for Reactive Action Planning and Execution in Robotics

We propose a hybrid combination of active inference and behavior trees (BTs) for reactive action planning and execution in dynamic environments. We show how robotic tasks can be formulated as a free-energy minimization problem, bringing the neuroscientific theory of active inference on a mobile manipulator. In this framework, the general nominal behavior is specified offline through BTs. A new type of leaf node, the \textit{prior} node, is introduced to specify the desired state to be achieved rather than an action to be executed as in classical BTs. The decision of which action to execute to reach the desired state is performed online through active inference. The resulting hybrid combination improves the robustness of plans against unexpected contingencies while considerably reducing the number of nodes in a BT. The properties of our algorithm, such as the convergence and robustness, are thoroughly analyzed and outperform classical BT solutions. The theoretical results are validated using a mobile manipulator in a retail environment.