ENACT: An Efficient and Extensible Entity-Actor Framework for Modular Robotics Software Components

Abstract:
Recent advances towards generic robot programs require efficient sharing of copious information among many functional modules. Current solutions favor extensibility over efficiency and thus limit attainable functionality. In contrast, we contribute a software framework that efficiently implements the extensible entity-aspect-actor paradigm. Our framework remains safe for multi-threading, scales to distributed systems, handles inconsistent data and supports history logging. Evaluation through scenarios in industrial and service robotics attests claimed benefits. We conclude that our framework is fit for use in robotics, with significance also for computer games, simulations, and VR applications.