RoboCat: A category theoretic framework for robot interoperability using goal-oriented programming.
Angeline Aguinaldo, Blake S. Pollard, Arquimedes Canedo, Gustavo Quiros, William Regli
This paper presents RoboCat, a category theoretic framework for robot interoperability using goal- oriented programming. RobotCat uses a high-level, goal-oriented declarative language for robot programming that allows workers from a variety of skill levels to specify what the robot should do. Automated tools then translate these goal-oriented functional specifications into procedural code to execute these tasks in various robot platforms. The use of a goal-oriented programming paradigm allows the reusability of manufacturing, production, and robotics knowledge. Our approach brings the principle of compositionality to the forefront when considering functions and their relationships. This has led us to rigorous mathematical representations found in category theory as a means of formally modeling modularity and behavioral knowledge. This category theoretic approach enables: (i) a top-down and bottom-up mapping from goal-oriented programs to low-level robot commands, and (ii) reuse of programs across a range of robotic platforms. By enforcing the described model, we can define hierarchical interfaces that require only local modifications when new robot software or world models are introduced; thus, making robot programming more interoperable.
IEEE Transactions on Automation Science and Engineering
, Pollard, B.
, Canedo, A.
, Quiros, G.
and Regli, W.
RoboCat: A category theoretic framework for robot interoperability using goal-oriented programming., IEEE Transactions on Automation Science and Engineering, [online], https://doi.org/10.1109/TASE.2021.3094055
(Accessed December 4, 2023)