Flexible Manipulation: Finite State Machine-based Collaborative Manipulation

Abstract

This paper presents a finite state machine-based approach to collaborative manipulation using the Robot Operating System (ROS) compatible MoveIt! planning system and the Flexible Behavior Engine (FlexBE). The system provides new enhanced MoveIt! capabilities and associated FlexBE state implementations that interact via standard ROS ActionLib interfaces. The system enables finite state machine-based planning using MoveIt! and trajectory execution using the standard ROS controllers framework. The software is directly applicable to any robot manipulator modeled with ROS and MoveIt!. In addition to planning and trajectory execution, state implementations also provide access to common functions such as inverse kinematics. The use of FlexBE directly enables adjustable autonomy levels and collaborative autonomy where the operator can verify transitions as needed. The software, which is available open source, is demonstrated on a 6 DOF Kinova MICO <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> arm and small Nao humanoid robot.