Communicating human intent to a robotic companion by multi-type gesture sentences

Human-Robot collaboration in home and industrial workspaces is on the rise. However, the communication between robots and humans is a bottleneck. Although people use a combination of different types of gestures to complement speech, only a few robotic systems utilize gestures for communication. In this paper, we propose a gesture pseudo-language and show how multiple types of gestures can be combined to express human intent to a robot (i.e., expressing both the desired action and its parameters - e.g., pointing to an object and showing that the object should be emptied into a bowl). The demonstrated gestures and the perceived table-top scene (object poses detected by CosyPose) are processed in real-time) to extract the human's intent. We utilize behavior trees to generate reactive robot behavior that handles various possible states of the world (e.g., a drawer has to be opened before an object is placed into it) and recovers from errors (e.g., when the scene changes). Furthermore, our system enables switching between direct teleoperation of the end-effector and high-level operation using the proposed gesture sentences. The system is evaluated on increasingly complex tasks using a real 7-DoF Franka Emika Panda manipulator. Controlling the robot via action gestures lowered the execution time by up to 60%, compared to direct teleoperation.

On Partial Opimality by Auxiliary Submodular Problems

In this work, we prove several relations between three different energy minimization techniques. A recently proposed methods for determining a provably optimal partial assignment of variables by Ivan Kovtun (IK), the linear programming relaxation approach (LP) and the popular expansion move algorithm by Yuri Boykov. We propose a novel sufficient condition of optimal partial assignment, which is based on LP relaxation and called LP-autarky. We show that methods of Kovtun, which build auxiliary submodular problems, fulfill this sufficient condition. The following link is thus established: LP relaxation cannot be tightened by IK. For non-submodular problems this is a non-trivial result. In the case of two labels, LP relaxation provides optimal partial assignment, known as persistency, which, as we show, dominates IK. Relating IK with expansion move, we show that the set of fixed points of expansion move with any "truncation" rule for the initial problem and the problem restricted by one-vs-all method of IK would coincide -- i.e. expansion move cannot be improved by this method. In the case of two labels, expansion move with a particular truncation rule coincide with one-vs-all method.