Improving Human-Robot Object Exchange by Online Force Classification

Abstract

Handing an object over to a human is a challenging task for a robot to perform, especially when the human partner has no experience interacting with robots. This paper presents our work to enable a robot to learn how to achieve this task with wrist force/torque sensing. Firstly, we present a device to record the data, then we discuss the techniques used for the teaching. We choose to focus on the classification problem defined to enable the robot to find the finger opening movement. The main challenge is that the classification should be run online, at a comparable rate to the controller. To achieve a computationally efficient classifier, the Wavelet Packet Transformation is used for feature extraction, and then we used the Fisher criterion to reduce the dimension of features. A Relevance Vector Machine is used for the continuous classification procedure mainly for its sparsity. Some recorded data and the results from dimension reduction are shown, then we discuss the accuracy and sparsity of the classification by Relevance Vector Machine in this application. The software of continuous classification on forces is then tested on the robot for interactive object exchange between human and robot, which gives promising results.

The where of handovers by humans: Effect of partner characteristics, distance and visual feedback

“…A human handover starts from the time an intention to pass or receive an object is generated [2], which is followed by the hand movements both by the giver and the receiver to a particular position where the object transfer will take place. Finally, there is a physical interaction in which the object is transferred from one hand to the other [3]. Previous studies have examined issues regarding the velocity [4, 5] and trajectory [6, 7] of handover movements, and the grip force [3, 8] and grasp points [9, 10], and location [11, 12] and configuration [13, 14] of the exchanged object.…”

“…Finally, there is a physical interaction in which the object is transferred from one hand to the other [3]. Previous studies have examined issues regarding the velocity [4, 5] and trajectory [6, 7] of handover movements, and the grip force [3, 8] and grasp points [9, 10], and location [11, 12] and configuration [13, 14] of the exchanged object. Here we were interested in how the specific physical and social characteristics of a partner influences handovers.…”