Variable-Speed Teaching-Playback as Real-World Data Augmentation for Imitation Learning

Imitation learning’s inevitable reliance on human demonstrations in hard-to-simulate settings has resulted in a shortage of training data, even with a simple change in speed. Although the field of data augmentation has addressed the lack of data, conventional methods of data augmentation for robot manipulation are limited to simulation-based methods or downsampling for position control. This paper proposes a novel method of data augmentation that is applicable to force control and preserves the advantages of real-world datasets. We applied teaching–playback at variable speeds as real-world data augmentation to increase both the quantity and quality of environmental reactions at variable speeds. An experiment was conducted on bilateral control-based imitation learning using a method of imitation learning equipped with position–force control. We evaluated the effect of real-world data augmentation on two tasks, pick-and-place and wiping, at variable speeds, each from two human demonstrations at fixed speed. The results showed that generation of variable-speed movement from human demonstrations with less than 10% variety of durations is feasible, and improved success rate and accuracy along the duration/frequency command by gathering environmental reactions at variable speeds, rather than simply altering speeds of gathered data. GRAPHICAL ABSTRACT