Speaker
Description
Human motion is of interest to industrial simulations for process optimisation, ergonomic evaluation and visualisation of digital-twin environments. Furthermore, it is of interest for simulation-based reinforcement learning in human-robot interaction and humanoid robotics for industrial scenarios. Human motion data created algorithmically lacks the variability and naturalness of real human motions; therefore, using motion-capture-based or human skeletal data extracted from RGB data is the preferred method for gathering human motion data for generation. However, current human motion generation models are trained on datasets that focus on activities of daily living, thereby lacking the terminology and body movements required in industrial environments. This work represents the first step towards identifying the requirements of human motion generation models. It provides a working example of industrial datasets for human motion generation with the instance of LARa.
