ARTICLE Giusti_RAM_2018/IDIAP Flexible Automation Driven by Demonstration: Leveraging Strategies that Simplify Robotics Giusti, A. Zeestraten, M. Icer, E. Pereira, A. Caldwell, D. G. Calinon, Sylvain Althoff, M. learning from demonstration EXTERNAL https://publications.idiap.ch/attachments/papers/2018/Giusti_RAM_2018.pdf PUBLIC IEEE Robotics and Automation Magazine (RAM) 25 2 18-27 2018 https://ieeexplore.ieee.org/document/8356588/ URL 10.1109/MRA.2018.2810543 doi How do I automate this task? This is a crucial question for production engineers. In classical industrial automation (think of a car factory) robots perform a small set of tasks for long periods of time. They were selected because their kinematic structure and strength suited the task requirements, and their motions were pre-programmed by a skilled programmer. In flexible manufacturing environments, tasks may change daily or hourly. The classical approach to automation is less suitable here: buying a dedicated robot for each set of tasks is uneconomical and coding each task is too time consuming. Further, qualified programmers with the requisite knowledge might not be available. To solve these problems, we propose a framework for modular robots that determines their structure and program based on human demonstrations.