Title: “Model-based Optimization for Humanoid Walking on Challenging Terrain.”
Planning locomotion trajectories in challenging environment, such as rough terrain, stairs, step stone or steep slopes, is a difficult task. Based on the assumption that human gait is optimal because of evolution and personal learning, we follow the approach to transfer motion generation principles from humans to humanoid robots based on the underlying optimality criteria. To identify these criteria also for challenging walking situations, human locomotion is analyzed by means of sophisticated mathematical models and optimal control techniques. Template models which reproduce characteristic center of mass behavior and swing foot trajectories are considered, such that the same model can be used for both, the analysis of human gait on the one hand and the generation of humanoid walking patterns on the other.
Bio: Debora Clever is a postdoctoral researcher at the University of Heidelberg within the EU-project KoroiBot. She has studied techno-mathematics with major on numerical analysis, computer science and cybernetics at the Technische Universitaet Darmstadt where she also received her PhD in applied mathematics in 2013. Her research focus is on inverse optimal control for human gait analysis and on optimal control for humanoid gait generation.