Detailseite
Projekt Druckansicht

Adaptives Laufen durch Multi-Kontakt Stabilisierung und Nutzung von Teilkontakten für humanoide Roboter

Fachliche Zuordnung Automatisierungstechnik, Mechatronik, Regelungssysteme, Intelligente Technische Systeme, Robotik
Bild- und Sprachverarbeitung, Computergraphik und Visualisierung, Human Computer Interaction, Ubiquitous und Wearable Computing
Förderung Förderung von 2018 bis 2022
Projektkennung Deutsche Forschungsgemeinschaft (DFG) - Projektnummer 407378162
 
Erstellungsjahr 2022

Zusammenfassung der Projektergebnisse

Although research on life-sized humanoid robots started already 50 years ago, an actual real world application still sounds like science fiction to most of us. However, in the last couple of years significant progress was made which led to multiple impressive humanoids presented by high-tech companies. Although they appear to be very versatile, robust, and even acrobatic in promoted videos, one has to keep in mind that even these high end systems have to be extensively tuned for their demonstrations and they still do not fulfill the high requirements of a real world application. While wheeled robots have proven their effectiveness in various fields, reaching from autonomous carriers in warehouse logistics to robot vacuum cleaners, legged systems still have not arrived in the consumer market yet. Although a legged system tends to be much more versatile, the design and control of such systems is rather complex. Especially biped robots are difficult to balance, which brings them into the focus of current research. Apart from the scientific challenges, the main motivation of investigating full-sized humanoid robots is that they are expected to have the best locomotion capabilities in environments made for humans, e.g. when stepping over obstacles or climbing stairs is considered. This paves the way for various potential future applications such as robot assistants in health and elderly care, the last mile in parcel delivery, or the entertainment industry. Moreover, many findings which originate from research on humanoid robots can be directly transferred to different applications, such as autonomous driving. The goal of this particular project was to make locomotion of humanoid robots more versatile and robust. For this, the robot should better adapt its motion to the environment, primarily by using explicit hand-environment contacts for additional support during difficult maneuvers. Potential support can be provided by walls, tables, or handrails when climbing stairs. Apart from this, a secondary goal of this project was to enable partial contacts of the feet, by which we mean that only the toe segment of the foot is in contact. This allows the robot to climb stairs with short treads, but also enables tiptoe contacts which significantly increases the kinematic capabilities during stepping down. The aforementioned topics have been investigated using the already existing research prototype L OLA, which was developed at the Technical University of Munich (TUM) (funded mainly by the DFG). Within the scope of this project, the robot has received substantial upgrades to its hardware, such as a complete redesign of the upper body, but also to its software, in particular the computer vision, contact planning, motion generation, and stabilization systems. Each of these challenges represent an area of research on their own, which emphasizes the multi-disciplinary character of this project. Apart from various scientific papers which originated from this project, also multiple videos have been published which explain the developed methods and showcase our results. The most prominent ones are youtu.be/gUNZ0AmLiWU and youtu.be/ovG2Rz9-1p8 which give a good overview of the achievements of this project.

Projektbezogene Publikationen (Auswahl)

  • “Quintic Spline Collocation for Real-Time Biped Walking-Pattern Generation with variable Torso Height”, in: IEEE-RAS 19th International Conference on Humanoid Robots (Humanoids), Toronto, Canada, 2019
    Seiwald, P., Sygulla, F., Staufenberg, N.-S., and Rixen, D.
    (Siehe online unter https://doi.org/10.1109/Humanoids43949.2019.9035076)
  • “A force-control scheme for biped robots to walk over uneven terrain including partial footholds”, in: International Journal of Advanced Robotic Systems 17.1 (2020)
    Sygulla, F. and Rixen, D.
    (Siehe online unter https://doi.org/10.1177/1729881419897472)
  • “Experimental Analysis of Structural Vibration Problems of a Biped Walking Robot”, in: IEEE International Conference on Robotics and Automation (ICRA), Paris, France, 2020
    Berninger, T. F. C., Sygulla, F., Fuderer, S., and Rixen, D. J.
    (Siehe online unter https://doi.org/10.1109/ICRA40945.2020.9197282)
  • “Fast Approximation of Over-Determined Second-Order Linear Boundary Value Problems by Cubic and Quintic Spline Collocation”, in: Robotics 9.2 (2020)
    Seiwald, P. and Rixen, D.
    (Siehe online unter https://doi.org/10.3390/robotics9020048)
  • “LOLA v1.1 – An Upgrade in Hardware and Software Design for Dynamic Multi-Contact Locomotion”, in: IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids), Munich, Germany, 2021
    Seiwald, P., Wu, S.-C., Sygulla, F., Berninger, T. F. C., Staufenberg, N.-S., Sattler, M. F., Neuburger, N., Rixen, D., and Tombari, F.
    (Siehe online unter https://doi.org/10.1109/HUMANOIDS47582.2021.9555790)
  • “SceneGraphFusion: Incremental 3D Scene Graph Prediction from RGB- D Sequences”, in: IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021
    Wu, S.-C., Wald, J., Tateno, K., Navab, N., and Tombari, F.
    (Siehe online unter https://doi.org/10.1109/CVPR46437.2021.00743)
 
 

Zusatzinformationen

Textvergrößerung und Kontrastanpassung