A Virtual Musculoskeletal Model for Variable Compliance and Joint Stabilization of a Walking Hexapod
-
1
Department for Computational Neuroscience Bernstein Center for Computational Neuroscience Georg-August-Universität Göttingen, Germany
It has long been acknowledged that muscles work as motors. In fact, they can also perform as dampers, springs, and struts. While under tension or during motion, they may lengthen, shorten, or remain the same. Due to their elastic property, they can generate compliant motions and achieve joint stabilization. These two characters facilitate interactions between robots and the environment. For example, compliant joint motions can prevent robots and human from damages owing to the ‘soft’ property.
Many traditional methods imitate muscle functionalities to produce compliant joint motions and achieve joint stabilization. Series Elastic Actuators (SEAs) enable robots to lower impact force and achieve joint stabilization by physical spring and damper mechanisms. But tunable SEAs are usually bulky, heavy and mechanically complex. This is unsuitable for small walking robots. A software control method is an alternative for joint compliance and stabilization. Active compliance uses force control (also known as impedance control) to change stiffness of actuators for compliance and stabilization of the robots. Active compliance, however, relies on position and torque feedback. This property slows down joint reactions to external loads. In addition to active compliance, VMC (Virtual Model Control) is another simple force control approach. It applies simulations of virtual mechanical mechanisms to yield real actuator torques. The torques produces compliant behaviors as if physical mechanisms are connected to the real robot. However, its major disadvantage is that many parameters need to be tuned by hand, which mainly limit structures of robots and over-depend experiences of developers.
In this study we propose a virtual musculoskeletal model for a biologically inspired six-legged walking machine, i.e. AMOS (Advanced MObility Sensor driven-walking device). This virtual musculoskeletal model enables AMOS to easily adjust joint stiffness and reduce leg contact force when confronted with external disturbances. Each AMOS joint can be virtually driven by a pair of antagonistic mechanisms. ‘Virtual’ here means joint can mimic antagonist muscle behaviors without physical spring and damper components. With external excitation, these virtual mechanisms drive joints as if the joint is driven by agonist and antagonist muscles. The agonist and antagonist muscles are mathematically modeled by a spring-damper mechanism.
The AMOS joint stiffness can be variably tuned by a stiffness coefficient when being excited by a sensory feedback, e.g. a contact force feedback. This simple stiffness regulation differs from mechanically complex SEAs and sophisticated VMC (Virtual Model Control) controllers. Moreover, only a contact force feedback can excite joint movement, which distinguishes from active compliance overdepending on position and torque feedback. The virtual musculoskeletal model has been tested on AMOS leading to variable compliance with simple tuning, joint stabilization to external perturbations (e.g. hand pushing and releasing AMOS body, object dropping on AMOS body), as well as a reduction of contact force during AMOS walking.
Acknowledgements
This research was supported by the Emmy Noether Program of the Deutsche Forschungsgemeinschaft (DFG, MA4464/3-1), and the Federal Ministry of Education and Research (BMBF) by a grant to the Bernstein Center for Computational Neuroscience II Göttingen (01GQ1005A, project D1).
Keywords:
Antagonistic Mechanisms,
Impedance Control,
Muscle Functionalities,
Stiffness Tuning
Conference:
Bernstein Conference 2012, Munich, Germany, 12 Sep - 14 Sep, 2012.
Presentation Type:
Poster
Topic:
Motor control, movement, navigation
Citation:
Xiong
X,
Wörgötter
F and
Manoonpong
P
(2012). A Virtual Musculoskeletal Model for Variable Compliance and Joint Stabilization of a Walking Hexapod.
Front. Comput. Neurosci.
Conference Abstract:
Bernstein Conference 2012.
doi: 10.3389/conf.fncom.2012.55.00238
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
09 May 2012;
Published Online:
12 Sep 2012.
*
Correspondence:
Mr. Xiaofeng Xiong, Department for Computational Neuroscience Bernstein Center for Computational Neuroscience Georg-August-Universität Göttingen, Göttingen, Germany, xizi@mmmi.sdu.dk