From Locomotion to Cognition: "Cheap" Rapid Legged Locomotion

• Abstract:
  

  A lot of our research in the past has been on relatively low-level sensory-motor tasks such as locomotion (e.g. walking, running, jumping), navigation, and grasping. While this research is of interest in itself, in the context of artificial intelligence and cognitive science, this leads to the question of what these kinds of tasks have to do with higher levels of cognition, or to put it more provocatively, "What does walking have to do with thinking?" This question is of course reminiscent of the notorious "symbol grounding problem". In contrast to most of the research on symbol grounding, we propose to exploit the dynamic interaction between the embodied agent and the environment as the basis for grounding. We use the term "morphological computation" to designate the fact that some of the control or computation can be taken over by the dynamic interaction derived from morphological properties (e.g. the passive forward swing of the leg in walking, the spring-like properties of the muscles, and the weight distribution). By taking morphological computation into account, an agent will be able to achieve not only faster, more robust, and more energy-efficient behavior, but also more situated exploration by the agent for the comprehensive understanding of the environment.

  The main objective of this project is to explore the design principles of biologically inspired legged running robots. In particular this project focuses on a minimalistic model of rapid locomotion of quadruped robots inspired by biomechanics studies. The goal of this project is, therefore, to achieve technology for a form of rapid legged locomotion as well as to obtain our further understanding of locomotion mechanisms in biological systems. By exploiting morphological properties, we have developed the running quadruped robot "Puppy", which is capable of relatively robust rapid legged locomotion by using intrinsic body dynamics induced by spring-like like property, weight distribution, and body dimentions. Owing to the use of body dynamics, the control of the robot is extremely simple and, moreover, it has rich behavioral diversity.

• People:

• Matej Hoffmann
• Juan Pablo Carbajal
• Lijin Aryananda
• Max Lungarella
• Fumiya Iida

• Available student projects: (also listed in Open Student Projects)
  
  
  • If you are interested in one of the listed projects, or if you want to arrange your own research projects (semester thesis, bachelor thesis, master thesis, diploma thesis, SIROP project, laboratory practice, etc), please contact Matej Hoffmann, hoffmann at ifi.uzh.ch.

     

  • Exploration of body capabilities using chaotic dynamics (MSc. thesis) ( Detailed description)

    In this project, we attempt to answer the question of how an embodied system can autonomously explore and discover the action possibilities inherent to its body by exploiting the interaction of neural dynamics and body dynamics. Recent investigations have shown how chaotic oscillators can be used in a variety of embodied systems as a means of self-exploration of movement possibilities. The goal of this project is to explore these concepts in simulated and, above all, real robots. The outcome should be a control architecture that allows a robot to explore its own movement possibilities through environmentally triggered transitions from one behavior to another. The first part of the project will consist in realizing a simulation of a physical system to test potential control architectures. A starting point will be a simulation of our dog robot in Matlab and Webots. The core experiments will be performed with a real robot Puppy.