Search for: All records

This paper aims to identify in a practical manner unknown physical parameters, such as mechanical models of actuated robot links, which are critical in dynamical robotic tasks. Key features include the use of an off-the-shelf physics engine and the Bayesian optimization framework. The task being considered is locomotion with a high-dimensional, compliant Tensegrity robot. A key insight, in this case, is the need to project the space of models into an appropriate lower dimensional space for time efficiency. Comparisons with alternatives indicate that the proposed method can identify the parameters more accurately within the given time budget, which also results in more precise locomotion control. 

Tensegrity rovers incorporate design principles that give rise to many desirable properties, such as adaptability and robustness, while also creating challenges in terms of locomotion control. A recent milestone in this area combined reinforcement learning and optimal control to effect fixed-axis rolling of NASA’s 6-bar spherical tensegrity rover prototype, SUPERball, with use of 12 actuators. The new 24-actuator version of SUPERball presents the potential for greatly increased locomotive abilities, but at a drastic nominal increase in the size of the data-driven control problem. This paper is focused upon unlocking those abilities while crucially moderating data requirements by incorporating symmetry reduction into the controller design pipeline, along with other new considerations. Experiments in simulation and on the hardware prototype demonstrate the resulting capability for any-axis rolling on the 24-actuator version of SUPERball, such that it may utilize diverse ground-contact patterns to smoothly locomote in arbitrary directions. 

Effective robotic systems must be able to produce desired motion in a sufficiently broad variety of robot states and environmental contexts. Classic control and planning methods achieve such coverage through the synthesis of model-based components. New applications and platforms, such as soft robots, present novel challenges, ranging from richer dynamical behaviors to increasingly unstructured environments. In these setups, derived models frequently fail to express important real-world subtleties. An increasingly popular approach to deal with this issue corresponds to end-to-end machine learning architectures, which adapt to such complexities through a data-driven process. Unfortunately, however, data are not always available for all regions of the operational space, which complicates the extensibility of these solutions. In light of these issues, this paper proposes a reconciliation of classic motion synthesis with modern data-driven tools towards the objective of ``deep coverage''. This notion utilizes the concept of composability, a feature of traditional control and planning methods, over data-derived ``motion elements'', towards generalizable and scalable solutions that adapt to real-world experience.