Learning theory and empirical potentials for modeling discrete mechanics

Dadashi, Shirin; McClelland, Hunter G.; Kurdila, Andrew J.

doi:10.23919/acc.2017.7963643

Cited by 4 publications

(1 citation statement)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper, equation ( 16) is the kernel with which the empirical potential energy is learned. The derivation of the empirical potential is outside the scope of this paper, so the reader is directed to [10,12,33] for a complete discussion. The result, however, is that the empirical potential function is identified as…”

Section: Identifying the Empirical Potential Functionsmentioning

confidence: 99%

Learning bioinspired joint geometry from motion capture data of bat flight

Bender¹,

Guo²,

Powell³

et al. 2019

Bioinspir. Biomim.

Self Cite

View full text Add to dashboard Cite

Bioinspired robotic systems are often designed by assuming that the kinematics of the biological system of interest are well approximated by an open kinematic chain subject to box constraints. These box-constrains are typically generated by designer interpretation of biomotion and anatomical studies or pragmatic fabrication constraints. In contrast to this standard design paradigm, this paper presents a methodology for learning joint geometry which restricts the range of motion of generic ball joints to only the reachable set observed during biomotion experiments. This reachable set is identified by constructing an analytical-empirical potential energy function over the experimental observations. The energy function effectively 'pushes' configurations close to the set of observations. This energy function is then thresholded to identify the zero-potential (ZP) configuration set, and joint geometry is constructed using the resulting contour. We construct an entire bat wing using this method, and demonstrate through motion capture experiments that the learned geometry successfully restricts motions to the ZP set.

show abstract

Section: Identifying the Empirical Potential Functionsmentioning

confidence: 99%

Learning bioinspired joint geometry from motion capture data of bat flight

Bender¹,

Guo²,

Powell³

et al. 2019

Bioinspir. Biomim.

Self Cite

View full text Add to dashboard Cite

show abstract

A Data-Driven Method Based on Quadratic Programming for Distance-Based Formation Control of Euler–Lagrange Systems

Rodríguez

Fernández

Moreno

et al. 2021

IEEE Control Syst. Lett.

View full text Add to dashboard Cite

Empirical Potential Functions for Driving Bioinspired Joint Design

Bender

George

Powell

et al. 2018

Journal of Dynamic Systems, Measurement, and Control

View full text Add to dashboard Cite

Bioinspired design of robotic systems can offer many potential advantages in comparison to traditional architectures including improved adaptability, maneuverability, or efficiency. Substantial progress has been made in the design and fabrication of bioinspired systems. While many of these systems are bioinspired at a system architecture level, the design of linkage connections often assumes that motion is well approximated by ideal joints subject to designer-specified box constraints. However, such constraints can allow a robot to achieve unnatural and potentially unstable configurations. In contrast, this paper develops a methodology, which identifies the set of admissible configurations from experimental observations and optimizes a compliant structure around the joint such that motions evolve on or close to the observed configuration set. This approach formulates an analytical-empirical (AE) potential energy field, which “pushes” system trajectories toward the set of observations. Then, the strain energy of a compliant structure is optimized to approximate this energy field. While our approach requires that kinematics of a joint be specified by a designer, the optimized compliant structure enforces constraints on joint motion without requiring an explicit definition of box-constraints. To validate our approach, we construct a single degree-of-freedom elbow joint, which closely matches the AE and optimal potential energy functions and admissible motions remain within the observation set.

show abstract

Learning theory and empirical potentials for modeling discrete mechanics

Cited by 4 publications

References 21 publications

Learning bioinspired joint geometry from motion capture data of bat flight

Learning bioinspired joint geometry from motion capture data of bat flight

A Data-Driven Method Based on Quadratic Programming for Distance-Based Formation Control of Euler–Lagrange Systems

Empirical Potential Functions for Driving Bioinspired Joint Design

Contact Info

Product

Resources

About