Search Menu
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information
Search Scholarly Publications

Title: An Approach to Realize Generalized Optimal Motion Primitives Using Physics Informed Neural Networks

Journal Article · · ASME Letters in Dynamic Systems and Control
DOI: https://doi.org/10.1115/1.4066627 · OSTI ID:2476627
 [1];  [1];  [2];  [1]
  1. Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
  2. Old Dominion Univ., Norfolk, VA (United States)
+ Show Author Affiliations

Autonomous manipulation is a challenging problem in field robotics due to uncertainty in object properties, constraints, and coupling phenomenon with robot control systems. Humans learn motion primitives over time to effectively interact with the environment. We postulate that autonomous manipulation can be enabled by basic sets of motion primitives as well, but do not necessitate mimicking human motion primitives. Here, this work presents an approach to generalized optimal motion primitives using physics-informed neural networks. Our simulated and experimental results demonstrate that optimality is notionally maintained where the mean maximum observed final position percent error was 0.564% and the average mean error for all the trajectories was 1.53%. These results indicate that notional generalization is attained using a physics-informed neural network approach that enables near optimal real-time adaptation of primitive motion profiles.

Research Organization:
Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States)
Sponsoring Organization:
USDOE Laboratory Directed Research and Development (LDRD) Program
Grant/Contract Number:
NA0003525
OSTI ID:
2476627
Report Number(s):
SAND--2024-15169J
Journal Information:
ASME Letters in Dynamic Systems and Control, Journal Name: ASME Letters in Dynamic Systems and Control Journal Issue: 2 Vol. 5; ISSN 2689-6117
Publisher:
ASMECopyright Statement
Country of Publication:
United States
Language:
English

References (13)

Physical Interaction via Dynamic Primitives book January 2017
Optimal control of differentially flat systems is surprisingly easy journal January 2024
On path following control of nonholonomic port-Hamiltonian systems via generalized canonical transformations journal January 2019
Passivity Analysis of Quadrotor Aircraft for Physical Interactions conference October 2021
Motion primitive based random planning for loco-manipulation tasks conference November 2016
Planning for Manipulation with Adaptive Motion Primitives conference May 2011
Manipulation Planning Among Movable Obstacles Using Physics-Based Adaptive Motion Primitives conference May 2021
Deep Lagrangian Networks for end-to-end learning of energy-based control for under-actuated systems conference November 2019
Search-based planning for manipulation with motion primitives conference May 2010
Port-Hamiltonian Neural ODE Networks on Lie Groups for Robot Dynamics Learning and Control journal January 2024
Sliding Mode Impedance Control of a Hydraulic Artificial Muscle
  • Slightam, Jonathon E.; Nagurka, Mark L.; Barth, Eric J.
  • Volume 1: Advances in Control Design Methods; Advances in Nonlinear Control; Advances in Robotics; Assistive and Rehabilitation Robotics; Automotive Dynamics and Emerging Powertrain Technologies; Automotive Systems; Bio Engineering Applications; Bio-Mechatronics and Physical Human Robot Interaction; Biomedical and Neural Systems; Biomedical and Neural Systems Modeling, Diagnostics, and Healthcare https://doi.org/10.1115/DSCC2018-9186
conference September 2018
Sliding Mode Impedance and Stiffness Control of a Pneumatic Cylinder
  • Slightam, Jonathon E.; Barth, Eric J.; Nagurka, Mark L.
  • Volume 1: Advanced Driver Assistance and Autonomous Technologies; Advances in Control Design Methods; Advances in Robotics; Automotive Systems; Design, Modeling, Analysis, and Control of Assistive and Rehabilitation Devices; Diagnostics and Detection; Dynamics and Control of Human-Robot Systems; Energy Optimization for Intelligent Vehicle Systems; Estimation and Identification; Manufacturing https://doi.org/10.1115/DSCC2019-9009
conference October 2019
Combining physics and deep learning to learn continuous-time dynamics models journal March 2023

Similar Records

Intelligent control and adaptive systems; Proceedings of the Meeting, Philadelphia, PA, Nov. 7, 8, 1989
Conference · 1989 · OSTI ID:6251437
Hierarchical intelligent control for robotic motion
Journal Article · 1994 · IEEE Transactions on Neural Networks (Institute of Electrical and Electronics Engineers); (United States) · OSTI ID:6818187
Utilizing Reinforcement Learning to Continuously Improve a Primitive-Based Motion Planner
Journal Article · 2022 · Journal of Aerospace Information Systems · OSTI ID:1845399
+2 more

Related Subjects

97 MATHEMATICS AND COMPUTING
machine learning
neural networks
optimization algorithms
robotics