• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 2
  • Tagged with
  • 3
  • 3
  • 2
  • 2
  • 2
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

Analysis of Advanced Control Methods for Quadrotor Trajectory Tracking

Milburn, Tyler 08 October 2018 (has links)
No description available.
2

Design, modeling and control of inherently compliant actuators with a special consideration on agonist-anthropomorphic configuration / Conception, modélisation et contrôle d'actionneurs intrinsèquement conformes avec une considération spéciale sur la configuration anthropomorphe agoniste-antagoniste

Hari shankar lal das, Ganesh kumar 22 December 2016 (has links)
Conception, modélisation et contrôle des actionneurs intrinsèquement conformes avec une considération particulière sur la configuration anthropomorphe agoniste-antagoniste "La recherche vise à la conception, la modélisation et le contrôle des actionneurs intrinsèquement conformes pour les systèmes anthropomorphes.La première partie du travail se concentre sur l'étude de divers Existants et rechercher la possibilité d'autres actionneurs autres que les moteurs électriques conventionnels.Une attention particulière est accordée aux actionneurs souples à base de polymères élctroactifs qui ont un bon potentiel dans les futures applications robotiques. Parallèlement, on a synthétisé un modèle de la dynamique de l'actionneur et du contrôleur basé sur le modèle (MPC et contrôle optimal) pour un bras anthropomorphe 7 Dofs actionné par une paire antagoniste-agoniste de Muscles Artificiels Pneumatiques (PAM) à chaque articulation. Ce modèle et contrôleur est alors intégré dans l'environnement logiciel développé par l'équipe. En utilisant le bras manipulateur anthropomorphe basé sur PAM et le simulateur numérique, des tests sont effectués afin d'évaluer le potentiel de cet actionneur et de comparer avec les capacités du corps humain. / Design, modeling and control of inherently compliant actuators with a special consideration on agonist- antagonist anthropomorphic configuration" The research aims at the design, modeling and control of inherently compliant actuators for anthropomorphic systems. The first part of the work focuses on the study of various existing designs and look for the possibility of alternative actuators other than the conventional electric motors. Special attention is given to elctroactive polymer based soft actuators which have good potential in future robotic applications. In parallel, a model of the actuator dynamics and the model-based controller (MPC and optimal control) have been synthesized for an anthropomorphic 7 Dofs arm actuated by antagonist-agonist pair of Pneumatic Artificial Muscles (PAMs) at each joint. Such model and controller is then integrated within the software environment developed by the team. Using the PAMs based anthropomorphic manipulator arm and the numerical simulator, tests are done in order to evaluate the potential of this actuator and compare with the human body capabilities.
3

Utilizing Trajectory Optimization in the Training of Neural Network Controllers

Kimball, Nicholas 01 September 2019 (has links) (PDF)
Applying reinforcement learning to control systems enables the use of machine learning to develop elegant and efficient control laws. Coupled with the representational power of neural networks, reinforcement learning algorithms can learn complex policies that can be difficult to emulate using traditional control system design approaches. In this thesis, three different model-free reinforcement learning algorithms, including Monte Carlo Control, REINFORCE with baseline, and Guided Policy Search are compared in simulated, continuous action-space environments. The results show that the Guided Policy Search algorithm is able to learn a desired control policy much faster than the other algorithms. In the inverted pendulum system, it learns an effective policy up to three times faster than the other algorithms. In the cartpole system, it learns an effective policy up to nearly fifteen times faster than the other algorithms.

Page generated in 0.0233 seconds