Parameter optimisation of human-simulated intelligent controller for a cart-double pendulum system

[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] The Overhead Crane has evolved in scope since its inception in the late 1800's. Its early use as a hoist for material transport is now proceeded by new found applications, such as in the Active Response Gravity Offload System (ARGOS) at the NASA Johnson Space Center. ARGOS is an astronaut training facility designed to simulate reduced gravity environments such as Lunar, Martian, or microgravity. By industry standards, it is essentially a repurposed Overhead Crane; in academia it can be conceptualized as a cart-double pendulum system. Anti-sway control of cart-pendulum systems has been heavily researched; however, these methods are not typically designed for space simulation. The goal of this research is to design a controller that provides both energy and error minimization for the cart-pendulum, so that its payload moves as if it were floating freely in a microgravity environment (according to Newton's 1st law). The Euler-Lagrange equation is used to model the system and an optimal control technique called the [alpha]-shift is used to control the system. Most treatments on optimal linear control do not include the [alpha]-shift, but its addition allows one to stabilize the system faster and provides an extra tuning parameter while maintaining the simplicity of the solution. Numerical experiments show that the [alpha]-shift method significantly improves the cart-pendulum's ability to control its payload; especially for payloads in the cart-double-pendulum case.

Download Full-text

Control of a nonaffine double-pendulum system via dynamic inversion and time-scale separation

2006 American Control Conference ◽

10.1109/acc.2006.1656484 ◽

2006 ◽

Cited By ~ 13

Author(s):

A. Young ◽

Chengyu Cao ◽

N. Hovakimyan ◽

E. Lavretsky

Keyword(s):

Time Scale ◽

Dynamic Inversion ◽

Double Pendulum ◽

Scale Separation ◽

Pendulum System ◽

Time Scale Separation

Download Full-text

Soft-Target-Based Predictive Fuzzy Control for a Cart-Pendulum System with Dynamic Constraints

Journal of Advanced Computational Intelligence and Intelligent Informatics ◽

10.20965/jaciii.2007.p0931 ◽

2007 ◽

Vol 11 (8) ◽

pp. 931-936 ◽

Cited By ~ 2

Author(s):

Yougen Chen ◽

◽

Seiji Yasunobu

Keyword(s):

Self Regulation ◽

Target Setting ◽

Pendulum System ◽

Dynamic Constraints ◽

Intelligent Controller ◽

Stabilization Control ◽

Single Target ◽

Self Adaptation ◽

Soft Target ◽

Target Sets

Human decisions to act are based on broad targets and respond flexibly in different situations. Such, self-adaptation to dynamic constraints is difficult but important for autonomous control. Conventional control usually uses a single target that results in inflexibility in responding to dynamic environments such as changes in constraints. We propose a predictive fuzzy intelligent controller based on soft targets defined as a series of target sets that include many target elements with different satisfaction grades and are converted to target setting knowledge by fuzzy logic. This controller was applied to upswing and stabilization control of a cart-pendulum system with dynamic changing limit positions as constraints to realize situational self-adaptation and target self-regulation. Simulation and experiments demonstrated the feasibility of our soft-target-based intelligent controller.

Download Full-text

Numerical Location of Painlevé Paradox-Associated Jam and Lift-Off in a Double-Pendulum Mechanism

Journal of Computational and Nonlinear Dynamics ◽

10.1115/1.4037033 ◽

2017 ◽

Vol 12 (6) ◽

Cited By ~ 1

Author(s):

Shane J. Burns ◽

Petri T. Piiroinen

Keyword(s):

Numerical Simulation ◽

Mechanical System ◽

Coefficient Of Friction ◽

Double Pendulum ◽

Painlevé Paradox ◽

Forward Solution ◽

Pendulum System ◽

Graphical Technique ◽

Lift Off

In this article, we will introduce the phenomenon known as the Painlevé paradox and further discuss the associated coupled phenomena, jam and lift-off. We analyze under what conditions the Painlevé paradox can occur for a general two-body collision using a framework that can be easily used with a variety of impact laws, however, in order to visualize jam and lift-off in a numerical simulation, we choose to use a recently developed energetic impact law as it is capable of achieving a unique forward solution in time. Further, we will use this framework to derive the criteria under which the Painlevé paradox can occur in a forced double-pendulum mechanical system. First, using a graphical technique, we will show that it is possible to achieve the Painlevé paradox for relatively low coefficient of friction values, and second we will use the energetic impact law to numerically show the occurrence of the Painlevé paradox in the double-pendulum system.

Download Full-text