Linearizations for mechanical systems in generalized coordinates

Elliot R. Johnson, Todd David Murphey

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Scopus citations


We describe an algorithm for calculating the linearization of the dynamics for arbitrary constrained mechanical systems in generalized coordinates without using symbolic equations. Linearizations of dynamics are useful tools for controllability and stability analysis and can be used to generate locally stabilizing controllers for linear and non-linear systems. However, the computational expense for finding linearizations of complex mechanical systems is often cited as a limiting factor that prevents their use. Recent work has introduced new methods of calculating the dynamics of arbitrary mechanical systems in generalized coordinates without deriving large, system-specific equations of motion. This paper extends that approach to calculate the linearizations of the dynamics without using the symbolic equations of motion. Using these ideas, it becomes practical to both simulate, analyze, and control more complex mechanical systems without sacrificing the benefits of generalized coordinates. Furthermore, this method addresses systems with closed kinematic chains, constraints, and external non-conservative forcing. The technique is applied to an example system with a closed kinematic chain and the resulting linearization agrees with results found by symbolically differentiating the full equations of motion.

Original languageEnglish (US)
Title of host publicationProceedings of the 2010 American Control Conference, ACC 2010
Number of pages5
StatePublished - Oct 15 2010
Event2010 American Control Conference, ACC 2010 - Baltimore, MD, United States
Duration: Jun 30 2010Jul 2 2010


Other2010 American Control Conference, ACC 2010
Country/TerritoryUnited States
CityBaltimore, MD

ASJC Scopus subject areas

  • Control and Systems Engineering


Dive into the research topics of 'Linearizations for mechanical systems in generalized coordinates'. Together they form a unique fingerprint.

Cite this