Modelling and simulation of advanced non-linear autopilot designs

Joseph Brindley, John Counsell, John G Pearce

Research output: Contribution to conferencePaper

193 Downloads (Pure)

Abstract

This paper presents the simulation in ESL of a non-linear 6 degree-of-freedom missile model with an advanced, non-linear, multivariable autopilot designed using Rate Actuated Inverse Dynamics (RAID) methods. High performance control of non-linear systems requires the design of advanced, non-linear control systems, such as those used in autopilot design. Traditional linear control system design and analysis techniques are not sufficient for non-linear systems and current non-linear analysis methods are extremely limited. Therefore, the only method available to fully assess the performance of non-linear controller designs is simulation of the non-linear system. For this reason it is an essential part of the analysis and design process of these types of controllers. Non-linear dynamics can be continuous or discontinuous, the aerodynamics of a missile are non-linear but since they are continuous they do not represent a simulation challenge. However, there are multiple sets of discontinuous dynamics present in both the missile control surface model and the autopilot which can lead to multiple discontinuities being reached simultaneously, providing a challenging modeling exercise. The paper demonstrates how this kind of behavior can be successfully modeled and simulated within ESL using a simple switching logic.
Original languageEnglish
PagesSession 9: Control & Structure
Number of pages6
Publication statusPublished - 11 Jul 2010
EventGrand Challenges in Modelling and Simulation Conference - Ottawa, Canada
Duration: 11 Jul 201014 Jul 2010

Conference

ConferenceGrand Challenges in Modelling and Simulation Conference
Country/TerritoryCanada
CityOttawa
Period11/07/1014/07/10

Keywords

  • modelling
  • simulation
  • non-linear
  • autopilot design

Fingerprint

Dive into the research topics of 'Modelling and simulation of advanced non-linear autopilot designs'. Together they form a unique fingerprint.

Cite this