Scholars Central

Conference abstractPublished 1 Sep 2014

Linear model identification for autonomous sailboat development

Authors: Shawn S Keshmiri, Katrina M Legursky, Richard Hale

Ask about this article

Abstract

Select any words in the title or abstract, or press this and choose a sentence.

Introduction: Sailing yachts have great potential to act as future long-term oceanic observing platforms, yet to date there have not been complete autonomous sailing systems robust enough to handle long term operation in the harsh ocean environment. The basis of control system design is a model capable of describing and capturing the necessary behavior of the system to be controlled. A common method employed in the aerospace industry when designing autonomous autopilots is parameter estimation where a physics-based model dependent upon motion variables, control inputs, and aerodynamic parameters is derived and then full-scale measurements are taken and used to estimate the aerodynamic parameters. This work presents results of using one-shot least squares estimation to fit a five degree of freedom linear state space sailing yacht model dynamic model. Experimental set-up: A Precision 23 sailboat, Avanti, has been outfitted with an adequate sensor system to do parameter estimation. The sensors provide boat speed through the water, apparent wind speed, wind angle, GPS position, velocity, heading, rudder angle, sail angle, Euler angles (roll/heel, pitch, and heading) and rates, and accelerations. Data is collected by a Labview Virtual Instrument (VI) PC interface which uniformly records and timestamps the data. Sailing tests are performed on Clinton Lake in Lawrence, KS. In addition steady state sailing, the data includes maneuvers designed to characterize the dynamic response of the vehicle to individual inputs, including rudder inputs, changes in sail angles, and responses to wind gust and wind shift. The data analyzed here is that of the boat in the medium wind range with the jib foresail and the full mainsail. One-Shot Least squares estimation for linear models: The LS algorithm may be solved in one step for linear systems, thus, computationally it goes very fast and can be applied to several large sets of data and yield estimated models very quickly. The estimated models are simulated in open loop with independent test data, and the fit is evaluated using R2 and the Theil inequality coefficient, UT. A total of five models have an overall fit with R2 > 0.92 and UT < 0.17. These models will be presented and discussed in this paper.

Ask about this article

Answers come only from this article's title, abstract and keywords, each sentence pointing to where it comes from. Your question is sent to OpenAI to write the answer. Scholars Central does not store it.

Try:

Similar by topic

More from this journal

The newest in Journal of Marine Science: Research & Development.

All its articles

Follow

New articles from what you follow go into your email digest, once you agree to alert emails.

  • JournalJournal of Marine Science: Research & Development
  • BylineArticles whose byline prints “Shawn S Keshmiri”
  • BylineArticles whose byline prints “Katrina M Legursky”
  • BylineArticles whose byline prints “Richard Hale”

For authors and organisers

Publish, present and keep up with your field

Find the right journal for your paper and submit your manuscript.

Photo: Pexels