Motion Prediction for Watercraft Operations and Recovery
Abstract
Simple probabilistic approach: Input roll, pitch, roll velocity, and pitch velocity past time history (or whatever relevant variables characterize the system) and nondimensionalize with each variables standard deviation. Search non-dimensional past time history for n neighbours nearest to the point of interest (point of interest being the time from which we wish to approximate forward, and n for this work was selected as 10). Note the actual dimensional roll, pitch, roll velocity and pitch velocity trajectories for the duration of interest immediately following each of the 10 nearest neighbours. Generate 1, 2, and 3 standard deviation (1, 2, 3) envelope curves of predicted motions based upon the mean value 1, 2, 3 at each time step from the neighbour time histories.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 08, 2010
- Accession Number
- ADA528162
Entities
People
- Brook Sherman
- Leigh Mccue
- Zhiliang Xing
Organizations
- Virginia Tech