Preplanning for high performance autonomous traverse of desert terrain exploiting a priori knowledge to optimize speeds and to detail paths

A. Gutierrez, T. Galatali, J.P. Gonzalez, C. Urmson, and W.L. Whittaker

Back to index

Summary

Describes "preplanning" technique to preview a route before driving. Claims Stanford, Oshkosh, and Ensco used preplanning. USGS data, 1m satellite images, and detailed topology from LADAR scans on a vehicle were used to aid preplanning. Interesting DGC 04 vs 05 statical graph comparision.

Preplanning is not consistant with the spirit of the DGC rules, hence the low rating.

Methods

Waypoints are parsed to produce a path detail using spline interpolation. Additional "fine waypoints" were added based on speed and turn radius by humans. A verification step is performed in iterations. Fine waypoints were 1m apart. Human assign risk level to area of course, which determines speed. Results indicate very small margin of error after preplanning.

Keywords

preplanning

Rating

-1

Bibtex Entry

@article{ debra94,

author = "A. Gutierrez, T. Galatali, J.P. Gonzalez, C. Urmson, and W.L. Whittaker",

title = "Preplanning for high performance autonomous traverse of desert terrain exploiting a priori knowledge to optimize speeds and to detail paths",

journal = "tech. report CMU-RI-TR-05-54",

year = "2005",

url = "http://www.ri.cmu.edu/pubs/pub_5282.html"

}

 

Back to index