Statistical Reliability Measures for GPS

Thursday, August 17, 2000 - 3:20pm - 4:15pm
Keller 3-180
Gerard Lachapelle (University of Calgary)
Joint work with S. Ryan.

Reliability refers to the controlability of observations, namely the ability of a system to detect blunders and to estimate the effects of undetectable blunders on the estimated parameters. Internal reliability quantifies the smallest blunder that can be detected through statistical testing of measurement residuals obtained through the least-squares solution of a system with redundant measurements. External reliability quantifies the impact that an undetected blunder can have on the unknown parameters. The fundamental approach for the one blunder case is described, together with numerous examples involving a standalone GPS and a system consisting of GPS augmented with external measurements. The theory is extended to the two simultaneous blunder case, together with additional numerical examples. The results illustrate the problems faced by GPS users requiring a high level of reliability for their positioning and navigation applications, and the advantages and limitations of reliability theory to deal with these problems.