Comparison of Low-Cost Computer Algorithms for Global Positioning System Users.

摘要

The Global Positioning System (GPS) is a navigation system which relies on range and range-rate measurements between satellites and the GPS-user in order to determine his position and velocity. Using extensive support equipment, GPS is anticipated to achieve extremely accurate results. However, since many users will not require high accuracies, this study compares the performance of three simulated computer algorithms for possible low-cost use. The three approaches to processing the measurement information are (1) a direct solution of the nonlinear measurement equations, (2) linearization of the equations prior to solution, and (3) a Kalman filter solution based on the linearized model. Velocity information was used to continually update position estimates (dead-reckoning) for the first two approaches. The Kalman filter was modelled with fixed covariance matrices. The approaches were tested against a 'real world' dynamic example (C5A) flight simulation) with error sources (receiver errors, atmospheric delay errors, etc.) modelled as gaussian white sequences.