Analysis of Specified and Hypothetical GPS IIIC Integrity for LPV 200 Operations

Curtis A. Shively

Abstract:	Modernized GPS IIIC will broadcast an integrity assured user range accuracy (URA) parameter describing a level of satellite ranging error which will only be exceeded with a small probability (10-8 per hour) without timely notification. Consequently, there is great interest in the potential of GPS IIIC to provide Localizer Performance with Vertical Guidance to a decision altitude of 200 ft (LPV 200), without need for additional augmentation. This paper addresses the potential integrity performance of GPS IIIC to meet the fault-based integrity requirement for LPV 200 operations. This requirement states that the magnitude of the user’s vertical position error in the presence of an undetected satellite range fault cannot exceed 15 m with a probability larger than 10-5. The analysis of achieved integrity performance is based on two alternative concepts: 1) a “specified” concept in which the satellite range error probability is bounded by points (discrete or a continuum) related to a Gaussian probability density function with standard deviation equal to URA and 2) a “hypothetical” monitor concept in which GPS IIIC integrity is modeled to meet a monitor missed detection probability consistent with an assumed prior probability of fault and integrity assured performance. For both concepts, the resulting risk depends on the vertical alert limit (VAL) which the user applies to restrict the satellite geometry and therefore also the translation of any single satellite fault into vertical position error. The analysis finds the largest VAL that will permit the faulted requirement to be met assuming a URA of 0.7 m. To maximize availability it would be desirable to permit the same maximum VAL of 35 m as for LPV 200 operations using the Wide Area Augmentation System (WAAS). For the “specified” concept, if only two discrete points are defined (4.42 and 5.73´URA) as in the current GPS IIIC specifications, the VAL must be limited to about 10.6 m. If, however, a continuous Gaussian bound could be assumed by users, the VAL could be increased to about 19.2 m. For the “hypothetical” integrity performance based on a monitor with Gaussian noise statistics and threshold set for a reasonable false alarm probability, the VAL could be increased to about 24.4 m. Based on these observations it is recommended that: 1) the requirement be revised to describe a continuous Gaussian bound, 2) planned operational tests include data collection at many more points than just the two currently specified and 3) the same analysis concept used for the “hypothetical” monitor be applied to the actual monitor after the detailed monitoring system design of GPS IIIC becomes more extensively defined.
Published in:	Proceedings of the 2010 International Technical Meeting of The Institute of Navigation January 25 - 27, 2010 Catamaran Resort Hotel San Diego, CA
Pages:	975 - 986
Cite this article:	Shively, Curtis A., "Analysis of Specified and Hypothetical GPS IIIC Integrity for LPV 200 Operations," Proceedings of the 2010 International Technical Meeting of The Institute of Navigation, San Diego, CA, January 2010, pp. 975-986.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In