Loran-C User Position Software (LUPS) Navigation Performance with the August 2001 Cross-Country/Alaska Flight Test Data Jaime Y. Cruz and Robert Stoeckly Illgen Simulation Technologies, Inc. Email: firstname.lastname@example.org or email@example.com Presented by R. Stoeckly at ILA31 30 October 2002
Background • Effort is part of the overall program to show that the Loran-C component of a GPS/Loran system can meet Non-Precision Approach (RNP-0.3) requirements during loss of the GPS signal • LUPS is a Loran-C navigation software with key features: • Propagation delay model based on the FCC M3 ground conductivity database with extension to Alaska • Measurement fault detection and exclusion (FDE) techniques • Optimal weighting of all-in-view signals • LUPS was used to post-process SatMate and DDC receiver data from the August 2001 cross-country and Alaska flight tests conducted by the FAA Technical Center • Aircraft GPS positions used as truth reference • Results in terms of accuracy, integrity, continuity, availability
Error Models • Measurement noise • Inversely proportional to the measured signal-to-noise amplitude ratio • Standard deviation at 0 dB SNR varies from 43 to 56 ns (13 to 17 m), depending on GRI • Propagation delay spatial error • Propagation delay error due to seasonal variation • Transmitter clock error • Standard error of master clocks with respect to UTC issmaster =1100 ns(330 m)(based on differences between time-of-arrival measurements of dual-rated signals) • Error due to other sources:sother = 130 ns(40 m)
Conductivity (S/m) 0.003 0.010 0.030 5.000 Propagation Delay Error Models Northeast, Midwest, Great Plains South, Alaska West Coast Rocky Mountains Standard error (ns) LUPS (red) and Wenzel and LORIPP (blue) models for error in time-averaged ASF (left) and error due to seasonal variation of ASF (right). The Wenzel model for ASF error is 0.15*ASF, plotted for four conductivity values.
Signal Edit Criteria • Signal reasonability edit criteria • U.S. Coast Guard Academy DDC receiver • -10 < ECD < 10 msec • SNR > -15 dB • Locus SatMate receiver • Use only signals not flagged by the receiver • Use only signals used in at least one of the SatMate-native solutions
LUPS Navigation Results(using SatMate receiver data) EASTERN CONUS 95% Error = 234 m Max Error = 337 m Ave # signals = 17.6 WESTERN CONUS 95% Error = 630 m Max Error = 854 m Ave # signals = 9.3
LUPS Navigation Results(smaster = 100 ns -> wrong model!) EASTERN CONUS 95% Error = 667 m Max Error = 1332 m Ave # signals = 11.8 WESTERN CONUS 95% Error = 667 m Max Error = 1654 m Ave # signals = 8.0
LUPS vs. SatMate-Native Navigation Errors Ave. # signals used SatMate: 11 LUPS: 19 Atlantic City to Sioux Falls
LUPS vs. DDC-NativeNavigation Errors Ave. # signals used DDC: 7.2 LUPS: 7.3 Atlantic City to Sioux Falls
LUPS Navigation Results(Alaskan flights, SatMate data) ALASKAN FLIGHTS 95% Error = 881 m Ave # signals used = 7.1 (2.9 chains, 5.1 stations) Ave # signals available = 8.6 • Seattle to Juneau • Juneau to Anchorage • Anchorage to Sitka • Sitka to Sacramento
Causes of Difficulty in Alaska • Possible causes of the weak performance in Alaska: • Use of constant conductivity in Alaska • Low estimate of ASF error in Alaska • Apparent miscalibration of emission delay for secondary(ies) at Narrow Cape (or Port Clarence ?) • Use of smooth-earth propagation model (Millington-Pressey) in rugged Alaskan terrain • Weak geometry in SE Alaskan panhandle (NE-SW direction) • Small number of signals received
Summary • Current observing geometry and LUPS design yield RNP-0.3 and RNP-0.5 level of performance in eastern and western CONUS, respectively • Key performance parameters • Minimum of 10 signals in good geometry • Use of conductivity database, FDE techniques, optimal weighting of observations • Fault-mode integrity will improve significantly with the planned change to control synchronization of master clocks by Time of Transmission Monitors (TTM) • Alaskan navigation needs more investigation
Recommendation • Further LUPS development could include: • Implement and exercise error and integrity models developed by Loran Integrity Performance Panel (LORIPP) • Improve ASF and ASF-error models to support planned ASF calibrations by the FAA • Improve conductivity database in Alaska and Canada using the FCC R2 database • Develop visualization software for using LUPS diagnostic output and presentation of results.
Chain Timing Biases (from TTM data) Chain timing biases based on USCG-NAVCEN Time-of-Transmission Monitor measurements. *Timing biases for Canadian chains are estimates from measured dual-rate TOA measurements.