M. E. Cannon

An Overview of Multi-Reference Station Methods for Cm-Level Positioning

Over the past few years, a significant amount of research has been conducted on the formulation of carrier phase corrections in order to enhance ambiguity resolution and to increase the distances over which precise positioning can be achieved. Recently the use of a network of multiple GPS reference stations for generating carrier phase-based corrections has emerged with great promise for use in real-time environments. However, little research has been conducted on the distribution of these corrections to potential GPS users located within, and surrounding, the network coverage area. This is an integral part of real-time kinematic DGPS, and it must be adequately addressed before a practical realization of the multireference station concept is implemented. The focus of this paper is to present a comprehensive summary of some of the multiple reference station methods, with specific attention directed toward the correction generation and dissemination processes. More specifically, the various multi-reference station methodologies have been categorized according to their underlying correction generation framework, but will be discussed in terms of the correction dissemination options presented by the various authors. The for main categories of methods investigated in this paper are: (a) partial derivative algorithms, (b) linear interpolation algorithms, (c) condition adjustment algorithms, and (d) virtual reference station methodologies.

A consistency test of airborne GPS using multiple monitor stations

In October 1990, several airborne GPS tests were conducted in the Ottawa region by the Canada Centre for Surveying (CCS) and the Canada Centre for Remote Sensing (CCRS). Ashtech XII receivers were located at up to three monitor stations with baseline lengths to the aircraft ranging from 1–200 km. Approximately two hours of airborne data, collected at a 2 Hz rate, were available for each of the three test days. Post-processing of the differential data was done using the University of Calgarys SEMIKIN package which utilizes a Kalman filter algorithm to estimate both the remote receivers position and velocity. Comparisons were made between the aircraft position and velocity determined from each of the monitor stations to assess the consistency of differential GPS when different reference stations are used. Results show that the degree of consistency is dependent upon the distance to the monitor stations. Agreement at the decimetre-level is achieved in position when the baseline lengths are within 100 km. Agreement in velocity is usually better than 1 cm s−1 (RMS).

Maintaining high accuracy GPS positioning 'on the fly'

A series of airborne GPS (Global Positioning System) tests using the Ashtech XII receivers was used to demonstrate the effect of cycle slips and erroneous data on the positioning accuracy. The overall achievable accuracy is in the order 20 cm; however, errors of several decimeters were detected in some instances. Traverse closures and residual analysis are used to assess the occurrence and magnitude of these errors. The addition of an inertial navigation system in the aircraft, which provides accurate relative positions, is also used to detect cycle slips and outliers in the GPS data. Strategies to achieve consistent accuracies are discussed, and results for different alternatives are presented.<<ETX>>

Assessment of the LandStar Real-Time DGPS Service under Several Operational Conditions

LandStar is a differential global positioning service (DGPS) that provides 24-h real-time positioning for various applications on land, water, and air in North America, Australia, New Zealand, Europe, and Africa. Its focus is on real-time applications requiring a submeter positioning capability such as agriculture, forestry, Geospatial Information Systems (GIS), survey/mapping, and land/vehicular navigation. LandStar uses a Wide Area Network of reference stations to derive DGPS corrections to model the variation of GPS error sources over a large area. These model parameters are used by the Virtual Reference Station processors to calculate standard corrections that are available for all predefined locations in the network. The corrections are transmitted to the user by L-band satellite communication in the standard RTCM SC104 DGPS correction format. This article investigates the performance of the LandStar Mk III system under various operational conditions and assesses its performance in both static and kinematic modes. Four field tests were conducted during 12 months that tested the sysem in clear static and kinematic conditions as well as suboptimal environments associated with low and heavy foliage conditions. Both the accuracy and availability of the system under these conditions is investigated, with an emphasis on whether the above variables are caused by the LandStar system differential corrections, the GPS measurements, or a combination of both.