Register      Login
Exploration Geophysics Exploration Geophysics Society
Journal of the Australian Society of Exploration Geophysicists
Shopping Cart: ( empty )
You are here: Home > Journals > EG > EG07012
RESEARCH ARTICLE
Contents Vol 38(2)

Real-time kinematic tracking of towed AEM birds *

Terence Kratzer 1 Julian Vrbancich 1 2
+ Author Affiliations
- Author Affiliations

1 DSTO Maritime Operations Division, PO Box 44, Pyrmont NSW 2009, Australia.

2 Corresponding author. Email: julian.vrbancich@dsto.defence.gov.au

Exploration Geophysics 38(2) 132-143 https://doi.org/10.1071/EG07012
Submitted: 31 July 2006  Accepted: 2 February 2007   Published: 15 June 2007

Abstract

In the absence of attitude and altitude sensors directly attached to the bird, helicopter airborne electromagnetic (AEM) data are typically interpreted assuming that the sensor bird maintains a fixed attitude as well as fixed vertical and horizontal offsets relative to the helicopter during survey. Laser altimeters fitted to the bird can be used for measuring bird-height over land, but these altimeters do not necessarily function over seawater, and in this case a fixed vertical offset is subtracted from the helicopter altimetry to estimate the height of the bird above sea level. With current navigation technology, these assumptions could be overcome by incorporating suitable altimetry and navigation sensors into AEM systems. We constructed an airborne testing rig to represent an AEM bird and fitted GPS, inertial navigation, and altimetry sensors to accurately measure bird attitude and height above seawater (as required for bathymetric mapping) during typical and atypical AEM-survey flight conditions. Bird height above sea level was measured with radar and laser altimeters, and was also estimated from the GPS receiver height. Bird attitude was obtained from the inertial navigation unit (INU) data and was compared with attitude data derived from a triangular configuration of three GPS antennas. Each antenna was linked to a pair of GPS receivers to allow comparison between dual-frequency, high-fidelity and single-frequency, low-fidelity measurements.

Bird attitude and altimetry measurements were recorded during surveys flown offshore Tickera Bay (Spencer Gulf, SA) and within Broken Bay (Sydney, NSW). These surveys were flown at a maximum altitude of 180 m, with bird roll angles up to about ± 40°. Using dual-frequency GPS receivers, the agreement between heights derived from GPS and laser (radar) altimeter data is typically ~0.3 m (0.6 m). GPS antenna separations computed during flight from measured GPS positions gave agreement to within 0.4% (typically 0.2%) of measured values. The agreement between pitch and roll angles computed from GPS antenna positions and INU measurements was within ~1° and 2° respectively, neglecting the effects of any offsets in the alignment of coordinate axes between the two systems. A comparison of pitch and roll angles obtained from single and dual-frequency GPS receivers showed that the accuracy of pitch and roll angles obtained from single-frequency GPS receivers is generally about ± 2°. The discrepancy in results from the different GPS receivers increases at various points along the flight path. This increase was attributed to a decrease in accuracy in results from the single frequency GPS receivers. Roll and pitch angle profiles show oscillations consistent with harmonic (pendulum) motion of the bird at the end of the tow cable connecting the bird to the helicopter. We recommend the use of inertial navigation interfaced to a single dual-frequency GPS receiver for accurate attitude and position measurements, combined with laser and radar altimetry sensors. The future implications of this study are that we expect to accurately measure attitude and altitude of a towed bird over seawater and that consequently these altimetry and attitude sensors will be implemented on a new helicopter TEM system (SeaTEM) specifically designed for bathymetric mapping.

Key words: AEM, GPS, inertial navigation, airborne tracking, altimetry.


Acknowledgments

We acknowledge the Australian Hydrographic Office for permission to display a portion of chart AUS778, and for supporting this study. J.V. also gratefully acknowledges D. Vandemark (NASA Goddard Flight Center) for helpful discussions concerning the use of the Riegl laser distance meter as a suitable altimeter over seawater, helpful discussions with G. Pearce (Flinders Ports Pty Ltd, South Australia) concerning MSL, LAT, and AHD datum for Wallaroo, and G. Johnston (Geoscience Australia) for providing AHD-geoid data from Seismic Mark SA15 station. We gratefully acknowledge the contributions made by Keith Mathews (Kayar Pty Ltd) towards this study in constructing the airborne test rig, and the reviewers (Richard Lane and Daniel Sattel) for their helpful comments.


References

Annan A. P. , 1983, Effect of differential transmitter/receiver motion on airborne transient EM interpretation: 53rd Annual International Meeting of the Society for Exploration Geophysicists, Expanded Abstracts, 622–623.

Davis, A. C., Macnae, J., and Robb, T., 2006, Pendulum motion observed in HEM systems Exploration Geophysics 37, 355–362.
Johnston G. M. , and Featherstone W. E. , 1998, AUSGEOID98: a new gravimetric geoid for Australia: 24th National Surveying Conference of the Institution of Engineering and Mining Surveyors. http://www.auslig.gov/techpap/iemsgary.htm

Sattel D. , Lane R. , Pears G. , and Vrbancich J. , 2004, Novel ways to process and model GEOTEM data: 17th Geophysical Conference of the Australian Society of Exploration Geophysicists, Extended Abstracts.

Smith, R. S., 2001, Tracking the transmitting-receiving offset in fixed-wing transient EM systems: methodology and applications Exploration Geophysics 32, 14–19.
Vrbancich J. , Kratzer T. , and Boyd G. , 2006, Real-time kinematic tracking of towed AEM birds: Australian Earth Sciences Convention 2006 and 18th Geophysical Conference and Exhibition of the Australian Society of Exploration Geophysicists, Extended Abstracts.

Wolfgram, P., and Vrbancich, J., 2007, Layered earth inversions of AEM bathymetry data incorporating bird attitude and offset – a case study of Torres Strait Exploration Geophysics 38, 144–149.
in Vrbancich et al., 2006, for a description of these logs and associated sampling times). The various processing methods used to provide the GPS position solutions that were relevant to this study are described briefly below, e.g., see Figure 2. (‘Single Point’ and ‘INU Single Point’ solutions are not presented here because the solutions were inferior to those obtained using the other methods.)

  1. ‘Single Point’ uses the C/A (coarse acquisition) pseudorange. It does not use real time kinematic (RTK), differential or INU corrections. (The pseudorange is the apparent distance from satellite to receiver, calculated using signal propagation time only, and contains many sources of error, such as ionospheric delay, tropospheric delay, ephemeris errors, and satellite and receiver clock errors).

  2. ‘INU Single Point’ uses the INU to improve the single point solution precision.

  3. ‘Pseudorange Differential’, also known as Differential GPS (DGPS), employs pseudorange corrections transmitted from a fixed (known) location (i.e., a base station) to improve positional accuracy. It does not require real-time transmissions.

  4. ‘INU Pseudorange Differential’, as above, but also integrates INU data.

  5. ‘L1 Float’ is an RTK solution, which uses carrier-phase information transmitted in real-time from a base station. Carrier-phase measurements are like pseudoranges given in number of wavelengths of the carrier signal between satellite and receiver; consisting of an integer and fractional component. The fractional component, and any change in the integer component, is fairly easy to calculate, but the integer component (known as the carrier-phase ambiguity) is initially unknown, and must be found by analysing many combinations to get the correct solution. L1 float is so called because it uses only the single (L1) carrier, and because the carrier-phase ambiguity is floating (or continuous).

  6. ‘Ionosphere Free Carrier Phase’ is a solution computed by combining L1 and L2 measurements, thereby estimating the delay caused by the ionosphere and removing it, providing a better carrier-phase ambiguity than a normal direct measurement.

  7. ‘Wide Integer’ is a type of RTK solution that a dual frequency (L1/L2) receiver can calculate by using combinations of the phase measurements made at the two frequencies, allowing a more rapid solution than at a single frequency. The carrier-phase ambiguity in this case is no longer floating: it has been fixed to an integer value.

  8. ‘Narrow Float’ is a more accurate solution that the receiver can solve for by refining a Wide Integer solution. The carrier-phase ambiguity is floating, however.

  9. ‘Narrow Integer’ is a solution in which the carrier-phase ambiguity has been computed to a fixed integer. This is potentially the most accurate solution, down to 2 cm.

  10. ‘INU RTK Float’ is similar to ‘Wide/Narrow Float’, but instead of starting with a wide solution, and refining it to get a narrow solution, the solution is initialised from the INU.

  11. ‘INU RTK Fixed’ is similar to ‘Wide/Narrow Integer’, but the solution is initialised from the INU as above in point 10.

GPS and INU data analysis

The post-processing software imports the raw base station GPS data, the base station’s known position, and the raw rover GPS data, and attempts to find a differential positional solution for the rover. After this is complete, the INU data can be added to the differential solution to obtain a higher degree of accuracy, and to obtain attitude data. In order to process the INU data, the velocity profile of the differential GPS solution is interpreted by the user to find a period of relatively little movement, i.e., less than 1 m/s (while on the ground), over which to perform a static alignment of the INU. The INU processing time is set to commence at the start of this period of minimal movement, and an alignment time is provided by the user to cover this period. The alignment time, typically about five minutes, occurs at the commencement of the run.

The process for acquiring positional data before take off is as follows.

  1. The base station GPS coordinate fix is obtained by sampling over a period of at least 12 h, typically 18 to 24 h, before the survey. This data is uploaded via a web base to the AUSPOS Online GPS Processing Service (Geoscience Australia) to receive ‘rapid-turnaround’ (typically 2 days) International Terrestrial Reference Frame (ITRF) and Geodetic Datum of Australia (GDA94) coordinates. The accuracy of the horizontal and vertical coordinates is less than 10 mm and 10–20 mm respectively.

  2. The base station is set to transmit corrections to the rover GPS receivers via a radio link. A HR900 spread spectrum (902–928 MHz) data radio modem (1 W) was used set to 57 600 kbps, with Forward Error Correction (FEC) enabled. Maximum range at Tickera Bay was estimated to be ~10 km using omnidirectional transmit and receive antennas. Range (at 1 W transmitter power) could be increased by using a directional receive antenna, but this is impractical for our application.

  3. The vertical and horizontal displacement of the INU relative to the antenna of the GPS receiver interfaced to the INU is measured (i.e., the ‘lever arm’) and stored in the rover GPS receiver. This is necessary for a real-time solution. The entire rover assembly is then kept stationary for ~5 min, while recording the various logs (see Appendix in Vrbancich et al., 2006; examples of logs can be obtained from the ‘OEM4 Family User Manuals Vols. 1, 2’, available online at http://www.novatel.com). This stationary period ensures sufficient static alignment time when post-processing data. Note that GPS positional solutions are referenced to the location of the GPS antennas, whereas combined GPS-INU positional solutions are referenced to the INU reference frame.

  4. Finally, the rover assembly undergoes a short (~60 s) manoeuvre to allow the GPS receiver interfaced to the INU to validate the lever arm.