Fieldwork in 1998

Summary reports on fieldwork activities undertaken by members of the MCFE project team and our collaborators during 1998.

Contents

• Canterbury
• Sweetwaters, Kenya, July-September 1998
• Falerii Novi, Lazio, Italy, September 1998
• Castillo de Mulva, Andalusia, Spain, October 1998
• Valloire, French Alps, Easter 1998

Canterbury

Several final year Computer Science and MSc students have undertaken projects based on related topics ranging from differential postprocessing of GPS data to context-aware campus guides.

Sweetwaters, Kenya, July-September 1998

This summer's fieldwork had three main objectives:

1. To perform a field trial of the prototypes with ecologist fieldworkers.
2. To explore how the user interface could be optimised for the fieldwork environment.
3. To develop a rhino footprint identification program.

Each of the these is summarised in a paragraph below.

The Field Trial

Unlike the 1997 field trial which involved supporting animal behaviour observations, the 1998 trial involved supporting an ecologist in a vegetation study. The aim of this vegetation study was to try to establish the damage that elephants were inflicting on fever trees growing along the river that runs through the Sweetwaters game reserve. This involved selecting different sample sites along the river bank and performing a ten nearest trees sample. A central tree is measured followed by the measurements of the 10 nearest trees and their distance from that central tree.

The PalmPilot software provided a convenient and consistent method of recording these measurements in a more efficient manner (much less time was taken in organising and transcribing the recorded data). This particular field trial also pointed to how very useful accurate GPS measurements would be in some fieldwork settings. Unlike the giraffe observations of the year before where location was of secondary concern and a hundred metre accuracy not an issue, in this year's vegetation work location, distance and height were of central importance. The ecologist spent the majority of time on a sample site battling through the undergrowth with tape measures and a compass in order to map out the relative positions and distances between fever trees. Accurate GPS measurements could significantly reduce the time taken in collecting such data, increase the accuracy of the collected data, and make it much easier to relocate the trees at a future date to continue the study (baboons hamper current efforts by stealing the tags!).

The Minimal Attention User Interface (MAUI)

One of the features that had a particularly positive response in our 1997 field trial was the facility to edit numeric data using two hardware buttons on the PalmPilot to increment or decrement the value by a predefined step amount. This highlighted the desire of fieldworkers to be able to operate a data collection device with minimal distraction, in effect maximising their focus on their observational work whilst minimising their attention on the computer interaction. Often it is not necessarily the duration of the interaction that is important but rather how much attention needs to be expended on the device. For example, the numeric buttons were frequently used but did not distract from the observational task at hand, so time was not an issue.

In the 1998 trial we explored this minimal attention model further in the development of a one-handed interface. Our aim was to let the ecologist record their data with only the use of one hand and with a simple interface that could be operated with only the occasional glance. The solution we devised involved a question-by-question mode of data collection rather than the forms we had been previously using, and sectioning the touch screen into large areas that could be operated with the thumb of the hand in which the PalmPilot was held. The resulting interface does not necessarily reduce the amount of time to collect the data (in fact, it may increase it a little) but the user's other hand is free to use binoculars etc. and the attention required to operate the device is greatly reduced. The interface has not been extensively used as yet, but the results of initial tests have been promising.

Rhino Identification

The ecologists' work often involves the identification of animals in order to associate and compare behavioural observations of different individuals. The ecologists were particularly interested in using rhino footprints to identify individuals as it would provide much extra sighting data on these seldom seen animals and perhaps allow for a remote form of tracking. Additionally, local trackers seem often to be able to identify rhinos from their footprints by comparing them with their own hands, but there is a question as to how accurate this is considering the differing soil conditions, etc.

The solution we developed reused the existing PalmPilot programs to collect the rhino footprint data with a additional new program that explored the concept of triggering by calculating and listing the available rhino information notes in order of closeness of match to the current footprint. The user could then select one of the listed rhino notes to view more information about the rhino whose footprint they are possibly inspecting. The triggering process matches using the various dimensions of the footprint and also takes into account the environmental factors which may deviate the measurements, e.g. substrate type, speed of rhino, etc. Home range data is also used to filter out rhinos that were unlikely to have been in the area.

The triggering process is quite flexible, allowing the degree of error that still constitutes a match to be defined for each of the dimensions and permits weightings to be expressed on the scoring of fields. Initial results of using the software look promising, but much more footprint data needs to be collected before any conclusions can be drawn on its effectiveness.

Jason Pascoe, 6th November 1998

Falerii Novi, Lazio, Italy, September 1998

A combination of techniques was applied at two archaeological field survey projects at the Roman town site of Falerii Novi in northern Lazio, Italy. This work incorporated a number of portable computing and specific GPS applications, supplementing geophysical survey using a Geoscan fluxgate gradiometer and topographic survey with a Leica TC805 total station.

The geophysics methodology employed at Falerii Novi, and indeed generally, does not fully utilise the wealth of information that is systematically observed by the geophysics team. Not only can valuable archaeological information be gained from this intensive ground coverage, but additional records can contribute directly to the interpretation of geophysical anomalies. The availability of Palm Pilot computers with predetermined data entry fields allowed this potential resource to be capitalised upon. Using the Stick-E suite of software designed and developed by the MCFE group, templates could be put together to meet these demands and then implemented in the field almost instantly. These enabled the geophysicists to record details regarding surface and environmental conditions, while at the same time producing a spatially-referenced record of surface scatters of archaeological material. By constraining the information to predefined fields these data could be rapidly and usefully ported to a database held on the site Toshiba laptop computer.

The combination of Palm Pilot and Sticke-E software also proved invaluable in the documentation of fieldwalking strategies employed on the site, replacing the need for pen and paper in the (often muddy) field. Since the survey strategy necessitated the precise location of sample points in the field, an uncorrected GPS solution was not appropriate in this context, but it was immediately clear that the Palm Pilot provided a preferable alternative to paper fieldwalking proforma. The supervisor was able to design a recording template and complete "notes" for each unit of survey. This process of detailed sampling of the site was accompanied by a larger scale survey of surface materials. Conducted by a single Roman finds expert, using the Apple Newton MessagePad running the TrackMan and Tracker software and attached Trimble GPS with hat mounted antenna, a reasonable coverage of the 30 hectare site was obtained in 4 days. It is hoped that these results, once differentially processed by Nick Ryan, will contribute to interpretations of town layout and chronology at this site.

The Falerii Novi Hinterland project is committed to the enhancement of field survey methodologies. In particular, techniques that allow for rapid, yet accurate gathering of data are of crucial importance for the future of large scale regional survey. Within the framework of this project, the use of portable computers has proved useful. Specifically, a new technique for palmtop-based fieldwalking was tested at the site. This uses GPS in place of labour-intensive survey methods for the accurate location of fieldwalking units.

The portable computing solutions developed by the MCFE project adapted extremely well to their new uses at Falerii Novi. The archaeologists were quick to see the potential of the technology, and were enthusiastic to learn how to operate the systems. It was widely felt that the use of portable computing in the field has the potential to give some of the responsibility and skill back to the fieldworkers, in an otherwise repetitive daily routine of survey tasks. While the Palm Pilot was limited to use in contexts where the subject of recording was already spatially referenced through traditional survey techniques, the DGPS MessagePad/Trimble unit has a great deal to offer to archaeological field survey, integrating the highly labour intensive two-stage process of survey and recording into a seemless whole. It only remains to see how accurate the readings will be once processed.

The Falerii Novi Hinterland project is based in the Universities of Durham, Southampton and North Carolina. The Falerii Novi project is based in the Universities of Durham and Southampton. Further information regarding the projects is available from Sarah Poppy.

Graeme Earl & Sarah Poppy, 27th November 1998

Castillo de Mulva, Andalusia, Spain, October 1998

The MFCE equipment was also taken on a brief period of fieldwork in Andalusia, conducted by Southampton University research postgraduates Graeme Earl and Sarah Poppy. In particular the MessagePad and Trimble combination was used to record the location from which key photographs had been taken at the well-preserved Roman site of Castillo de Mulva (Roman Munigua), using the Photograph recording template written by the MCFE group. The creation of a well-documented photographic archive is the first stage in the 3D computer reconstruction of Roman town houses at this site. In addition the differential GPS equipment was used to record a well-preserved length of important Roman road, spatial information which is poorly documented elsewhere. It is planned that the processed results of this will be incorporated into a Geographic Information System (GIS) at the University of Southampton.

Graeme Earl & Sarah Poppy, 27th November 1998

Valloire, French Alps, Easter 1998

Jason Dykes, a member of the JTAP funded VFC (Virtual Field Course) project, uses a PalmPilot, Sticke-E software and a Garmin GPS receiver to record the location, bearing, field of view and other information about digital images and video. Jason tells us that "...the wires were a bit heavy and restrictive for snowboarding!", so perhaps this is another application for the MAUI (Minimal Attention User Interface) described above? Below is a picture of the VFC program panoraMap in action, with linked georeferenced panoramic images, multimedia and a map base. The GPS routes on the map load directly from the Garmin download, and you just click on the map and open a file to georeference any media. There is a GUI for selecting media players to associate with file extensions too... then just click on the symbol on the map to play the file.

Once panoramas are entered with FOV and Bearing the software adds arrrows in each of the views which link to other panoramas, so you can tour through the images. The red arrow is a link between the panorama and map views. Panoramas can be scrolled by clicking and dragging the image or the FOV on the map. The VFC site has information on a suite of software for more serious fieldwork applications than Jason's holiday snaps!

	MCFE Home
	Fieldwork in 1996/7