Time study is an important research tool used in comparing productivity of forest harvesting systems across varying conditions. Unfortunately, it has been an expensive tool to apply, involving travel and fieldwork by a crew of technicians plus time in the office to reduce raw data to usable information. In the interests of safety and cost reduction, it would be preferable to have a means of performing autonomous time study that did not involve fieldwork and that produced detailed summaries of machine or system performance over long periods of time. Automated tracking of machine productivity is currently possible on advanced harvesting equipment, notably cut-to-length harvesters, but is not for the predominant tree-length logging systems used in the US south. This paper reports on a data acquisition system used to convert movement and positional data collected using a GPS receiver mounted on tree-length harvesting equipment, primarily skidders, into time study information. The conversion was performed in two stages: reduction of raw position data to a set of measurable, simple events; interpretation of sequences of simple events as machine functions. Simple events were extracted from the raw position data along with time and distance accumulated since the previously occurring event, then passed to the interpretive stage of processing. The stream of simple events was evaluated using a pattern-matching system that combined events into machine functions. Patterns were rules specified using a regular expression syntax that defined machine-specific operational characteristics. Field trials with the data acquisition system on skidders showed it was capable of reproducing measurements obtained from field crews. In two tests measuring total skid cycle time (46 cycles), the automated time study system missed identifying fewer than 10% of the cycles, and of those identified, the difference in cycle time averaged less than 2% (5 s). Although small, the bias was consistent (automated system shorter) and significantly different from 0. Elemental time study was also possible. A three-element ∗ Corresponding author. E-mail addresses: mcdontp@auburn.edu (T.P. McDonald), fultojp@auburn.edu (J.P. Fulton). 0168-1699/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.compag.2005.01.004 20 T.P. McDonald, J.P. Fulton / Computers and Electronics in Agriculture 48 (2005) 19–37 cycle (travel empty, grapple, and travel loaded) was identified in 33 of the 36 cycles with differences in element times between clock and automated system measurements being generally less than 10%. © 2005 Elsevier B.V. All rights reserved.
[1]
Bryce J. Stokes,et al.
Productivity and Costs of Three Harvesting Methods
,
1994
.
[2]
Charles L. A. Clarke,et al.
On the use of regular expressions for searching text
,
1997,
TOPL.
[3]
S. Mattsson,et al.
Real-time monitoring of thinning performance using automatic data collection and GPS
,
1997
.
[4]
Robert D. Grisso,et al.
ANALYSIS OF TRAFFIC PATTERNS AND YIELD MONITOR DATA FOR FIELD EFFICIENCY DETERMINATION
,
2002
.
[5]
Timothy P. McDonald,et al.
Using the global positioning system to map disturbance patterns of forest harvesting machinery
,
2002
.