Datasets for Statistical Analysis

Tractor guidance systems

Keywords:

Categories:

science and engineering;
regression, time series

Description:

The data are taken from an experiment conducted by Mr Rodger Hill under the supervision of Mr Ivan Wolski. Two receivers ("Beeline" and "Trimble") were placed on the equipment and used GPS technology, while (simulating the) ploughing of a paddock.

The Beeline system was used to autonomously guide the tractor. The Trimble system was used to track the path of the tractor (or the plough in Experiment 3).

The three data files are for three variations of the experiment:

• In the first experiment, the Trimble receiver was placed on the tractor while only the tractor travelled up and down the paddock;
• In the second experiment, the Trimble receiver was again placed on the tractor while the tractor pulled an implement (a chisel plough) up and down the paddock;
• In the third experiment, the Trimble receiver was placed on the implement while the tractor pulled the implement (a plough) up and down the paddock (the Beeline remained on the tractor to auto-guide the tractor).

The data give the offsets (in metres) from the intended path of the tractor or implement. The data are recorded approximately every second with the Beeline receiver, and more frequently with the Trimble. In any case, the measurements are taken at fixed intervals.

The data are recorded by Swath: A swath is a run up or down the paddock between turns of the tractor at either end of the paddock.

There are a number of questions of interest:

• One assumes that placing an implement on the tractor, especially a plough, might affect the accuracy and/or precision. Can this be verified and quantified?
• Similarly, one might assume that placing the receiver on the implement might affect accuracy and/or precision differently than when it is on the tractor pulling an implement. Can this be verified and quantified?
• There is probably autocorrelation in the data which may affect any modelling. Does the accuracy decrease over time, for example?
• Does the swath have any effect? For example, does the accuracy depend on the direction of the tractor? (For swaths 1, 3, and 5, for example, the tractor travels in the opposite direction to swaths 2, 4, 6.)

For more information about the data and experiment, see the reference below.

Variables:

Receiver	the receiver used: Either Beeline or Trimble
Swath	the swath number of the tractor's path
Offset	the offset from the intended path of the tractor in metres

Data Quality:

There are no missing values. For more infomration about the data collection, see the reference below.

Source:

From Mr Rodger Hill who conducted the experiment as part of his project in the Bachelor of Engineering degree at the University of Southern Queensland.

Notes:

References:

Dunn, Peter K., Powierski, A., and Hill, R. Statistical evaluation of data from tractor guidance systems. Accepted for publication in Precision Agriculture.