Methodology for a Labor Extensive and Semi-Automated Field Trial Design Using Autoguidance and Conventional Machinery R.N. Jørgensen, C.G. Sørensen, H.T. Søgaard, K. Kristensen, O. Green, S. Christensen The Danish Institute of Agricultural Sciences, Denmark Rasmus.Joergensen@agrsci.dk Abstract The aim of this paper is a methodological proof of concept for a labor extensive and semi- automated field trial design using auto guidance and conventional machinery. A modified strip plot design was created involving two principal treatments, soil tillage and plant density. The different tillage treatments were randomized to columns within groups of columns and the different plant densities were randomized to rows within groups of rows. All applied machinery was guided using a commercial AutoFarm RTK Autosteer™ system. In spring, the soil treatments involved 10 cm deep tillage one week prior to seeding, 5 cm deep tillage immediately before seeding, and no tillage in the stubble. The plant densities were 25, 50, 75, 100, and 150% of the norm for spring barley. In total, 840 net plots of 5.85 m 2 were placed within an area of 150 x 150 meters. The harvest was performed with a dedicated plot harvester collecting total grain weight and one grain sample per plot. Each grain sample were analyzed for moisture, thousand grain weight and cleaning loss. Based on the log files from the AutoFarm RTK Autosteer™ on the combine, the exact area of each net harvested plot was estimated. Due to the number of replicates, Least Significant Differences (LSD) in yields were lower than normally seen in plot trials. Keywords: Cross strip plot trials, auto steering system, GPS Introduction Plot design of trials are a standardized method of investigating different treatments on plants or soils, both in scientific and commercial agricultural research. However, the conduction of plot trials is laborious and costly. Heisel et al. (1999) showed that weed sampling required a large amount of time in terms of manual counting and positioning. Furthermore, manual conducted trials incur the risk of introducing imprecise data acquisition which will compromise the reproducibility and repeatability of the trial results (e.g. Hicks & Turner, 1999). Hence, automated trial establishment, treatment and data acquisition are needed in order to reach an improved reproducibility and repeatability as well as better precision and accuracy. The benefits of automatically guided agricultural field machinery have been well established (Tucker et al., 2002; Dunn et al., 2006). GPS auto-guidance systems have the ability to reduce application overlap, increase operational performance, increase situational awareness, etc. Specifically, the increased potential of controlled repeatability of actions enhances plot design creation and enables an increase in the number of repetitions and hence, increase the value of the statistically analysis of the acquired plot data. Experiences from using auto-steering systems show the working enviroment of the tractor driver is