Feedlot Adoption - bunk scanner field trials

Summary

Feed bunk management is the process of determining feed allocation for pens of feedlot cattle for a 24-hour feeding cycle. Objectives of bunk management focus on consistently maximising feed intake, whilst minimising feed wastage and digestive disorders (bloat and acidosis).

Bunk calling is a critical input for this process, and the human callers' actions directly determine feed intake and carcase weight gain of pens of feedlot cattle.

With advancements in mapping, sensors, and robotics technologies, it is now feasible to automate feed bunk management. Enabled by levy-funds, our recent MLA project B.FLT.0166 achieved world-first determinations of the precision and accuracy of humans when predicting feed remaining in feed bunks versus a bespoke prototype automatic bunk calling solution based on appropriate technologies. The prototype's performance was assessed through a rigorous and varied experimental process, under non-operational conditions, providing highly repeatable and accurate predictions for all feed remaining quantities assessed. Moreover, the prototype system was consistently more accurate and precise than human predictions.
Feedlot trials were subsequently undertaken to assess the prototype system against humans in an operational environment under normal conditions. Due to logistics of mapping empty feed bunks in the commercial feedlot, feed remaining predictions were determined through post-processing of lidar point-cloud data to complete project B.FLT.0166.

To improve likelihoods of acceptance and adoption by all stakeholders for the prototype automatic bunk calling system, further in-field experiments were required across three diverse feedlot operations with different locations, bunk designs, and road conditions to validate the prototype system's technology robustness. This report presents the outcomes of these experiments and demonstrates that under normal operating conditions the prototype system predicted feed remaining in bunks accurately and repeatedly, outperforming human callers in both criteria, thereby representing a high-value outcome for the Australian red meat industry.flt