Optimization of Weighted Voronoi-based Workspace Partitioning for Fleet of Autonomous Tractors: A Simulation Study

The automation of agricultural systems is essential for improving productivity and ensuring sustainable food production, and the introduction of autonomous tractors represents a significant advancement in agricultural automation. In this paper, we propose a method for optimizing workspace partitioning based on a weighted Voronoi diagram for fleet of autonomous tractors. The proposed method (1) acquires an image through UAV (Unmanned Aerial Vehicle) for the entire workspace and generates a node map for partitioning the workspace. (2) To calculate the center point of the workspace to be partitioned on the generated node map, the center point is calculated using the 𝑘-means clustering algorithm. (3) The workspace is differentially partitioned through the voronoi diagram algorithm and the performance-based weights for each tractor. (4) In order to minimize the deviation of task time, the workspace partitioning is optimized by adjusting the weights to reflect their relative importance. The proposed approach was validated through simulation. A comparison of the results before and after the optimization reveals a reduction in task time deviation. The results of the study indicate that optimized partitioning of the workspace can significantly enhance the operational efficiency of fleet of autonomous tractors, thereby improving agricultural productivity.