Analyzing Terrain Surfaces to Synthesize and Visualize Optimal-Coverage Tractor Paths for Conservation Farming

There are competing economic goals when plowing a farm field. The rows must be level, to reduce erosion. The rows should also perfectly abut each other, so that no land is wasted on gaps and no fuel is wasted on overlapping areas. With todays GPS-guided tractors, it has become worthwhile to try to produce an optimal tractor path, knowing that a tractor can actually be programmed to follow it. This paper describes the use of visualization methods to analyze a farm field represented by a topologically organized neural network characterized by a shunting neural equation. A complete coverage tractor path is autonomously generated from the dynamic activity landscape of the neural network using GPU programming and framebuffer computing. The effectiveness of the proposed approach is verified through computer simulation.