Generalized data-driven optimal path planning framework for uniform coverage missions using crop spraying UAVs

Abstract Unmanned aerial vehicle (UAV) based crop spraying has become a popular alternative in the field of precision agriculture. One key goals UAV is achieving spray coverage that as uniform possible to ensure maximum efficacy. Most existing studies literature focus on analysing effects parameters uniformity distribution using experimental studies. However, this work, we propose novel generalized data-driven optimal path-planning framework aimed at finding operational flight (flight speed and pass widths) for lawnmower path plan meet specified rate while ensuring uniformity. The takes model an input computes minimize non-uniformity without making any assumptions type. Furthermore, also neural network structure Gaussian kernel neurons design data. makes no assumption about type UAV, onboard nozzle placement, or parameters. accuracy modelling solution only depends quality training In other words, higher diversity data terms would result more representative consequently improve obtained from proposed optimization framework. present case study demonstrate use test performance via simulation experiments DJI AGRAS-T10 drone. results showed pass-width solutions low forward speeds were similar optimizing positioning nozzles boom sprayer achieve coverage. Whereas, high speeds, was comparatively spread effective over each increased. A discussion contextualized provided highlight salient features limitations