Effect of operational parameters of UAV spraying simulation platform on droplet characteristics and prediction using PSO-ANN for maize crop

• UAV simulation platform enables precise maize spraying parameter optimization indoors • Flow rate, forward speed, and canopy zone strongly impact spray coverage. • ANN-PSO model predicts and optimizes UAV spray parameters with R²=0.897. • Optimal spraying was achieved at 2 L/min, 2500 rpm, 2.5 km/h and at top canopy zone Aerial spraying with unmanned aerial vehicles (UAVs) offers several advantages over conventional ground-based methods but requires precise optimization of spraying parameters. Real-world flight tests are time-intensive, weather-dependent, equipment-risking, and constrained by continuous power needs. UAV spray simulation platforms overcome these limitations by replicating spraying in controlled environments, enabling detailed analysis of deposition, coverage, and droplet density under diverse operational conditions. This study investigates the effect of operational parameters i.e., flow rate, propeller speed, and forward speed on droplet deposition characteristics in maize crops on UAV spraying simulation platform. Using specially developed UAV simulation platform, experiments were conducted on maize plants. Spray coverage, deposition, and droplet density were quantified using water-sensitive papers and image analysis. Results showed that flow rate, forward speed, and canopy zone significantly influenced coverage and deposition, while droplet density was mainly affected by flow rate and canopy zone. Optimal performance was achieved at flow rate of 2.0 L/min, propeller speed of 2500 rpm, and forward speed of 2.5 km/h, particularly in the top canopy zone. At these parameters highest spray coverage (26.84 ± 0.93%, mean ± SE), deposition (2.17 ± 0.24 µL/cm²), and droplet density (152.67 ± 14.28 droplets/cm²) were achieved. Additionally, an artificial neural network coupled with particle swarm optimization was employed to predict and optimize spray parameters, yielding high prediction accuracy (R² = 0.897) and minimal deviation from observed values. This integrated approach highlights the potential of UAV spraying systems, combined with machine learning, to enhance precision pesticide application in maize, improving efficiency and sustainability.