Development of a Simulated Annealing-Optimized Adaptive Neuro-Fuzzy Inference System (SA-ANFIS) for Sorghum Seed Planting Parameter Tuning

Abstract

Precision agriculture requires adaptive control systems to optimize planting operations, ensuring consistent planting depth under varying soil and operational conditions. This paper proposes a novel Simulated Annealing-optimized Adaptive Neuro-Fuzzy Inference System (SA-ANFIS) controller for real-time adjustment of planting parameters based on soil moisture and planting speed. The ANFIS framework combines fuzzy logic and neural networks to model nonlinear relationships, while Simulated Annealing (SA) optimizes membership functions and rule bases to enhance control accuracy. Experimental validation demonstrates that the SA-ANFIS controller significantly improves adaptive performance compared to standalone ANFIS controllers, reducing root mean square error (RMSE) by ~18% in dynamic field conditions. The proposed system offers a robust solution for precision planting machinery, enhancing crop yield and resource efficiency. It is, therefore, recommended for autonomous planters