Autonomous mobility of a fleet of vehicles for precision pesticide application

Abstract

Sub-optimal pesticide application not only results in wastage and low yield but is also associated with harmful effects to the environment. To ensure optimal application of pesticide, a dynamic approach for pesticide management is required. Existing works do not consider the dynamic demand of pesticide and are based on static models. To solve these issues, this paper proposes a technique for the vehicle routing problem (VRP) in pesticide application. The proposed technique is based on dynamic predictive control, i.e., at each time step, the current demand is fed into a model predictive control algorithm which optimizes the vehicular assignment as well as routing to a specific area in an agricultural field. The proposed technique considers charging as well as capacity constraints for the autonomous vehicles. Simulations in MATLAB/Simulink show that the proposed technique not only resulted in at least 20% faster field coverage time but also consumed at least 22% less charge as compared to existing techniques for agricultural VRP.