Optimal multi-robot path planning for pesticide spraying in agricultural fields

Abstract

In this paper, we formulate the problem of spraying pesticides in agricultural fields using robots with constraints on the capacity of pesticides that can be carried and the objective of minimizing the distance travelled. We allow different levels of infection in different regions and assume that the amount of infection and the corresponding amount of pesticide to be sprayed is known a priori. However, since, there is a bound on the amount of pesticide that can be carried, the robots may need to visit a refilling station. The problem can be reduced to a multiple traveling salesmen problem with certain constraints on their travel as guided by the pesticide carrying capacity of the robots. We first perform a graph transformation, that removes these constraints and instead reduces to a “generalized” multiple traveling salesmen problem, where the salesmen are not required to visit each city, but only one city in each group of cities. Finally, we reduce the problem to a mixed integer linear programming problem. We provide simulation results as well as experimental results to demonstrate the feasibility of our approach.