Path Planning Algorithms for Agricultural Machines

Abstract

If the field plot shape is not rectangular and if it contains obstacles, the coverage path planning problem is hard to solve for a non-omnidirectional machine. Scientists have developed several algorithms to solve this coverage path planning problem, but all of them have pros and cons. If the machines were omnidirectional and turning times were decreased to insignificant, the problem would be quite easy to solve using known robotic path planning methods. Traditional agricultural machines, like tractors, tractor-trailer combinations, self-propelled harvesters and other man-driven machines are slow to turn at headlands. This is the most differentiating property of the problem formulation compared to traditional robotic coverage path planning, which has dealt mainly with omnidirectional kinematics. In this article two different algorithms are presented to solve the coverage path planning problem for agricultural machines. The first algorithm is a higher level algorithm to split a complex shaped field plot to smaller parts is presented. The higher level splitting algorithm is presented in detail in this article. The algorithm can handle any field, including obstacles. The algorithm is based on trapezoidal split, merge and search. The algorithm is suited to any kind of vehicle, which is described with a few