Robust Leader–Follower Control for Cooperative Harvesting Operation of a Tractor-Trailer and a Combine Harvester Considering Confined Space

The cooperative grain unloading of a tractor-trailer and a combine harvester is a key means to improve the efficiency of crop harvesting operations. In view of the problem, this paper proposes a robust leader-follower control method for a tractor-trailer to track a combine harvester in a confined farmland environment for the cooperative harvesting operation. Firstly, considering the model uncertainties, the kinematic and dynamic models of the autonomous tractor-trailer system are established. Subsequently, based on the relative kinematics and nonlinear transformations, a second-order Euler-Lagrange error dynamic model for cooperative harvesting system of a tractor-trailer and a combine harvester is proposed. The effective combination of sliding mode control (SMC) and radial basis function (RBF) neural network compensates for model uncertainties in the control system, and achieves the robust leader-follower control of a tractor-trailer and a combine harvester. Considering the confined operation environment due to irregular farmland boundaries or obstacles, an obstacle avoidance strategy is proposed based on an improved artificial potential field (APF) method to ensure that the tractor-trailer can also pass smoothly even in confined working space. The results demonstrate that the proposed robust leader-follower control method can successfully achieve the cooperative control of a tractor-trailer and a combine harvester in crop harvesting operations. Moreover, in confined farmland environments, the proposed obstacle avoidance strategy can ensure that the tractor-trailer successfully avoids both dynamic and static obstacles while tracking the combine harvester, and the required obstacle avoidance space is smaller.

I. Introduction

Due to limited working time windows, multiple agricultural machines are frequently needed to work constantly at the same time in certain large-scale agricultural productions [1]. For example, in order to improve the efficiency of crop harvesting operations in farmland, it is usually necessary to adopt a simultaneous harvesting and unloading operation plan. Hence, it is necessary to maintain a certain relative distance and relative orientation between the grain truck and the combine harvester to achieve cooperative grain unloading operation. In crop harvesting operations, the grain truck typically uses a tractor-trailer system. This is a nonlinear multi-body system, and its coupling increases the complexity and difficulty of kinematic and dynamic modeling. Unfortunately, the agricultural environment is usually very harsh [2], and the tractor-trailer system are also affected by various uncertain factors, such as unknown vehicle parameters, time-varying mass and moment of inertia, soil resistance, uneven ground, and external disturbances etc., [3]. These make the cooperative operation control of a tractor-trailer and a combine harvester very challenging. In addition, there are static and dynamic obstacles in the agricultural environment. Therefore, considering the obstacle avoidance control of tractor-trailer system is of great significance for the reliable application of the cooperative operation system between a tractor-trailer and a combine harvester.