Research on path tracking method for rotary tillage ridging operation based on robust model predictive control

• Path-tracking accuracy would degrade in tillage-ridging task due to external disturbance. • The disturbance was identified to be related with force data on the implement. • A control method using a simple model was developed to handle the disturbance. • The method was applied on a tractor platform and verified in field. Tractor is a non-holonomic system whose path-tracking accuracy would degrade when subjected to external disturbance. When using a rotary-tillage-ridging implement for ridging, its interaction with soil would introduce random disturbance to the tractor system, especially on an unleveled ground surface. To achieve accurate navigation in this situation, a robust model predictive (RMPC) method was proposed. The method was based on the bicycle model and the prediction error caused by the simplified model was compensated using a closed-loop structure. Field test results showed that compared with a model predictive control (MPC) method, the mean absolute errors of lateral error and yaw-angle error were reduced by 21% and 24%, respectively.