Control Strategies for Reducing Ground Tear in Agricultural Vehicles with Overactuated Steering

Sammanfattning

This thesis investigates how two different steering configurations, tire steering andpivot steering, as well as control strategies can be utilized to reduce ground impactcaused by shear forces in different scenarios. The vehicle dynamics for the two steeringconfigurations and their ground interaction on deformable soil is modelled andsimulated using Matlab. A state space formulation for each steering setup is derivedand used for the development of an LQR and MPC controller. The control objectiveis reference tracking of the longitudinal velocity, and front and rear steering angleswith the ultimate goal of reducing excessive shear forces. The developed modelbasedcontrollers utilize input rate penalties to promote smooth control actions, aswell as weighting terms on slip ratio and slip angle, which are directly related tothe generated shear forces. In addition, the MPC formulation incorporates explicitconstraints, allowing tire slip to be directly limited. A PI controller is also developedas a benchmark for comparison with the model-based controllers.Both the steering configurations and developed controllers are evaluated in simulationunder both uniform and asymmetric terrain conditions. The pivot steeringmodel combined with the implemented controller shows a reduction in total shearforce compared to the tire steering model. However, it shows slightly reduced capabilitiesin reference tracking performance compared to tire steering. The modelbased controllers were successfully implemented and shows improved performancecompared with the PI controller in terms of ground impact reduction as a result ofthe input rate limitations. Furthermore, the model based controllers could successfullyincorporate slip ratio weights for reduction of longitudinal shear forces. TheMPC also provided promising results in incorporating weights on slip angle and directconstraints on tire slip. Both steering configurations successfully incorporatesan extended Kalman filter for state estimation, although with reduced performance