The agricultural industry is facing a major technological change with autonomousvehicles in focus. This follows the global trend, where the interest lies in increas-ing production, while reducing costs with the help of automation. Consideringthe vast amount of different agricultural machines on the market today, the pro-cess of automating these machines is long and needs to start on one machine.This thesis covers the process of developing an automatic control system for aseedbed tine harrow.

The seedbed tine harrow cultivates the soil at a certain depth in preparationfor planting. The different functions on the harrow are today manually controlledfrom the cab of the tractor, which means that the farmer must constantly moni-tor the process. The proposed control system uses radar sensors to measure andhydraulic systems to control the harrowing depth and the crossboards. The de-velopment of the control system consists of modeling the harrow, creating a sim-ulation environment, choosing a filtering strategy, and testing different controlalgorithms.

The resulting control algorithm, implemented and tested on the harrow, con-sisted of a Kalman filter with separate PD-controllers for each function, the har-rowing depth, and the angle of the crossboards. The crossboard controllers usean additional feedforward control from measured disturbance. The thesis alsoexplores a set of experimental control algorithms, for instance, cascade control.These are not possible to implement on this generation of the harrow but showpromising potential from simulation.