A new yaw dynamic model for improved high speed control of a farm tractor

This paper presents the system identification of a new model for the farm tractor's yaw dynamics in order to improve automatic control at higher speeds and understand controller limitations from neglecting these dynamics. As speed increases, higher order models are required to maintain accurate lateral control of the vehicle. Neglecting these dynamics can cause the controller to become unstable at the bandwidths required for accurate control at higher speeds. The yaw dynamic model, which is found to be dominated by a second order response, is identified for multiple speeds to determine the effect of velocity on the model. The second order yaw dynamics cannot be represented by the traditional bicycle model. An analytical derivation shows that the model characteristics can, however, be captured by a model consisting of a significant (non-negligible) relaxation length in the front tire. Experimental results are presented showing that the new yaw dynamic model can provide lateral control of the tractor to within 4 cm (1sigma) at speeds up to 8 m/s. These results are shown to be an improvement, at high speeds, over controllers based on models (such as a kinematic model) previously used for control of farm equipment.