Design and implementation of a model predictive control algorithm for trajectory planning vehicle environment

Abstract

Driver assistance systems and automate driving functions are being developed and implemented on vehicles in order to improve safety and comfort. Trajectory planning is a cornerstone in this field because it is the gateway for all the decisions related to motion of the car and the safety of the vehicle relies on those decisions. In the present master thesis an algorithm for trajectory planning in dynamic environments is proposed in a model predictive control framework. A linear, varying parameter model is used to simulate the vehicle behavior. It is assumed that a reference trajectory through the static environment, which may not be obstacle-free, is known in advance. The evolution of the dynamic environment and its uncertainty for a determined prediction horizon are estimated based on initial data of the position, orientation and velocity of the obstacles. Taking into account all this information, an optimization problem is formulated whose goal is to generate a collision-free trajectory for the current horizon. This problem is periodically reformulated and fed with new data as new prediction horizons are considered and the desired trajectory is recalculated subject to the new varying environment restrictions.