Diseño De Un Mpc Para El Seguimiento De Una Trayectoria De Un Vehículo Articulado

Abstract

In Colombia, the mass passenger transportation systems of some main cities have articulated vehicles in their fleet, which travel from an origin to a specific destination. After they have finished their workday, these vehicles must be parked in a workshop yard in a small space. This action has been identified as repetitive and must be performed at a low speed. This is why, to reduce stress and fatigue to the drivers of these vehicles in the face of these monotonous events, this work describes a case study related to the optimal planning of the trajectory of an articulated vehicle, for which it has been taken as a reference the mass transportation system of the city of Barranquilla: Transmetro. The parking scenario (highest yard) has been recreated in MATLAB with their respective fixed obstacles: walls or other vehicles of equal dimensions. The trajectory from the initial point to the end has a high degree of complexity, which is why the required equality and inequality restrictions have been configured. To find the route, a multistage Model Predictive Controller (MPC) for nonlinear systems has been used, which recognizes these obstacles as collision objects. Random uncertainty has been added to the initial values of the initial position, in addition to modifying the relative weighting of its cost function that depends on the system inputs: the longitudinal speed and the angle of rotation, so that the simulation is carried out in Closed-Loop and thus its optimization will be ensured.