Agentes de control de vehículos autónomos en entornos urbanos y autovías

Abstract

In the last years, mass-produced vehicle implementations have been done in the eld of intelligent Transportation System (ITS). The Advanced Driver Assistance Systems (ADAS), intelligent infrastructures and autonomous driving maneuvers have signi cantly contributed to the implementation of intelligent systems on the road and in urban areas. Thanks to research done by many groups and projects around the world, as well as the development of new technologies, it is possible to nd safer and more comfortable vehicles. Some examples of these implementations are: Antilock Brake System (ABS), Cruise control (CC), Automatic parking and Electronic Stability Control (ESC), among others. At this time, it is not a utopia to think that, in a close future, autonomous vehicle will coexist with other conventional vehicles, interacting with them. In this thesis, di erent control systems for autonomous vehicles have been developed, both individual and cooperative maneuvers in di erent urban and highway scenarios. First contributions are in the lateral and longitudinal control, using both classical and arti cial intelligent techniques, basically fuzzy and neuro-fuzzy controllers. In the second part of this work, di erent experiments that validate the proposed control system have been described. To realize these experiments, the di erent platforms of the AUTOPIA program have been used. In the same way, studies on the state of the art, as well as management of virtual environments have allowed to validate the results presented in this thesis. All the experiments, in their entirety have been tested in real vehicles, both dedicated test tracks and real roads. This control architecture for autonomous vehicles is independent of the vehicle and the scenario used. Electric and gasoline-propelled vehicles for urban environment, that is: straight and curve segments, two-way streets, roundabouts, blocked roads and commucations between infrastructure and vehicles have been used. Moreover, high speed experiments, using gasoline-propelled vehicles, have been presented. Thanks to these contributions, AUTOPIA team has participated in the rst autonomous vehicle competition at European level: GCDC 2011. The main contribution of this thesis is the cascade lateral control system for autonomous vehicles, which allows to transfer more e ciently the human knowledge for driving in urban environments. Furthermore, this control system is easy to tune, and it has been extended to di erent of maneuvers, such as reverse driving and roundabouts. On the other hand, the neuro-fuzzy longitudinal control system improves the results obtained with classical and fuzzy control, based on the experienced drivers and other new control variables.

Date of Award	2012
Original language	Spanish
Awarding Institution	Universidad Complutense (Madrid)