Control Methods For Sustainable Mobility In Multi-Class Freeway Networks

Everyday, economic globalization and climate change give rise to difficult challenges that institutions and government authorities must be able to address. In this context the responsibility of the transport system, in particular the road sector, is far from to be negligible. It is common knowledge that the growing needs of passenger movement and supply of goods have led to an incessant degradation of quality life in terms of congestion, and environmental damage, which is why further efforts toward green and sustainable mobility are essential. Indeed, in spite of recent technological advances in automotive industry, traffic emissions are still far from the ambitious limits imposed to safeguard the human health. Moreover recurrent and non-recurrent congestion, daily afflict the citizens, generating frustrations in the road users that lose confidence on the reliability of traffic systems. Within the vast panorama of surface transportation, freeway system is a interesting case of study where the necessity of several stakeholders converge. The aim of this thesis is therefore to propose models and control strategies in the freeway field, where besides the reduction of congestion (which is a typical goal of traffic control) the sustainable use of networks is addressed, introducing methodologies of traffic emissions reduction. Furthermore, in order to deal with the problems generated by increasingly heterogeneous traffic streams, multi-class models and multi-class control techniques have been developed, with particular attention to the crucial issue of freight transport. To this end, the dynamics of traffic flows is described by extending to the multi-class case the well known second-order macroscopic METANET model. Successively, different control strategies have been explored with the objective of defining specific control actions for each class of vehicles, in order to minimize the emission of pollutants and to smooth jammed traffic conditions. In particular subject of this study has been the application to the multi-class case of the ramp metering control strategy. The ramp metering is an efficient traffic management tool, which regulates the access of flows entering in the freeway through a traffic light installed at the on-ramps. In this manuscript, a local ramp metering strategy has been first of all proposed, that has been obtained adapting a widespread feedback control measure to the multi-class scheme and to the aforementioned environmental issues. Afterward, a more sophisticated control scheme, including a multi-objective nonlinear optimal control problem has been proposed. To seek the solution of the control problem, a specific solution algorithm has been developed. Such algorithm exploits a specific version of the feasible direction algorithm whose effectiveness in freeway context has been widely demonstrated in literature. Finally, a multi-class and multi-objective combined ramp metering and routing control strategy has been proposed to improve performance in the freeway traffic networks. The considered control scheme is a predictive feedback type, i.e. the control computed at each time step depends on the measured system state and on the prediction of the system evolution, considering both a traffic and an emission model. In the study of the aforementioned control strategies, different emission models with characteristics consistent with the purpose of online control, have been adopted. First, a simple average speed emission model has been included in the proposed methodologies, then a new multi-class macroscopic emission model has been developed. The innovative aspect introduced with this model consists in considering the on-ramp emissions, which are explicitly modeled for different traffic scenarios. In the manuscript are included dedicated sections on the analysis of simulation results, in which the effectiveness of the proposed methodologies has been demonstrated.