Traffic congestion has been one of the important issues in our daily life with the development of traffic network. Researchers in the field of traffic control have been working on this problem to improve the efficiency and the stability of traffic operations.
In this thesis, a novel method for traffic flow control is developed, based on ramp metering and speed regulation with an adaptive sliding mode control (ASMC) method and a deadzoned parameter adaptation law. Our strategy is proposed for the traffic environment at a stochastic macroscopic level, taking the influence of the density and speed disturbances into account in traffic dynamic equations. This work is to design a local traffic flow controller using both ramp metering and speed regulation derived from the ASMC, with aim to achieving the desired density and speed of vehicles for the maintenance of the maximum mainline throughput against disturbances in practice. The advantage of this proposed method is in that it can largely improve the traffic flow performance, compared to other traditional methods using only ramp metering, even with ramp storage limitation and disturbances in presence. Furthermore, unlike the conventional sliding mode controller, a prior knowledge of disturbance magnitude is not required during the controller designing process. A stability analysis presented in this article shows that the traffic system based on the proposed traffic flow control method can be guaranteed to be globally asymptotically stable. The validity of the proposed method is tested by varying the traffic situations, i.e., under different initial traffic status, indicating that the proposed control method is sufficient to stabilize traffic flow and is better than both the previously well-known Asservissement Lineaire d'Entree Autoroutiere (ALINEA) strategy and the feedback linearization control (FLC) method.
