Traffic control measures aim at improving traffic conditions. Traffic control measures have been implemented and tested in practice. One of the most effective control measures is ramp metering. Ramp metering is a traffic control measure which regulates the inflow from the on-ramp to the mainline of the freeway. Ramp metering has been implemented a lot in practice and recently even in a large Field Operational Test (FOT): the PraktijkProef Amsterdam (PPA). Ramp metering uses a certain algorithm to optimize the traffic throughput on the freeway and to prevent the capacity drop. Over the years several different algorithms have been developed and tested.
During the PPA an adaptive controlled algorithm has been used. Adaptive control means that the algorithm is adjusted every time step based on the actual traffic conditions to optimize performance of the ramp metering installation. Adaptive control should improve the throughput of the freeway compared to a normal controlled ramp metering installation.
There are a lot of different adaptive control approaches which have not been tested and developed for ramp metering. One possible way to control an algorithm adaptively is to adjust the control process and see if the controller performs better or worse. In case of better the controller should be adjusted in the same way again, in case of worse the controller should take a step back and be adjusted again. The objective of this thesis is to develop and compare different adaptive control approaches for the ramp metering algorithm towards the current ramp metering algorithm used for the PPA and a situation without any control.