Local Freeway Ramp Metering using Self-Adjusted Fuzzy Controller

Abstract

A self-adjusted fuzzy local ramp metering strategy is proposed to keep the mainline traffic state and the on-ramp queue length at reasonable levels. The fuzzy ramp metering strategy (FRMS) takes the following variables as inputs: error between desired density and measured density, change-in-error and on-ramp queue length. On-ramp metering flow is decided by these variables. It is difficult to construct fuzzy rules for a three-dimension inputs fuzzy controller based on expert knowledge, so the proposed FRMS generates fuzzy control rules by an analytic expression with correction factors. The correction factors reflect the weights upon linguistic variables of inputs and can be regulated according to actual traffic state of mainline and on-ramp. The proposed FRMS not only simplifies the process of rules definition for a multi-dimension fuzzy controller, but also has function of self-adjusted control rules. To examine the proposed FRMS, a freeway stretch in Los Angeles is simulated with distributed models. The proposed FRMS is also compared with an existing T-S FRMS and PI-ALINEA in the simulation experiments which cover different on-ramp inflow scenarios. Simulation results show the proposed FRMS provides improved adaptation to various scenarios and superiority in striking a balance between the mainline and on-ramp performances.