Evaluation of On-Ramp Control Algorithms

Evaluation of On-Ramp Control Algorithms
Michael Zhang, Taewan Kim, Xiaojian Nie,
Wenlong Jin, Lianyu Chu, Will Recker

Ramp metering is the application of control devices such as metering signals to limit the number of vehicles entering a freeway. The fundamental philosophy of ramp metering is that the corridor can maintain its optimal operation by regulating the freeway demand to be under its capacity. Maintaining the optimal operation of the corridor should provide congestion avoidance and accordingly travel time savings.

Our research project had three objectives:
• review existing ramp metering algorithms and choose a few attractive ones for further evaluation,
• develop a ramp metering evaluation framework using microscopic simulation, and
• compare the performances of the selected algorithms and make recommendations about future developments and field tests of ramp metering systems.

Our evaluation study found that:
• Ramp metering reduces total vehicle travel time up to 7% compared with no metering. The effectiveness of a ramp control algorithm depends on the level of traffic demand. As traffic demand increases, ramp metering tends to be more effective in reducing system travel time.
• No significant performance differences exist among ALINEA, modified Bottleneck, modified SWARM with 1 time-step-ahead prediction, and Zone algorithms under the tested scenarios.
• Modified SWARM with five-step-ahead prediction has the poorest performance of all tested algorithms due to the inaccuracy of the five-step-ahead prediction model. This indicates a good traffic prediction is the key to SWARM's performance.
• Coordinated ramp metering algorithms do not necessarily perform better than local control algorithms if some of their key parameters are not well calibrated. Well tuned parameters are critical for the good ramp metering performance.
• Ramp metering performance and parameter values are non-linearly related. There is a broad range of parameter values over which ramp metering performance does not change significantly. Outside of this range, however, ramp metering performance deteriorates quickly.
• Ramp metering seems to be more effective under certain demand patterns than others.

download full report

previous featured research

PATH is a collaboration between the California Department of Transportation (Caltrans), the University of California, other public and private academic institutions, and private industry. PATH's mission: applying advanced technology to increase highway capacity and safety, and to reduce traffic congestion, air pollution, and energy consumption.

Visitors since 1996: