AN INTEGRATED CONTROL MODEL FOR FREEWAY INTERCHANGES
Schonfeld, Paul M.
MetadataShow full item record
This dissertation proposes an integrated control framework to deal with traffic congestion at freeway interchanges. In the neighborhood of freeway interchanges, there are six potential problems that could cause severe congestion, namely lane-blockage, link-blockage, green time starvation, on-ramp queue spillback to the upstream arterial, off-ramp queue spillback to the upstream freeway segments, and freeway mainline queue spillback to the upstream interchange. The congestion problem around freeway interchanges cannot be solved separately either on the freeways or on the arterials side. To eliminate this congestion, we should balance the delays of freeways and arterials and improve the overall system performance instead of individual subsystem performance. This dissertation proposes an integrated framework which handles interchange congestion according to its severity level with different models. These models can generate effective control strategies to achieve near optimal system performance by balancing the freeway and arterial delays. The following key contributions were made in this dissertation: 1. Formulated the lane-blockage problem between the movements of an arterial intersection approach as an linear program with the proposed sub-cell concept, and proposed an arterial signal optimization model under oversaturated traffic conditions; 2. Formulated the traffic dynamics of a freeway segment with cell-transmission concept, while considering the exit queue effects on its neighboring through lane traffic with the proposed capacity model, which is able to take the lateral friction into account; 3. Developed an integrated control model for multiple freeway interchanges, which can capture the off-ramp spillback, freeway mainline spillback, and arterial lane and link blockage simultaneously; 4. Explored the effectiveness of different solution algorithms (GA, SA, and SA-GA) for the proposed integrated control models, and conducted a statistical goodness check for the proposed algorithms, which has demonstrated the advantages of the proposed model; 5. Conducted intensive numerical experiments for the proposed control models, and compared the performance of the optimized signal timings from the proposed models with those from Transyt-7F by CORSIM simulations. These comparisons have demonstrated the advantages of the proposed models, especially under oversaturated traffic conditions.