California Partners for Advanced Transportation Technology

Dynamic traffic assignment through analytical modeling and optimization has been widely accepted by the IVHS R&D community as a promising traffic control tool for understanding and relieving traffic congestion on conventional highways and city streets. Due to the completely controlled nature of AHS traffic, dynamic assignment of AHS traffic is even more promising. One added dimension of complexity associated with AHS dynamic traffic assignment is lane assignment.  Lane changes, for fully utilizing AHS capacity or for exiting, incur disturbances to and hence reduction of longitudinal flow. Intelligent lane assignment is necessary to ensure a high rate of successful exiting, to another highway or to city streets, while minimizing the disturbances to the longitudinal flow. Merging maneuvers occurring at merge points such as on-ramps, highway-to-highway interchanges and locations where one lane is being dropped may also introduce disturbances to the longitudinal traffic flow on AHS. Intelligent metering and flow control at on-ramps and other merge locations is necessary to maximize the capacity while balancing the need to serve local demands with the need to optimize system throughput. This paper identifies a general class of equations/inequalities to represent the impact of lane changes and merges on AHS longitudinal flow and develop mathematical models for system-optimal dynamic traffic assignment on a two-lane (one direction) automated highway. The amount of impedance to the longitudinal flow depends on the operating scenario and traffic conditions and can be highly stochastic. This paper focuses on deterministic impedance. This is a necessary step towards a complete study of stochastic impedance and the experience gained with the proposed focus is indispensable for studying the stochastic impedance.