Abstract

Level-of-service (LOS) is defined in the Highway Capacity Manual (HCM) as a qualitative measure that needs to reflect user perceptions of quality of service, comfort and convenience. Delay has been chosen as the only measure for determining signalized intersection LOS but a behavioral investigation and justification for this choice is lacking.

A web-based survey was conducted to find how road users perceive LOS at signalized intersections. More than 1,300 responses revealed that, besides signal efficiency, left-turn safety was important. Drivers making left-turns preferred to use exclusive left-turn lanes. Drivers and pedestrians would be willing to wait a longer time at signalized intersections in exchange for protected left-turn signals. It was also found that conveniences such as the clarity of pavement markings were important to drivers at signalized intersections. The findings suggest that delay may need to be supplemented by a number of quantifiable attributes of signalized intersections for a representation of LOS that corresponds better to user perceptions.

Three alternative methodologies were developed by applying the potential conflict technique and fuzzy set theory to determine signalized intersection LOS that accounts for user perceptions, the potential conflict analysis method, the fuzzy weighted average method, and the fuzzy logic method. Each method has its appropriate usage. The potential conflict analysis method models the tradeoff between safety and efficiency explicitly by combining delay with collision risk to achieve a LOS measure incorporating safety risk. Rigid LOS thresholds were replaced with fuzzy numbers to make it more consistent with driver perceptions in the fuzzy weighted average method and the fuzzy logic method. To reflect the fuzziness of user perceptions, membership grades were used to assign confidence levels for each LOS category. Moreover, the fuzzy logic method simulates the human reasoning process for the evaluation of signalized intersection LOS.

This dissertation met its goal to find how road users evaluate signalized intersection LOS and to develop methodologies to determine signalized intersection LOS that accounts for user perceptions. Overall, the results of this research improve our knowledge of how road users perceive LOS at signalized intersections and provide useable methodologies of application.