The research defines an optimization method for locating vehicle detectors within an urban network. Locating traffic detectors within a network is currently based on a haphazard method while adaptive signal control systems incorporate saturation detection throughout a network. However, the cost of such complete detection coverage renders adaptive signal control expensive. The effects of network congestion, link traffic flow level, detector coverage and link location are evaluated for use with a flow estimation model.; A 20-intersection network located in downtown Salt Lake City, Utah provides the test network. A Monte Carlo simulation, validated with one week of observed network flow information, is utilized to estimate link and turning movement flows and develop the sets of “known” flows. These “known” flows provide the baseline data to determine the affects of different detection placement strategies. The Monte Carlo simulation allows investigation of a wide range of flow volumes that would normally be difficult and expensive to collect.; While supported with theoretical supposition, the most compelling support for this work is the enumeration process that has investigated over 5,000 modeling runs for a range of network flows. A systematic evaluation of detecting individual links to determine the impacts of detector location placement on the overall model performance provides a method for determining the relative relation between flow, congestion and link location. The systematic approach was selected over the more elegant dynamic optimization techniques in order to ensure global optimal feasibility and eliminate the potential for local optima solution.; The result is a “Utility Function” that allows each link in a network to be ranked by detection importance based on a relationship that is a function of link flow and location rating within the network. The Utility Function places the average link within 10% of its ranking based on enumeration modeling. Testing various detection patterns, using multiple detectors, supports the Utility Function results.; This research provides a tool for helping transportation engineers locate vehicle detection with the specific application of estimating flows in support of a real-time adaptive
展开▼
