Confidence interval estimation for path flow estimator

The uncertainty of an origin–destination (O–D) trip table estimate is affected by two factors: (i) the multiplicity of solutions due to the underspecified nature of the problem, and (ii) the errors of traffic counts. In this paper, a confidence interval estimation procedure for path flow estimator (PFE) is developed for assessing the quality of O–D trip tables estimated from traffic counts. The confidence interval estimation consists of two parts: (i) a generalized demand scale (GDS) measure for quantifying the intrinsic underspecified nature of the O–D estimation problem at various spatial levels, and (ii) an error bound to quantify the contribution of input errors (traffic counts) to the estimation results. Numerical results using PFE as the O–D estimator show that the proposed confidence interval estimation procedure is able to separate the two sources of uncertainty in constructing the confidence intervals at various spatial levels. Simulation results also confirm that the proposed quality measure indeed contain the true estimates within the defined confidence intervals.

► A confidence interval estimation procedure is developed for assessing the quality of O–D estimates.
► The confidence interval consists of a generalized demand scale (GDS) measure and an error bound.
► The GDS measure assesses the uncertainty of the estimation results due to multiplicity of solutions.
► The error bound quantifies the contribution of input errors to the estimation results.
► The confidence interval can quantify the uncertainty of the estimation results at various spatial levels.