Common use of SINGLE-LOOP ALGORITHM TEST SUMMARY Clause in Contracts

SINGLE-LOOP ALGORITHM TEST SUMMARY. The algorithm includes a revised region growing method for speed estimation and a method for LV volume estimation based on the Nearest Neighbor Decision rule approach. The revised region growing method identifies data intervals that do not contain any LVs for a particular time period. Volume and occupancy data from these intervals, together with a g factor derived from the mean of the SV population reported by ▇▇▇▇ and Nihan (2003), are used to estimate average vehicle speed for the period. This speed is then used in a distance-weighted Nearest Neighbor routine to estimate the number of LVs present in each interval. The algorithm outperformed the traditional algorithm even when the value for the parameter g used in the traditional algorithm was calibrated with night-time data. Further improvements to the proposed algorithm can be addressed in future research studies. For example, the algorithm’s performance should be tested at the onset and dissipation of congested periods, when the constant average speed assumption is clearly violated. Also, the underlying vehicle distributions used in the proposed algorithm should be tested for spatial and temporal transferability to ensure validity in applying the algorithm to other testing locations.