CONTACT US: Contact info
|
Citation |
Slavik MM. Transp. Res. Rec. 1985; 1050: 46-52. |
|
Copyright |
(Copyright © 1985, Transportation Research Board, National Research Council, National Academy of Sciences USA, Publisher SAGE Publishing) |
|
DOI |
unavailable |
PMID |
unavailable |
|
Abstract |
If the number-plate-matching technique is used to carry out an origin-destination survey, there is, among other dangers, the risk of finding spurious matches and thereby overestimating the amount of through traffic. A spurious match is a pair of identical entries, one recorded upstream and the other downstream, that belong to two different vehicles. Because it is difficult to calculate the number of spurious matches exactly, a sufficiently accurate approximate method has been introduced and used to assess the magnitude of the problem caused by spurious matches. When survey data are arranged and compared in small blocks, the problem is usually nonexistent or negligible. In the case of bigger blocks of data, however, the overestimation of through traffic may be surprisingly large--far too large to be ignored. The four main ways of combating spurious matches are reducing the number of entries per block; moving the upstream and downstream observation posts closer together; recording more symbols or using letters instead of digits, or both; and locating the survey on a route carrying as much through traffic as possible. Of these four, reducing the number of entries per block appears to be the most powerful strategy. If, for practical reasons, large blocks or few symbols must be used, or if the survey must be carried out over a large distance on a route with little through traffic, then the results obtained should be corrected by breaking up the matches found (M) into spurious (S) and genuine (G) ones and by using G instead of M for the evaluation of through traffic. However, when such surveys are planned, situations that call for a drastic correction of results should be identified and avoided. Language: en |
|
Keywords |
TRAFFIC SURVEYS; STREET TRAFFIC CONTROL; MATHEMATICAL STATISTICS |