Improved velocity estimation using single loop detectors Benjamin Coifman * Ohio State University, 470 Hitchcock Hall, 2070 Neil Avenue, Columbus, OH 43210-1275, USA Received 15 November 1999; received in revised form 18 April 2000; accepted 4 May 2000 Abstract This paper develops an improved algorithm for estimating velocity from single loop detector data. Unlikeprecedingworks,thealgorithmissimpleenoughthatitcanbeimplementedusingexistingcontroller hardware. The discussion shows how the bene®ts of this work extend to automated tests of detector data quality at dual loop speed traps. Finally, this paper refutes an earlier study that found conventional single loop velocity estimates are biased. Ó 2001 Elsevier Science Ltd. All rights reserved. Keywords: Trac surveillance; Loop detectors; Velocity estimation; Data screening 1. Introduction Loop detectors are the preeminent vehicle detector for freeway trac surveillance. They are frequently deployed as single detectors, i.e., one loop per lane per detector station. Although single loops have been used for decades, debate continues on how to interpret the measurements and how to calibrate the detectors. In conventional practice, the single loop measurements are verynoisyandmanyresearchershavesoughtsophisticated®lteringmethods,e.g.,Mikhalkinetal. 1972); Pushkar et al. 1994); Dailey 1999). Unfortunately, most of the preceding eorts focused on complicated models without explicitly identifying the sources of error. The earlier works also lose sight of the end goal: to produce an algorithm that can be deployed on a simple processor, such as a Model 170 controller. Thispaperprovidesanewperspectivebyclarifyingthesourceofseveralerrorsandsuggesting ways to reduce the impacts. The body of this work emphasizes velocity estimation, but it has implications for tests of detector data quality as well. The ®rst section reviews the state of the practice for parameter measurement and estimation from single loop detectors. The next section Transportation Research Part A 35 2001) 863±880 www.elsevier.com/locate/tra * Corresponding author. Tel.: +1-614-292-4282; web.: http://www-ceg.eng.ohio-state.edu/coifman. E-mail address: coifman.1@osu.edu B. Coifman). 0965-8564/01/$ - see front matter Ó 2001 Elsevier Science Ltd. All rights reserved. PII:S0965-856400)00028-8