Improving the Accuracy of Vehicle Reidentification Algorithms by Solving the Assignment Problem

Cetin, Mecit; Nichols, Andrew P

doi:10.3141/2129-01

Cited by 8 publications

(13 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the first stage, for each vehicle in D a match is found in U. This is accomplished by a Bayesian method developed by the authors previously (Cetin and Nichols, 2009). However, the method is modified (simplified) as explained in the subsequent section.…”

Section: The Reidentification Problem and The Proposed Frameworkmentioning

confidence: 99%

“…These measurements can either be the actual physical attributes of vehicles such as length (Coifman and Cassidy, 2002) and axle spacing (Cetin and Nichols, 2009) or some characteristics of the sensor waveform or inductive vehicle signature (Sun et al, 1999). Researchers have developed various methods such as lexicographic optimization (Oh, Ritchie, and Jeng, 2007;Sun et al, 1999), decision trees (Tawfik, Abdulhai, Peng, and Tabib, 2004), to reidentify vehicles.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Bayesian Models for Reidentification of Trucks Over Long Distances on the Basis of Axle Measurement Data

Cetin

Monsere

Nichols

2011

Journal of Intelligent Transportation Systems

Self Cite

View full text Add to dashboard Cite

Vehicle reidentification methods can be used to anonymously match vehicles crossing 2 different locations on the basis of vehicle attribute data. In this article, reidentification methods are developed to match commercial vehicles that cross 2 weigh-in-motion sites in Oregon that are separated by 145 miles. Using vehicle length and axle data as attributes to characterize vehicles, a Bayesian model is developed that uses probability density functions obtained by fitting Gaussian mixture models to a sample data set of matched vehicles. The reidentification model when applied to a test data set (where each downstream vehicle also crosses the upstream site) matches vehicles with an accuracy of 91% when both axle weight and axle spacings data are used. To account for the fact that not all vehicles in the downstream also cross the upstream site, an additional new step is developed to screen mismatched vehicles produced by the algorithm. For this step, several screening methods are developed that allow the user to trade off the total number of matched vehicles and error rate. For evaluating the effectiveness of the screening methods, 2 scenarios are considered. In the first scenario, only common vehicles that cross both the upstream and downstream sites are considered, whereas in the second scenario all downstream vehicles are considered. It is shown that the mismatch error can be reduced to as low as 1% and 5% at the expense of not matching about 25% of the common vehicles (crossing both sites) for the first and second scenarios, respectively.

show abstract

Section: The Reidentification Problem and The Proposed Frameworkmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Bayesian Models for Reidentification of Trucks Over Long Distances on the Basis of Axle Measurement Data

Cetin

Monsere

Nichols

2011

Journal of Intelligent Transportation Systems

Self Cite

View full text Add to dashboard Cite

show abstract

“…Recently, the authors of this report explored the use of axle spacing and axle weight data to reidentify commercial trucks at two WIM stations in Indiana where commercial trucks cross both stations (Cetin and Nichols 2009). They developed matching algorithms based on statistical mixture models and tested the performance of the algorithms on the data from these two WIM stations that are separated by one mile.…”

Section: Literature Reviewmentioning

confidence: 99%

“…These measurements can either be the actual physical attributes of vehicles such as length (Coifman and Cassidy 2002) and axle spacing (Cetin and Nichols 2009) or some characteristics of the sensor waveform or inductive vehicle signature (Sun et al 1999). Researchers have developed various methods, such as lexicographic optimization (Sun et al 1999;Oh et al 2007) and decision trees (Tawfik et al 2004) to re-identify vehicles.…”

Section: Literature Reviewmentioning

confidence: 99%

Exploratory Methods for Truck Re-identification in a Statewide Network Based on Axle Weight and Axle Spacing Data to Enhance Freight Metrics

Monsere¹

2011

View full text Add to dashboard Cite

Type of Report and Period CoveredFinal Report 14. Sponsoring Agency Code Supplementary Notes AbstractThe main objective of this project is to evaluate the feasibility of re-identifying commercial trucks based on vehicle-attribute data automatically collected by sensors installed at traffic data collection stations. To support this work, archived data from weigh-in-motion (WIM) stations in Oregon are used for developing, calibrating, and testing vehicle re-identification algorithms. The vehicle re-identification methods developed in this research consist of two main stages. In the first stage, each vehicle from the downstream station is matched to the most "similar" upstream vehicle by using a Bayesian model. In the second stage, several methods are introduced to screen out those vehicles that cross the downstream site but not the upstream site and to tradeoff accuracy versus the total number of vehicles being matched. These methods involve calculating both the highest and the second highest similarity measures for each vehicle being matched. It is demonstrated that the proposed screening approach improves the accuracy of the re-identification methods significantly. The models are applied to the truck data collected by WIM sensors at three stations in Oregon, which together create two different "links" that are 125 and 145 miles long, respectively. It is observed that the algorithms can match trucks with approximately 90% accuracy while the total number of trucks being matched at this accuracy level is about 95% of the actual common trucks that cross both upstream and downstream sites. These methods allow the user to trade-off the accuracy vs. total vehicles being matched by adjusting a threshold parameter. For example, trucks can be matched with 98% accuracy if one is willing to match about 40% of all common trucks. It is also found that when travel times of vehicles between the upstream and downstream sites exhibit larger variation, mismatch rate increases. Overall, for estimating travel times and origin-destination flows between two WIM sites, the methods developed in this project can be used to effectively match commercial vehicles crossing two data collection sites that are separated by long distances.

show abstract

“…While only 10 to 15 percent of vehicles were recognized from the axle data, the percentage proved high enough to obtain reliable travel time estimations. Recently, Cetin and Nichols (2009) explored the use of axle spacing and axle weight data to re-identify commercial trucks at two WIM stations in Indiana where commercial trucks cross both stations. They developed matching algorithms based on statistical mixture models and tested the performance of the algorithms on the data from these two WIM stations that are separated by one mile.…”

Section: Weight Signature Matchingmentioning

confidence: 99%

Developing Corridor-Level Truck Travel Time Estimates And Other Freight Performance Measures From Archived ITS Data

Monsere¹,

Wolfe²,

Alawakiel³

et al. 2009

View full text Add to dashboard Cite

Sponsoring Agency Code Supplementary Notes AbstractThe objectives of this research were to retrospectively study the feasibility for using truck transponder data to produce freight corridor performance measures (travel times) and real-time traveler information. To support this analysis, weighin-motion data from each of the twenty-two stations in Oregon were assembled, processed, and uploaded in the WIM data archive is housed under the Portland Transportation Archive Listing (PORTAL) umbrella at Portland State University's Intelligent Transportation Systems Lab. Nearly 42,000,000 truck records were successful uploaded to the archive dating back to July 2005. Two separate algorithms necessary for this research were scripted, tested, and validated. The closest stations are 38.3 miles apart; the greatest are 258 miles apart. The first algorithm matched transponders between of all vehicles in a time window between the upstream and downstream stations. The second algorithm filtered these matches for through trucks. The filter was validated by comparing estimated travel times during a winter weather-induced delay. The analysis showed that corridor-level travel times for trucks for 2007 and 2008 could be generated from the archived data. To explore the feasibility using these same data for real-time traveler information, ground truth probe vehicle data were collected. Travel time estimates from the WIM data and the probes were used to establish a simple linear relationship between passenger car and truck performance. It was concluded that the long distances between stations was a primary challenge to directly adapting the WIM data to real-time use. Recommendations were given on increased sensor spacing and filter improvement. Finally, potential performance metrics for station level, matched trucks, and filtered matched truck data were shown 17.

show abstract

Improving the Accuracy of Vehicle Reidentification Algorithms by Solving the Assignment Problem

Cited by 8 publications

References 7 publications

Bayesian Models for Reidentification of Trucks Over Long Distances on the Basis of Axle Measurement Data

Bayesian Models for Reidentification of Trucks Over Long Distances on the Basis of Axle Measurement Data

Exploratory Methods for Truck Re-identification in a Statewide Network Based on Axle Weight and Axle Spacing Data to Enhance Freight Metrics

Developing Corridor-Level Truck Travel Time Estimates And Other Freight Performance Measures From Archived ITS Data

Contact Info

Product

Resources

About