Title :
Calibration of Virtual Detection Zone and CCTV data for Cell Transmission Model
Author_Institution :
Faculty of Computer Science, Universitas Pelita Harapan, Thamrin Blvd., Lippo Village, Tangerang, 15811, Indonesia
Abstract :
This paper shows that our novel method, Virtual Detection Zone (VDZ) system with CCTV snap shots can provide the empirical data needed to construct Fundamental Diagrams and to calibrate a chosen model. VDZ system is better than the traditional system which uses loop detectors, as it is able to show zero speeds at totally jammed density, which is an essential parameter for macroscopic traffic model. Real data set have been obtained by using only seven VDZ agents who have carried GPS enabled smart phones, together with snapshots from fifteen existing CCTV, on Tangerang to Jakarta highway (a distance of 21 km). The test, carried out without the use of any intrusive sensors, has produced data which is in agreement with previous traditional method. By adopting and modifying Cell Transmission Model or CTM, a variant Macro model, it is shown that simulated speeds for AM cluster generated by CTM can be calibrated to follow real traffic data. Its Mean Absolute Percentage Error or MAPE can be improved from 93.8 to 48.8% and improved further by 5.1% using type test method.
Keywords :
"Sensors","Global Positioning System","Roads","Mathematical model","Vehicles","Data models","Smart phones"
Conference_Titel :
Sensing Technology (ICST), 2015 9th International Conference on
Electronic_ISBN :
2156-8073
DOI :
10.1109/ICSensT.2015.7438475