Browse Prior Art Database

Differential Weighting of Prediction Values in an Online Traffic Model

IP.com Disclosure Number: IPCOM000080097D
Original Publication Date: 1973-Oct-01
Included in the Prior Art Database: 2005-Feb-26
Document File: 1 page(s) / 11K

Publishing Venue

IBM

Related People

Cherrill, JG: AUTHOR [+2]

Abstract

A traffic control system employs a model which shows, for each road intersection, the predicted vehicle arrivals over each of a number of periods, for example, 32 four-second periods. The vehicle arrival figures are used, in conjunction with traffic light pattern indications to develop vehicle queue figures, and these are utilized to optimize the traffic light patterns to minimize delay at each intersection.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 100% of the total text.

Page 1 of 1

Differential Weighting of Prediction Values in an Online Traffic Model

A traffic control system employs a model which shows, for each road intersection, the predicted vehicle arrivals over each of a number of periods, for example, 32 four-second periods. The vehicle arrival figures are used, in conjunction with traffic light pattern indications to develop vehicle queue figures, and these are utilized to optimize the traffic light patterns to minimize delay at each intersection.

The model is constructed by projecting predicted patterns of vehicles leaving the intersections, to generate predicted vehicle arrivals at downstream intersections. When used in an online mode, the predicted vehicle arrival figures for the first few periods of each intersection arrival pattern are continuously replaced, by predicted vehicle arrival figures obtained from vehicle sensors located upstream of the intersections.

The arrival figures derived from the sensors are, of course, the most accurate, and the accuracy of each predicted arrival pattern decreases as the prediction period increases.

In order to compensate for this lowered accuracy for the long-term portions of the arrival predictions, the queue date derived from the arrival predictions is differentially weighted before it is used to optimize the light settings, these weightings diminishing from the earliest to the most future predictions.

1