This article needs additional citations for verification. (December 2007) |
COMPASS, also referred to as Freeway Traffic Management System, is a system run by the Ministry of Transportation of Ontario (MTO) to monitor and manage the flow of traffic on various roads (including 400-series highways) in Ontario.
COMPASS uses pairs of in-road sensors to detect the speed and density of traffic flow. This data is fed to a central computer at the MTO Downsview office and analyzed by operators, who also view the feeds of traffic cameras placed along the highways. Changeable Message Signs (CMS) then display messages to motorists on the highways, advising them of upcoming collisions, closures, detours and traffic flow.
The primary algorithm used by the Ministry is known as the McMaster algorithm, designed by Professor Fred Hall of McMaster University, in Hamilton, Ontario. Incident Detection algorithms have also been widely used throughout the COMPASS-enabled area.
Research on new algorithm developments and evaluations is performed at the ITS Centre and Testbed (ICAT), at the Civil Engineering department of the University of Toronto. The ICAT is equipped with direct fibre-optic links to the Ministry of Transportation, and received both traffic camera and loop detector data on a live basis. Visual data can be used to confirm the presence of incidents detected by the various algorithms.
Images from most COMPASS cameras are available online via MTO website.
COMPASS has some dedicated cameras used by MTO Enforcement Officers to monitor and manage truck queues at locations such as the Putman Commercial Vehicle Inspection Station. Images from these cameras are not available online.
Highways with COMPASS cameras:
(Fort Erie, Niagara Falls to St Catharines, Stoney Creek to Toronto)
(London and Windsor, Milton to Ajax, Kingston, Ivy Lea). One of the highest-volume highways in the world
(Vyner to Sarnia)
(Hamilton to Burlington, Oakville to Mississauga)
(Toronto to Markham)
(Niagara Falls)
(Toronto)
(Mississauga to Brampton)
(the Queensway, Ottawa)
(Toronto)
A false alarm for incident detection is not only highly undesirable, but seriously damages the confidence in the detection system. Therefore, a near 100% alarm accuracy is needed. This does not mean that 100% traffic parameter accuracy is required from the traffic sensors; however, the logical commands that analyze the change in traffic parameters need to be selected carefully in order to minimize the probability of false alarms yet detect all major incidents as well as a high percentage of all other incidents. Most importantly, confirmation of incident and evaluation of incident type by manual inspection of a video camera screen is probably the most significant incident detection technique.