Live video confers enormous advantage in the field of expressway incident detection. When an episode occurs, operators can see what is happening (type), where (location), how many vehicles are involved (severity), whether there are casualties and the overall situation.

Merging collateral data with imagery facilitates alarm verification. And, swifter discovery leads to faster intervention, avoidance of secondary collisions and consequential lane closures and delays. Technical approaches to automatic incident detection are divided into two basic categories: those employing a single sensor (station) versus comparing readings from two or more spatially separated stations.

While the comparison method usually is preferred over a single sensor because the latter tends toward excessive false alarms, the former depends on inter-station communication, increases cost and may reduce reliability. Algorithms which buttress incident detection are grouped into four general classes.

Comparative Algorithms attempt to recognize and differentiate unusual traffic patterns – stopped, slow-moving or wrong-way vehicles or dropped cargo – from normal conditions.

They are founded on the principle that any incident causes an increase in detector reporting levels upstream and a simultaneous downstream decrease. By evaluating certain parameters (volume, occupancy or speed) against pre-selected norms, an alarm is triggered when a measured value exceeds the preset threshold.

As the name implies, Statistical Algorithms utilize mathematical techniques to determine whether observed data diverge from estimated or predicted values. For example, lane occupancy mean and standard deviation can be calculated at one-minute intervals. An incident would be declared if current figures differ considerably from those for two prior, successive time periods or when they surpass a given setting.

Time Series/Filtering Algorithms analyze – or smooth – raw data over time to eliminate short-duration disturbances such as random fluctuations, traffic pulses and compression waves. Processed data are compared to forecasted values. When the discrepancy is significant, an incident is signaled.

Complex flow theories are used to describe and predict behavior by Traffic Modeling Algorithms. Again, when the actual situation departs from projected conditions, an incident is proclaimed.

Three effectiveness criteria are employed in assessing and cross-comparing performance of incident detection algorithms. All are indicators of relative accuracy:

Detection rate, customarily expressed as a percentage, is the ratio of reported incidents to all incidents, occurring over a specified interval;

False alarm rate, also expressed as a percentage, is the number of erroneous warnings versus all decisions made by the system under non-incident conditions during a fixed time period; and,

Mean time to detect is the average amount of time for a system to ascertain and advise of a legitimate incident.

Systems Deployed

It is important to realize that manufacturers decide which technologies and algorithms to embed into their products. Accordingly, we look at a sampling of commercial incident detection systems, each with its attendant underpinnings:

Developed in partnership with the Minnesota Department of Transportation, ADDCO Inc.’s (Saint Paul, Minn.) Virtual Transportation Operations Center (VTOC) software provides complete traffic management with data exchange over the Internet. Customized VTOC units collect data around-the-clock from a variety of sensors and CCTV cameras.

Diamond Consulting Services Limited (Aylesbury, Buckinghamshire, U.K.) designs the Idris Incident Detection System to give road hazard warnings, allowing traffic management/alarm systems to be deployed and help avoid follow-on episodes.

Data from outstations are combined to determine slow-moving traffic, an individual slow vehicle, congestion, vehicles traveling in the wrong direction or a single stopped vehicle anywhere along outfitted road sections.

Econolite Control Products, Inc. (Anaheim, Calif.) promotes an optional incident management module for their Pyramids system. Links within Pyramids can be assigned speed thresholds which trigger pop-up messages if an incident occurs (based on vehicle speed, derived from volume and occupancy). An event alert can be used to pre-position a remote-controlled camera to the appropriate region-of-interest.

EIS Electronic Integrated Systems Inc. (Toronto, Ontario, Canada) advances their Freeway Traffic Management System (FTMS). FTMS is a PC-based, low-cost, quick-deployment solution for automatic incident recognition – utilizing multiple Remote Traffic Microwave Sensors in side- or forward-looking configurations – with 0.5-km (0.31-mile) spacing between stations.

Data are collected at thirty-second intervals, displayed on several screens and analyzed in realtime. Detected incidents generate an alarm, are portrayed graphically, cue video cameras and prompt the operator to file an accounting, which is logged for subsequent review.

Excel Technology Group (Seventeen Mile Rocks, Queensland, Australia) offers incident detection/vehicle classification with supplemental functions which include vehicular volume, occupancy, speed and headway. Realtime incident data for traffic engineers are returned simultaneously with statistical information (vehicle length, weight, speed, usage pattern).

Autoscope Video Vehicle Detection Systems (Image Sensing Systems, Inc., Saint Paul, Minn.) automatically identify occurrences in tunnels and on open highways. The technology – relying primarily on vehicular speed – delivers a high-performance alternative to magneto-inductive loops and other approaches for junction control, incident detection and surveillance applications.

Iteris Inc. (Anaheim, Calif.) manufactures and sells their Vantage Video Detection Systems for highway management and intersection control. Vantage also imparts an information collection resource which expands Iteris’ capabilities into video incident detection and data measurement applications (vehicle presence, count, speed, occupancy).

The Peek Traffic (Palmetto, Fla.) subsidiary of Quixote Transportation Safety, Inc. announces VideoTrak Plus for cost-effective and accurate vehicle and incident detection and data collection. VideoTrak’s multi-resolution processor affords realtime video analysis to identify traffic conditions, adapt to various environments, monitor image quality and verify proper camera operation.

Patented tracking algorithms perform detection, while specialized shadow filtering, image stabilization and automatic gain adjustment minimize false positives as well as false negative calls.

Siemens Traffic Controls Limited’s (Poole, Dorset, U.K.) INGRID detects traffic incidents in urban areas. Relying on two classes of algorithms to minimize false alarms, one examines current data for sudden changes in traffic flow and occupancy.

The other uses ASTRID (the Automatic SCOOT TRaffic Information Database) to assess historical reference information against current SCOOT (Split Cycle and Offset Optimization Technique) data.

The patented Cetrac TMS4080 Wide Area Incident Detection system from Singapore Technologies Electronics Limited (Singapore) collects data at two adjacent points, spaced between 0.5 and 1.0 km (0.31 to 0.62 miles). An AI neural network performs non-linear mapping of raw data onto various traffic models.

At a traffic management center, operators receive audio or visual incident warnings with the relevant roadway section(s) highlighted in yellow. If the incident persists beyond a predefined limit, the display turns red to confirm legitimacy.

Traficon NV’s (Bissegem, Belgium and Chantilly, Va.) Video Image Processor (VIP/I) marries flow monitoring and incident detection in a single printed-circuit module. During set-up, the user can define conditions and types of events which, when detected, trigger an alarm.

Those include queue length, stopped vehicle, wrong-way driver, speed drop, smoke or fog and video failure. VIP/I monitors up to eight lanes and distinguishes five flow modes, based on speed and zone occupancy.

Approximately 50,000 vehicles per day traverse I-95 in urban Philadelphia. With daily use projections rising, the Pennsylvania Department of Transportation (Harrisburg) embarked on advanced traffic management along a thirteen-mile segment of the Interstate.

As part of that effort, Transdyn Controls, Inc. (Pleasanton, Calif.) furnished a complete Traffic and Incident Management System, consisting of three main components: non-intrusive, side-fired microwave units for sensing vehicle presence, speed, occupancy and volume; central computer hardware and incident detection software; and the CCTV sub-system for visual verification of incident response activities.

The presence and nature of expressway incidents as well as appropriate responses to them are highly distinctive. Given dynamically changing freeway conditions, the need for timely, adaptable and appropriate solutions is crucial to the performance, viability and public acceptance of any approach to incident detection.