Overview
Incidents are estimated to cause somewhere between 52% and 58% of total delay experienced by motorists in all urban area population groups, according to the 2010 Urban Mobility Report. The delay estimates differ depending on where and what data are collected, and how delay caused by incidents is separated from recurrent congestion delay. More importantly, this delay cannot be predicted, whereas recurrent delay, by definition, is predictable.
Among incidents that cause delay, there exists the so-called 10-90 rule: 10% of incidents account for 90% incidentinduced delay1 . The delay caused by the same type of an incident varies substantially depending on its exact location (e.g. relative to ramps, availability of shoulder) and, more importantly, on the state of traffic. The same incident, when traffic is very light or highly congested, contributes a smaller increase in delay than at other times. Clearance time significantly affects delay, especially when traffic is close to capacity: since queue lengths increase proportionally with clearance time, the freeway section will transition into congestion, and recovery will take longer. In many instances clearance time is prolonged for statutory reasons. For example, the Freeway Service Patrol is not allowed to move a vehicle to the shoulder if there is a serious accident: in case of injury, a Highway Patrol officer must be present before a vehicle can be moved2 . To reduce response (and hence clearance) time, it may be worth pre-positioning resources (Highway Patrol and tow trucks) to respond to incidents, but this is possible only if reliable statistical data of time and location of incidents are available.
Relteq Systems, Inc. develops software solutions to help traffic engineers and planners achieve superior operational performance on highways and urban arterials – as measured by traffic flows, absence of delays, lower fuel consumption and accident reduction. This Reliability IDEA project improved on technology created at the University of California, Berkeley to provide fast and reliable traffic simulation capabilities to traffic managers as an online service deployed in a cloud.
The resulting product, nicknamed Relteq Harmony, is a Decision Support System (DSS) that transportation network operators can use to determine the best possible response to traffic events. In particular, they can evaluate advanced operational strategies ranging from improved traffic signal timing or freeway ramp metering to variable speed limits, flexible lane assignments and alternative routing. Relteq Harmony functions either as a planning tool, or as a tactical tool that runs in real-time and lets a traffic control center to continuously optimize the application of available tactics, keeping roadways flowing at peak capacity. Relteq Harmony is developed following the paradigm of Active Traffic and Demand Management (ATDM) proposed by Federal Highway Administration (FHWA)3 .
The IDEA project enabled the development of key innovations that are now part of Relteq Harmony. Whereas existing traffic simulation software requires weeks or months of intense engineering work to establish parameters such as travel demand and network capacity, calibration of these parameters is automated in Relteq Harmony, feeding directly off traffic sensors4 . The solution runs entirely in the Amazon EC2 cloud, requiring no installation or maintenance on the part of transportation operators. This also means that extensive simulation resources are available on demand – multiple servers in the cloud can run in parallel to deliver faster results. Further, multiple government agencies can now collaborate on the same platform to provide even greater benefits to the traveling public.
Particular enhancements to Relteq Harmony resulting from the current project include:
1. Rich editing capabilities in Google Maps based network and scenario editor, allowing the user to place sensors, controllers and events on the road network;
2. Data handling mechanism for building operational scenarios from the available traffic and event data;
3. User Interface for launching simulations, single deterministic as well as batches of stochastic ones, and generate reports;
4. Software utility for automatic generation of a traffic model based on daily traffic data.