Cost Benefit: Utah traffic light scheme pays dividends
First publishedin ITS International
Traffic signal programme leads way in Utah © Aquamarine4 | Dreamstime.com
A traffic signal control scheme in Utah is being taken up by other US authorities. David Crawford finds out how the Beehive State is leading the way in DoT and driver savings
Growing numbers of US state departments of transportation (DoTs) and their road users are gaining real financial benefits from an advanced approach to traffic signal monitoring recently developed in Utah.
Central to the system is its use of automated traffic signal performance measures (ATSPM) technology, brought in to improve the speed of delivery and accessibility of content of relevant data originating from on-road traffic controllers. Examples include flow rates, signal coordination and split failures (when green time fails to meet volume demand). As of 30 November 2018, 1,966 intersections with controllers were integrated into Intelight MaxView, a statewide management system from which current data is continuously being received.
The controllers record their actions and time-stamp the enumerations at 1/10th resolution, 24/7, storing the results in temporary data files. At intervals - typically every 15 minutes – a traffic management centre server makes a file transfer protocol connection to retrieve the outcomes. Once collected, the system transfers these into a database, where they can be interpreted manually or with the aid of automated graphics, analysed visually or with an optimisation algorithm, and archived for later comparative analysis.
The system enables users to view easily the information they need, for example on traffic detector status, in order to control hardware or influence signal timing and coordination without their having to install special software, or manage network connections or firewalls. Utah DoT traffic signal operations engineer Mark Taylor told ITS International: “Staff can use ATSPMs to measure directly what they could previously only estimate and model. The new approach allows an agency such as ours effectively to manage and optimise its traffic signals without the cost of extensive fieldwork, and to visualise the data in a web-based and easy-to-understand format”.
The results are also being made available to the motoring public – Utah says it aims to be the most transparent DoT in the US. Detector status is a key output. Each day at 7am, Utah DoT and its partner agencies in the state receive an automated email detailing vehicle and pedestrian detectors that have malfunctioned since the last sending. An algorithm searches through five categories (including the absence of any ATSPM data), flags the intersections that have experienced problems and sends out an e-alert to enable diagnosis and remedial action - before motorists note what is happening and call in to complain.
Table 1: Total estimated savings made by UTAH DoT ASTPMs over 10 months
For Taylor, “the brilliance of the automated alerts is that the algorithm is detector vendor-neutral and will work for all types of equipment - for example, inductive loops, video, magnetometers or radar - as it uses the amount or lack of data during certain time periods to generate the report”.
Between 1 January and 8 November 2018, the automated system found 637 issues with detectors that needed to be repaired, mostly before notification by the travelling public. “It’s obvious that ATSPMs are being used multiple times each day to improve the Utah transportation network and demonstrate its worth.”
Table 1 shows the number of times a report was run for each of a range of key metrics over an earlier 10-month period. The detection-derived savings ($3.06m) were to motorists in their travel times, calculated as being worth $15 per hour. Taylor calculated that, if the equipment was not working properly and Utah DoT had no automated system for being notified, “it would probably take road users up to 30 days to make a complaint”. He estimates the savings, for example from not having to carry out traffic counts, at $160,000. Since 2012, the agency estimates that it has spent some 12,000 hours in developing the ATSPM open-source software.
Utah DoT traffic management centre
Sharing the ATSPM software – and its benefits - with other US agencies is integral to Utah DoT’s approach; it charges no usage fee, thus reducing followers’ costs. An early adopter was Virginia DoT, one of whose officials attended a 2015 Utah DoT workshop and then decided to trial the system on an 18-intersection corridor along a heavily trafficked section of US highway 29.
Deployments in other US states include that of Georgia DoT, which won it first place in the innovation category of the 2018 Best of ITS Awards given by ITS America’s Georgia state chapter. Take-up has continued to spread following the formal inclusion of ATSPMs in the US Federal Highway Administration (FHWA)’s Every Day Counts programme. Introduced in 2009, this offers a US state-oriented model that identifies, and enables the rapid deployment of proven but not yet widely-adopted innovations aimed at improving highway safety, reducing congestion and enhancing environmental sustainability.
A further incentive for agencies has emerged with the recent publication, by consultants Kittelson and Associates and Indiana’s Purdue University (see box), of a guidebook on Performance-Based Management of Traffic Signals. This gives agencies advice on the specification, adoption and application of ATSPMs for smaller as well as larger systems. Meanwhile, equipment vendors have begun to include the necessary metrics in their products.
A 2014 review by Purdue had already identified the longer-term consequences of a lack of adequate signal performance reporting, resulting in the actual quality of operations often being unknown, while its impacts are “often underestimated in the programming of budgets and staffing levels”. Even when adequate resources are available, it can still be challenging to allocate resources if there is limited knowledge about which locations are operating well and which are not, the review says.
“Signal controllers and detection systems typically feature some functions for reporting a variety of errors,” it goes on. “However, usually this information is not easy to acquire remotely and is not always logged”.
For the majority of the 300,000-plus signals operating across the US, retimings take place on a three- to five-year cycle, at a cost of around $4,500 per intersection, with complaints from road users as the primary source of information on interim performance.
As the US Federal Highway Administration’s Center for Accelerating Innovation stresses, the lack of ongoing information drives costs up by necessitating modelling software along with manually collected traffic data to carry out limited ad hoc before-and-after comparisons.
From the ground up
The story began in 2012 when Utah DoT, following its long-established practice of sending engineers to the annual Washington DC meeting of the US Transport Research Board to garner fresh ideas, learned about an ATSPM approach that had been developed by
Purdue University. Its personnel subsequently visited Utah DoT to explain their initiative, developed with partners including Econolite, Peek and Siemens, and encouraged the department to develop its own software - which it did from the ground up.
Purdue provided specialist guidance on individual metrics. Utah DoT was in the position of being able to build on its prior investments in communications networks, which made possible a rollout to 1027 intersections in 2013. Utah DoT traffic signal operations engineer Mark Taylor said: “Our technology is a good example of how such a system can be quickly deployed when there is an excellent data infrastructure in place.”
Between 1 August 2013 and 31 May 2014, while the concept of ATSPMs was still relatively new, Utah DoT tracked how often performance measures were being used to locate faulty vehicle detection, and when these measures were used to improve operations at an intersection - such as automated counts for model optimisation and making time-of-day adjustments to signal coordination.
See Table 1 above