Programs & Labs
2003 Annual Report
A report of research, education, and technology transfer activities, fiscal year 2002/2003, of the Intelligent Transportation Systems Institute at the University of Minnesota.
Research
Development of Dynamic Route Clearance Strategies for Emergency Vehicle Operations
Sensors mounted on or near traffic signals can alter a signal's timing cycle in response to an approaching emergency vehicle
From left: University police officer Josh Betts, Mn/DOT signal engineer Robert Betts, principal investigatior Eil Kwon, and research assistant Sangho Kim
For a heart attack victim, a few minutes of delay in receiving medical attention can mean the difference between life and death. The nationwide 911 system saves thousands of lives by reducing the response time of emergency vehicles. But after being dispatched, even highly skilled emergency response personnel can be slowed down by red lights or, worse, find themselves stuck in traffic as motorists struggle to clear the roadway.
At the ITS Institute, a research group led by Dr. Eil Kwon is exploring dynamic route clearance for emergency vehicles, a technology that promises to reduce emergency response times while minimizing traffic disruption for other motorists.
Eliminating red lights along the route of an emergency vehicle can give emergency response teams a critical speed boost. Not only is the emergency vehicle able to proceed without slowing or stopping at intersections, but traffic on crossing streets is prevented from entering the route, so traffic volume is effectively reduced around the emergency vehicle.
Several commercial systems are available to accomplish this goal. Known as Traffic Signal Preemption systems, they consist of some form of vehicle-mounted signal emitter (such as an infrared beam or low-powered radio transmitter) combined with sensors mounted on or near traffic signals. When triggered by the emergency vehicle's transmitter, these sensors activate a control mechanism that alters the traffic signals' timing cycle. The result is that traffic signals change to green more quickly when a transmitter-equipped vehicle is approaching, and stay green until the vehicle has cleared the intersection.
Intersection-based signal preemption with local detection, however, has several limitations. Chief among these is the fact that a vehicle-mounted signal transmitter must have a clear "line of sight" to the traffic signal in order to trigger the preemption routine. Further, it is possible to have unnecessary preemption for the intersection signals that are not on the emergency route if they happen to be located in the line of sight.
This problem becomes particularly acute in urban areas, where emergency vehicles must make frequent turns and where closely spaced buildings block the preemption signal at corners. And in congested city traffic, even when the preemption signal is received relatively early in the emergency vehicle's approach to a traffic signal, it may be difficult for motorists to clear the intersection in time.
Dynamic route clearance goes beyond intersection-based signal preemption by managing the entire route that the vehicle takes from dispatch to emergency scene. In a dynamic system, a network-monitoring module continually gathers traffic information and passes these data to a route selection subsystem. The subsystem then calculates an optimal route based on current conditions. As the emergency vehicle travels along this route, its location is monitored by the central system, which intervenes in signal timing as appropriate.
Mathematically determining the best (i.e., lowest travel time) route is an example of the "single-source shortest path" problem; to solve it, the route selection module employs the well-known Dijkstra's algorithm. In this case, the network monitoring module first computes travel time between points considering current traffic conditions, and the algorithm computes the "shortest" route in terms of time rather than space.
As soon as the emergency vehicle clears an intersection, the system initiates a signal-timing recovery procedure to return the signal to its original pattern. Kwon's system, designed around the signal policies of Minneapolis, adjusts the "Walk" interval of the crossing street (whose signal was blocked during preemption) to return the system to its original state.
To evaluate the route-based dynamic preemption system, Kwon and his team used VisSim™ microscopic traffic simulation software interfaced to an external virtual control center module, which was developed by the researchers using the C programming language. Key capabilities of VisSim for this application included its ability to model a set of detectors that can detect only emergency vehicles. A virtual intersection controller module was also developed to emulate different types of signal preemption strategies including the existing method with local detection. Because the current version of VisSim does not permit the route of an emergency vehicle to be specified during the simulation, testing the online route selection module was not performed in this study; this phase of testing focused purely on the effectiveness of dynamic signal preemption with a pre-specified route.
The streets around the University of Minnesota's Minneapolis campus were selected as the sample network for this evaluation. Geometric data on the test area were collected from aerial photographs, and detailed traffic data including volume, signal timing, and current preemption sequences were provided by the City of Minneapolis Traffic Operations Center.
For calibrating the simulation model, actual travel time data for emergency vehicles on three routes were collected in cooperation with the University Police Department. Two of the three routes chosen are equipped with the Opticom™ signal preemption system, which uses a vehicle-mounted light beam to activate the preemption routine.
Simulation of multiple distinct routes revealed that route-based dynamic signal preemption produced superior results on relatively long and complicated routes when compared to the existing intersection-based preemption method. For example, emergency vehicle travel times under the dynamic systems showed reductions ranging from 9 to 12 percent on a relatively straight route, and from 10 to 16 percent on a complicated route.
Network-wide traffic performance was also evaluated from the point at which the emergency vehicle was introduced into the simulation until the end of the simulation period 30 minutes later. Total vehicle-hours of travel and delay per vehicle data show comparable or better results under the dynamic preemption method when compared to intersection-by-intersection signal preemption, even while the emergency vehicle itself realizes a significant travel-time reduction.