Cutting Through the Noise: Sirens and Emergency Vehicle Detection

In a life-threatening situation, every second counts. The busy roads of a city like Boston can make it very difficult for an emergency vehicle to respond to incidents quickly. In 2017, the median response time for potentially life threatening (Priority 1) scenarios in Boston was 6.4 minutes. Ambulances, police vehicles, and fire trucks use three forms of alert: an audible warning, flashing lights and colored bodywork markings. Often the first alert system to be noticed is the audible warning, which is typically a sweeping siren pattern. Early localization of sirens by pedestrians and motorists provides them with time to move out of the way to allow emergency vehicles to maneuver faster through traffic. However, sirens no longer cut through the noise of everyday life as they once did. Today, cars are better soundproofed and often filled with loud music or other distractions. A driver may hear the siren, but have trouble determining where the emergency vehicle is coming from or how to tell the difference between emergency sirens. Some new developments in siren technology aim to solve these problems to improve response times and save more lives.

Difficulties of the Modern Noise Environment

Emergency vehicles must compete with the din of heavy traffic, car radios and, in the summer, blasting air conditioning systems with the windows rolled up. Modern cars are highly insulated against external noise, and some high-end vehicles even include an active noise cancellation feature. These luxury features are starting to become more affordable, so we can expect to see more cars equipped with similar technology on the road. There is a growing number of motorists distracted by cell phones. Pedestrians are often wearing headphones and are similarly engrossed in their phones. These factors are enough to mask alert sounds, making it hard for a driver to recognize the siren and clearly define the position and direction of the emergency vehicle. Limitations of the human auditory system make this task especially difficult.

The Human Auditory System

The human ear does not have a built in spatial-mapping, so the brain has to process a variety of aural information to determine the location of a sound. To determine the angle of a sound source in the horizontal plane, the auditory system uses binaural (two ear) cues. Because our ears are separated by the width of our head, a sound that originates from the left side will arrive at the left ear before it reaches the right ear. This temporal discrepancy is called the Interaural Time Difference (ITD). (Specialized neural circuits in the brainstem can distinguish ITDs as small as 10 microseconds!). That sinistral sound will also be louder in the left ear, representing the Interaural Level Difference (ILD).

At high frequencies, ITDs are not very effective, because certain time delays may result in the same phase difference at the ears. Conversely, at low frequencies ILDs are not effective because low frequency sounds have wavelengths larger than the diameter of the head, so little to no attenuation by the head occurs between ears. ITD and ILD are not enough on their own for localizing sounds; monaural (single ear) cues are used which occur as a result of sound reflecting off the ear’s pinnae and upper body, shaping the spectrum of the incoming sound wave according to the sound source direction. A number of studies have shown these spectral cues are important for the localization of sound sources above and below the listener, and help to discriminate sounds coming from the front and back.

Generally, human localization performance remains approximately constant for frequencies below 1 kHz, acuity degrades somewhat between 1 kHz and 3 kHz, above which it improves again. The reason for decreased performance in this frequency region is that ITDs start to become ambiguous above 1 kHz, whereas below 3 kHz the ILDs are not significant enough for a listener to locate a sound successfully.

Sirens, Old and New

One might think that hard-to-hear sirens could be fixed by turning up the volume, but louder does not always equal better. Too loud and a siren can damage the hearing of pedestrians, startle drivers into making poor driving decisions, and make it difficult for the emergency responder to hear other sirens, which can be very dangerous when they enter an intersection. Ideally, a siren needs to operate at a lower level, in a frequency that humans can accurately localize. Existing sirens typically use a sweeping pattern, with most energy between 500 Hz to 1500 Hz, however, this pattern is not necessarily the easiest source to localize. As such, some companies have tried to design new sirens that are easier to identify.

In Britain, Deborah Withington, a principal research fellow at Leeds University, came up with a white noise siren that covers a wider range of frequencies than the traditional siren. Sounds with broadband spectral content and strong onsets are known to be easier to localize accurately than narrowband or tonal sounds with gradual onsets, because the brain is able to compare the binaural cues from many different frequencies and receives better timing information from the sharp onsets. However, much of a citizen’s response to a siren is based on what their preconceived notions of what a siren sounds like. Hearing a foreign sound may leave them perplexed or startled, but they have not been conditioned to recognize it as an emergency and respond accordingly. Americans might have experienced something similar when traveling abroad and hearing the distinct two-toned call of a European siren. The white noise siren lacks the cultural recognition of an old-fashioned siren, which has been related with emergencies for decades. So Withington combined the two, resulting in a highly effective warning sound that combines the strengths of the old and new sirens.

Another new siren, The Rumbler, developed by Federal Signal Corporation, has been used on emergency vehicles in over a hundred police departments through North America including New York City and Washington, D.C. Similar to Withington’s siren, The Rumbler system combines a low frequency sound with the standard electronic siren. It is typically played in short bursts while an emergency vehicle is approaching and passing through an urban traffic intersection. The low frequency sound can travel further and has a greater ability to penetrate and induce structure-borne excitation in nearby vehicle cabins compared to other siren technology. This structure-borne noise incites a “feel factor” that can be sensed by the vehicle’ occupants, including those that are hard of hearing or wearing headphones. The result is a better ability to warn nearby vehicles and pedestrians, thus lessening the potential of emergency vehicle collisions. One downside of this system is that the low frequency sound can travel very far, bothering residences far from the emergency (reportedly as high as 15 stories in New York City).

In order for cars to get out of the way of emergency responders, they first have to hear them. The urban environment and the human auditory system both make this a complicated task. The new sirens described here show that we can take advantage of psychoacoustic principles to create noises that we can recognize and locate better in the midst of all the modern noise surrounding us. The city of Boston has a goal of a median response time to Priority 1 scenarios of 6 minutes, which they missed by only 24 seconds in 2017. Improvements in siren technology that can help save those seconds that will help save lives.