This week we are starting a project to examine the sounds that might be used for e-scooters to alert pedestrians of their presence. In this project we are working with a number of e-scooter manufacturers, led by TIER, Transport for London and a variety of groups representing people with various sensory challenges.
People seem either to love or loathe e-scooters! For some they are a new form of micro-mobility, convenient and easy to use; for others they are a torture, appearing as a surprise and creating stress from the shock of their sudden appearance. The former group uses them in increasing numbers and the latter is increasingly concerned, citing accidents and other worries, and this group includes a lot of people wit a range of capabilities - low vision, hearing problems, neurodiversity, reduced locomotion, and so on. This project is to see if, and if so, by how much, the concerns of the latter group might be assuaged by ensuring that the e-scooters announce their presence in a suitable way - in this case through the medium of sound.
But do we want to increase the loudness of noise in the urban environment? And what do we do about people who cannot hear the sound?
The answer to the first question is "no". This gives us a challenge - how might we make a sound that does not increase the already-too-loud noise levels? The project is just starting, so of course we do not yet have a precise answer to this question, but we do have some ideas about how to approach this challenge. The basic 'urban noise' is largely down to traffic noise, which is characterised as a sound with a wide frequency range, from around 20Hz to 3,000Hz - this makes it a rumbly noise with some high-ish frequency tones. Low frequency tones spread widely; high frequencies tend to die off within a relatively short distance. However, not all frequencies in the traffic noise are equally loud, and there is a loudness peak of around 60-70dB at around 1,000Hz, with the frequencies shorter and longer than that decreasing in loudness the farther away the frequencies are from that. This gives us an opportunity to exploit something called the "Cocktail Party Effect". This is a phenomenon that was recognised some time ago - as perhaps its name might imply - that people encounter when in a noisy room: despite the generally high noise level caused by all the people talking in there, the person can still hear someone on the other side of the room whispering the person's name.
The Cocktail Party Effect is an indoors version of a feature of human hearing that evolved thousands of years ago, and has remained in our hearing system. The human hearing system is very good at detecting a small sound far away and then making sense of it, either through recognising the sound through past lived experience, or by directing the vision system towards the location of the sound in order to obtain more information about its source ("is it a meal, or am I its meal?"). That sound, might be as small as the snap of a twig in the savanna. So we are looking for sounds that are quite small, but have the distinctiveness that works with the hearing system that makes them easy to detect. We can position these sounds in terms of frequency in one of the quieter parts of the frequency distribution of traffic noise so that it would be audible within, rather than above, the level of traffic noise. The nature of this sound is still being researched, but it is likely to be constructed from different parts of a sound - designing the onset, body and offset of the sound to work well for detection and location, as well as comprehension of the sound. We will be testing possible sounds against simulated urban noise of various forms, using the PEARL sound system to create and reproduce these in the laboratory. We will be testing them with people from a wide range of people with various capabilities, so that we can make sure, before testing any sounds on-street, that they are potentially acceptable to everyone.
But what about the second question, which is crucial for people who simply cannot hear well enough to detect the sounds? This requires investigation of other senses that might be able to aid detection of the e-scooter. Sound has the great advantage that we can hear all around, including behind, us, where our eyes cannot see anything unless they are directed there. But light can project onto surfaces, so maybe there is a possibility of using light to complement the sound to enable the e-scooter to be more visible, thus giving people more options for detecting the scooters, wherever they are coming from. All of his needs to be researched in a safe but realistic environment, and that is exactly what PEARL is for.
Watch - and listen to - this space!
A number of press items have picked up on this story. You can see some here (although it is a bit of a mystery as to where the idea that it would sound like a tuba came from - as a matter of observation, it is extremely unlikely that such a sound would work at all for this purpose: it would be far too low, and very difficult to locate!):