Use case: noise and sound
What’s the issue?
A significant amount of teaching is still delivered aurally. Face to face learning activities, for example collaborative group work, also typically use forms of communication between students that rely on a conducive acoustic environment. In addition, many students live in close proximity to each other, and engage in social activities that can create noise. So sounds are both an important element of learning and the university experience for most students and also a source of annoyance.
High levels of noise can be damaging to health, and noise pollution or noise nuisance reflect unreasonable amounts of noise within an environment. These can sometimes result in the local authority taking action against those creating the noise. However, in most cases noise is more of a distraction or annoyance, affecting your ability to carry out the activity you intended. The impacts can include disruptions to sleep, increased stress or anxiety, or simply poor concentration or productivity. Historically, responses to noise problems are infrequent and complaint driven, so have limited impact on the situation at the time.
What can be done?
The level of “annoyance” experienced as a result of the noise is subjective, and can be moderated by a number of factors including sensitivity to noise, the ability to control the noise, the predictability of the noise event, and other environmental “comfort” factors.
The Intelligent campus offers opportunities to provide control and information that can reduce the amount of annoyance. Equally, responding in real time to triggers of certain noise levels could result in additional mechanisms to reduce the noise at its source or by insulating the sound from those who are impacted.
The idea of quiet study spaces already exists, and even some halls of residence are designated as quieter spaces, although they are typically static, not responding to environmental changes, and relying on self-regulation or human monitoring and intervention. What would be more effective would be to tailor spaces to the individual or group’s activities and preferences, and respond immediately to noises that impact negatively.
Environmental sensors can continuously monitor the acoustic levels of both indoor and outdoor spaces, and report this information through a variety of interfaces including mobile devices. For example apps could show you the current noise levels in busy bars, study rooms or computer suites, enabling the individual to choose a location appropriate to their activity.
Newer sensors also sometimes have processing capability, including being able to identify and categorise different types of noise, for example recognising the difference between traffic noise, live music from a band in the student union, or the sound of shouting and screaming. This allows for a differentiated approach to various acoustic triggers.
Making it intelligent
Integrating different types of data can provide a rich source of information about noise events and their context, allowing more informed responses. Knowledge of event timings (eg music concerts), and the movement and congregation of people across campus (through location information), can start to build a picture of where noise might build up and what neighbouring groups or activities might be impacted (people sleeping, or doing quiet study).
Nudging strategies, for example recommending different routes or transport options could help to disperse noisy crowds, suggest less busy bars, or direct crowds away from exam rooms. Less direct than alternative routing could be signposting special offers or “things you like” nearby to entice you elsewhere – food in the cafe, a lunchtime club attended by people you know, a book you wanted just arrived in the library… Knowledge of the intended timing (and the actual happening) of noisy events or quiet study activities could also influence the schedule of maintenance works.
Noise cancelling technology exists in more sophisticated headphones, and could be applied to larger environments such as study rooms where external noise is detected and unwanted. The data identifying noise events could be combined with data on scheduled quiet events to effectively target noise cancelling, or even respond to mobile controls by individual users. Weather could also be incorporated into the decision strategies, for example if windows are more likely to be open on a hot day, allowing noise to carry further.
Are there any examples?
A research partnership between the City of Calgary and The University of Calgary in Canada used acoustic sensors to autonomously detect and categorise acoustic events. The technology was able to pinpoint the location and time of the noise, and use machine learning to differentiate between different noise types including music and construction. The authorities were notified when legal levels were exceeded, allowing action to be taken, and noise patterns over time were mapped and correlated to better understand acoustic impacts of different activities.
In Louisville Kentucky, similar technology allows a drone response to gunshots. Thankfully such incidents are rare in the UK, however other applications could be imagined, for example providing monitoring and aid to individuals in distress or subject to attacks.
In Singapore, at Nanyang Technological University, researchers have applied noise control technology (such as used in headphones) to cancel external noise. They formed a grid-like array of units on an open window, which reduced up to 50% of noise from traffic and construction. It detects the noise and models it in real time, producing an inverted waveform to cancel out the noise as it happens.
In some of the real examples above, the perception of the introduction of “surveillance” was opposed by some, but the advantages of accurately targeting serious incidents were argued as being better for privacy than blanket CCTV coverage. As with all examples involving sharing of data, the collection and interpretation of information about individuals, such as their location or behaviours, is sensitive and needs to follow a strict ethical code.
Who needs to be involved?
With the increasing integration of data to monitor, model and respond to acoustic events and the wider environment, a broad range of stakeholders and data could be involved. This includes academic services such as timetabling; estates services including catering outlets and buildings maintenance; student social activities such as clubs and events; campus events including concerts; local authorities for roadworks, construction activity, traffic management and transport operators.