Sensors have long been used in industrial settings to monitor and improve processes and systems. Associate Professor Flora Salim from RMIT University speaks about how researchers are using sensor technology to enhance the liveability of our cities.
Manufacturing plants and refineries are usually highly-structured environments. They typically feature machinery, automated systems, and computerised technologies and software operating within a delimited space administered by a particular company or administration and all geared towards particular goals – profitable and efficient production, for instance. These days, industrial processes are becoming increasingly predictive. A recent study by McKinsey, for example, has outlined how the combined uptake of big data, advanced analytics and the Internet of Things(IoT) is boosting industrial productivity by reducing unpredictability. Life in our large modern cities, by comparison, can appear chaotic, unstructured and
unpredictable. The activities carried out by individual citizens are by no means necessarily pursued with systematic goals or the activities of others in mind. Sometimes, in fact, the simultaneous pursuit of the same goal by many individuals can lead to mutual obstruction. An ordinary trip to a desirable beach location, for example, can lead to frustration when one finds that several hundred, if not thousands, of other people have exactly the same plan – roads are congested, the best picnic or bathing spots taken. All the while, another equally tempting location may lie empty and unused.
The rational and efficient use of the resources and infrastructure upon which peoples’ working and leisure time depends is something that frequently animates policy discussions around urban planning. The federal government, for example, has itself established a long-term “smart cities” plan that seeks to develop policies towards greater utilisation of real-time data and IoT technologies in urban settings. The use of digital data to measure human activity is something that has been around for a number of years now. Companies like Google and Facebook, for instance, have long captured user data to target consumers with products and services.
Capturing data on the everyday interactions of people within their urban environment and infrastructure, influencing those interactions in realtime to make cities more sustainable, efficient, accessible and liveable, is a goal of projects aimed at forging future smart cities. This means cities that employ sensor, data and internet technologies to provide better living and working environments for their populations.
In most people’s minds, sensors are associated with the monitoring, measuring and testing of industrial processes, helping maintain the smooth, safe and efficient functioning of systems ranging from manufacturing to gas, water and other utilities plants. But researchers are now also seeking to use sensor technology to not only measure human behaviours, but to also use the collected data to help individuals make better decisions about how they manage their day.
Preliminary steps in the direction of testing small-scale smart city technologies are being trialled in a collaboration between RMIT University and Mornington Peninsula Shire at the popular beachside town of Rye. It is hoped that the project, led by RMIT’s Associate Professor Flora Salim and Dr. Andy Song, will provide a glimpse at how utilising existing digital technologies can make cities more liveable in the future. Once the project is up and running, traffic sensors will feed into smart signs displaying real-time availability of parking, while also guiding traffic to the least congested route. Sensors will be placed on barbecues and in bins to let council workers know when they need attention, and air quality sensors set up at toilet blocks. The historic Rye Pier will have air and water quality sensors attached.
“Mornington Peninsula is a really hot tourist attraction – there are a lot of beaches with popular picnic and barbecue areas,” said Salim, a member of RMIT’s Centre for Information Discovery and Data Analytics. “During the holiday period, it is very congested in some of the tourist towns, so people find it hard to locate a place to park their cars. Even the local residents, who just want to go down the street to the shop, are having trouble finding parking. Also, at some places, there are very few barbeque pits. A group could go to a barbeque or picnic area and find that it is already occupied. So, sometimes it can be very hard for people to plan even a simple daytrip to Mornington Peninsula.”
Salim explained that there is currently no data that can provide any information on the usage of the locations and their amenities. Rolling out the sensors across the township will change that. Salim and her colleagues will develop models that can predict ahead of time the availability of these different services by five to 30 minutes. “For example, if someone is arriving at Mornington, they can know within 15 minutes prior to their arrival which parking area they should head to, based on the current availability and historical data, both of which will be collected via the
sensors,” Salim said. “We’ve had a lot of experience in building these kinds of predictive models. So, that is the first thing we will do. The goal is to ease up congestion and alleviate some of the consequences of that, such as having people endlessly driving around looking for somewhere to park or to have their picnic.”
One of the important features of the project is that the data will be publicly available. Within a year, the team will develop and release a daytrip planner app, enabling people to engage with the data in real-time. “If someone is planning their day, and they want to target three or four sites, then we can recommend the best route for them for the day. And this data will be updated throughout the day. You will be able to select a preferred region, day and time, and other preferences, such as whether you want to visit a quiet or less crowded area. Through the app, we will be able to provide a presentation of the best option on the basis of the available
data,” said Salim.
“In this project, we will observe when people start consuming this data, how it influences their behaviour and decision-making. This includes decisions about what time someone might visit a location or not. For example, a person might look at the data and determine that at a certain time it will be busy and decide to visit at a later time. We want to be able to determine whether it might actually enhance individual productivity.”
Beyond this, the data will inform public authorities and potentially change how operations are carried out. For instance, data from the bin sensors will provide information on when garbage trucks will need to empty bins at various sites.
Going forward, Salim said that she sees the application of sensor technology expanding to encompass more complex and expansive urban environments, providing data for the more efficient use of resources or infrastructure. Her research team has also been working with the City of Melbourne, utilising its live sensor data from the city, including parking sensors and pedestrian sensors. “Using data can help city planners to be more predictive. For example, they will be better able to anticipate the numbers of people coming and going and help them determine how many parking spaces will be required, or if more parking spaces are necessary. It can show how many people are overstaying in their parking spots in certain areas because they might require longer time periods to park, due to the scarcity of street parking in the CBD.”
Salim said that while much smart city work was being done in research settings, the difficulties in securing approval for more trial projects meant that developments towards genuinely data-driven smart cities were going to take time. “There are a lot of things happening in research, but in order for this to be trialled, you need to go through so many different layers of approvals,” she said. “For this reason, I am very excited about the Mornington Peninsula project, because it will not just exist at the research level; it is actually going to be a real trial of these technologies.”
Going forward, Salim said that research was being planned that would see researchers working with road and public transport authorities to be able to utilise data from the traffic to ease up traffic congestion in big cities like Sydney or Melbourne. “We hope to be able to use data to recognise deviations in movements of people across a city – for example, a delay in a public transport network or a major road. These impact on the daily trajectory of people’s lives, so the idea is to develop a system to warn ahead of time that there are problems, enabling individuals to respond in real-time,” she said.
For Salim, a smart city would be one where there is a high level of responsiveness and engagement between individuals and their decisions, technology and the authorities responsible for providing the infrastructure and services we rely on in our day-to-day lives. “In a smart city, the enabling technology, in a sense, has to disappear – it is there, but you are no longer aware of it. All you know is that it is a great city to live in, because you are engaged, you feel like you are part of the decision-making, you know things are running efficiently, and you feel involved,” Salim said.
“The Internet of Things is already embedded in everyone’s lives – the core of IoT is faster computation with sensors and devices that are miniaturised. And this doesn’t happen in an industrial environment, it’s taking place in everyday life – our smartphones are IoT devices, for instance – but in a more powerful way. If you can capture the dynamics and the pulse of the city, then that will really help us to understand how a city could be more sustainable.”