The Nokia Automated Analytics Solution (NAAS) for Access Control is capable of simultaneously scanning up to six people per second per camera, at an accuracy rate of 95% when scans are conducted under a light source of around 300 lumens.

The system employs a combination of thermal cameras, machine learning and analytics, centralised management and cloud-based wireless technologies to identify heightened temperatures in multiple individuals at once with a margin of error of around +/- 0.3°C.

According to Mr Amit Shah, Nokia’s vice-president of Analytics and Internet of Things (IoT), this makes the system useful for rail companies aiming to provide non-intrusive, effective precautions to prevent the spread of Covid-19.

“We can scan multiple lines at the same time simultaneously, and can have a display which tells the person to approach the camera and then direct them where to go,” Shah says. “Of course, typically you don't usually want six people to push at the same time, because you want them to be socially distanced. But this is very typical in case of a railway station, particularly when families travel together, so you can send them in at the same time.”

Being flexible

The NAAS is based on Nokia’s pre-existing SpaceTime scene analytics technology, which was developed to identify anomalous behaviour in live video. The system was designed to be easily adaptable to a broad array of circumstances and capacities, with an open system architecture and a broad suite of analytics and automation workflows.

“The system comes with several different deployment models,” Shah says. “You might use it in a building with one entrance, or you might have a building like in a railway station, where they have tens of gates. It can handle up to thousands of cameras, so it is extremely scalable.”

Data from the infrared cameras is collected on a central interface.

Shah says that the NAAS system could offer significant improvements to efficiency for rail operators and manufacturers which feel hindered by mandatory testing requirements. He highlights the positive effect which the system has had at Nokia’s Chennai factory where it was first installed, where 1200 employees were previously scanned manually.

“One thing that is very important for us is to show to business that it really reduces their cost,” Shah says. “The head of the Chennai factory says that it has saved him almost €80,000, there’s no longer a long queue of shift workers waiting to get in, and they don't have two or three people dedicated to taking temperatures using handheld thermometers. You protect the workers because you don’t have to get so close to take a temperature.”

Future-proofing

Shah also says that a key focus for the technology was to ensure that it was both adaptable to changes in coronavirus regulations and to meet the additional needs of clients, and to ensure easy repurposing once the pandemic ends.

“One thing we were keeping in mind was that we were trying to make it future-proof,” Shah says. “The system is made up of a multi-hosted, multi-tenant cloud server where everything happens. And you have the capability of adapting it with somebody else’s intellectual property.”

Shah highlights the development of another proof-of-concept based on the SpaceTime technology to monitor level crossings, which was developed for Japan’s Odakyu Electric Railway, as a key example of how the technology could be adapted post-Covid. The system was trialled on the Odakyu network between February and March.

“This is exactly the same system,” Shah says. “The system is also good for platforms. You can draw a virtual line, and if any object crosses that, you can raise an alert. One case that comes to mind is platform safety. Or, you can focus it on the tracks and monitor any unusual activity on the track.”

Regarding the potential risks to privacy which a system such as NAAS may pose, Shah says that the system is equipped with a set of privacy settings which are pre-configured to be enabled by default.

“Nokia has a very deep sense of privacy, and we comply with the EU’s GDPR privacy laws,” he says. “My concern was that we cannot have a solution which distinguishes between employees or the general population. So, we provide the technology to blur faces, for example. I think the biggest safeguard is to provide the ability to upgrade the privacy level and put it in the hands of the operator.”