IOST: The Internet of Safety Things
The hype over the Internet of Things has spread to the security sector, and there are many opinions surrounding the impact it will have on future system design. However, should we consider a more bespoke approach, bringing together security and safety elements to create an Internet of Safety Things? Benchmark looks at this idea with the help of Bjørn Eilertsen, Vice President of Corporate Products Business Unit, Milestone Systems.
It’s been hard to escape the topic of the Internet of Things (IoT) over the past few years. Like many technology trends, it’s one that seems to be mentioned far more often than it’s ever explained. The definition is vague, and implementing ‘physical objects embedded with software and sensors with the ability to exchange data over the Internet’ is a pretty broad vision.
In fact, given the way it’s often discussed, you’d be forgiven for thinking the Internet of Things was imminent and easily achievable. You’d think a utopia of intelligent objects, all living in peace, harmony and common understanding, communicating and working together, was just around the corner. But is that actually practical in the real world?
Many manufacturers have been following the IoT hype-cycle. There is an interest in driving innovation, so companies like to keep on top of the cutting edge of technology. But so far, where the IoT is concerned, there has been more hot air than pragmatic evidence of how the Internet of Things will work. Some, including Milestone, think it’s time to puncture that hype balloon. We need to get real, or the flimsy structure of the Internet of Things will simply get taken by the wind and drift away.
Before going any further it is worth clearing one issue up: a ‘thing’ connected to the Internet is not the same as the ‘Internet of Things’. The very vagueness of the IoT term is a barrier here. There’s a semantic issue, a language barrier, which prevents genuine understanding and discussion.
The same is true on a technical level. A common language is required. The connected devices need to be able to communicate something meaningful to each other, or the situation is no different to what already exists. We already have devices connected to the internet discretely, with a multitude of confusing protocols preventing connections being made.
Birds of Paradise
It’s the notion of semantic networks which prompts Milestone to coin the term ‘the Internet of Safety Things’ (IoST). In a human sense, we’re not just talking about sharing a common language; we’re sharing a common interest.
Many think the most important aspect of IoT is a common language, but it is more complex than that. For example, if two people speak English, but one is interested in talking about football while the other wants to discuss the elaborate mating dance of the Papua New Guinean Bird of Paradise, there’s little to no common ground. However, if these people were both interested in the mating dances of Birds of Paradise, it would be a different story, especially if they both had access to different sources of information about this bird and its dancing ways.
To translate that back into security technology terms, we need our Safety Things to speak a common language, but also to focus on the same topics. If a sensor on a door could communicate to a surveillance camera to say that it is opened, the camera could then automatically move to look at the door. However, that interaction would be even more powerful and would effectively generate more meaning if the interaction was combined with additional sources of information.
If we combine the information from the door sensor with the information from an access control system, the camera is able to make more sense of the situation. The message it receives might then be, ‘John Smith has just opened the door’. The software is then able to compare John Smith’s database image with that of the so-called John Smith at the door. If it’s a match, he gets through. If it’s not, the system is able to send a message to all doors telling them not to open for this impostor while alerting security staff too.
So, what are Safety Things? The way we see it, they’re primarily sensors in the broad notion of the word. Under that category could be cameras, access control points, pressure and vibration sensors, thermal and temperature sensors, audio, motion detectors and any other item which can provide safety and security information.
Individually, each element, each thing, is relatively simple. It’s good at doing its job, and that’s it. That’s the whole principle behind the Internet of Things: to provide specialist objects with additional power through the connections with other specialist objects. However, there is a requirement for a logical organising entity to decide what to do with the messages received from the connected things. If you think that sounds a bit like a VMS, we think so too.
In the IoST, an open platform VMS would operate in a similar way to an Operating System to connect different devices. This would be a significant advance on the VMS of today.
Consider a scenario involving first responders to a car accident. This would be an evolution of today’s emergency response systems; the next technological step forward.
There would be cameras monitoring the highway. These cameras are constantly updated with data on traffic congestion. They are able to increase their recorded and streaming frame rates when traffic congestion increases, as this is when accidents are most likely to occur and information about them is most urgently required.
If an accident does occur, that video would automatically be routed to the traffic operations centre via the management system. A recommendation can be made as to the number of ambulances and teams to send to the accident. Live video can be sent to the ambulance teams before they arrive at the scene via tablets or in-ambulance devices, providing them with urgent data.
All team members would be using body-worn video, and these images would be streamed directly from the accident scene to the hospital. Using IoST-enabled instruments and monitors, first responders can quickly and efficiently record vital data direct from accident victims, such as heart rate, blood pressure, etc., and stream this direct to the hospital. They could also gather information from victims’ medical bracelets or fobs.
Forearmed with this knowledge, those at the hospital are able to communicate and issue potentially crucial instructions to the first responders, meaning that in effect treatments can start up to 20 minutes before the patient arrives at the hospital. These valuable minutes could save many lives.
An important aspect in this evolution of emergency response systems is the way that the video cameras and devices are able to join the network in an ad hoc way, as they are needed. Proximity to other devices, or simple requests from other devices for information, mean they are brought into the system when the data they can supply is of optimum use.
This is already starting to happen. Municipalities are organising their accident response teams in this way, and technology is being used to allocate resources in real-time, filtered via intelligent analytics.
What’s worth noting, however, is that this is effectively a closed loop. In this situation, we’re not talking about these devices being connected to the wider internet. Like the practical, pragmatic application of any part of the Internet of Things, it has to organised in an effective way.
At present, the Internet and the majority of IT networks are metadata structures that are characterised by immature management. In effect, that means sophisticated protocols which would allow for automatic connection to a network are not in place. In the simplest terms, connections still need to be made in the same way that we connect devices to the Internet.
This is not always as simple as many people might think. In the case of the first responder situation, this would need to be via an encrypted wireless channel which requires that each device is approved and authorised.
To remain utterly pragmatic, it should be said that for any truly effective IoST application, real automatic connection may never happen. However, things change and they change rapidly. Those protocols may be in place quicker than we might imagine.
Is the Internet of Safety Things a lot closer to being a reality than the Internet of Things is in other industries? The answer is yes, because the fastest, most pragmatic advances tend to happen when they assist with business processes, and safety and security is a key business process.
Many people talk about the ‘Internet of Things’ as a single entity, but of course we are more likely to see several IoTs. It makes sense that the Internet of Safety Things wouldn’t be the only IoT. The logical approach is for a number of IoTs, each effectively managing their own semantic network, communicating about the subjects that are most important to them.
However, with regard to flexibility, the ‘Things’ – the various connected devices – could have a role in more than one IoT, governed by different rules. A camera may have a security or safety role in the IoST while at the same time having another role such as heat mapping or asset tracking in the Internet of Business Things.
In the Internet of Safety Things, detection of a person in a shop outside of working hours, without the access control system reporting the presence of an employee, would activate an alarm. In the Internet of Business Things this would have no meaning, other than to perhaps notify the business manager that there may be a disruption to opening hours.
During normal business hours, the camera might just count people or register dwell events or even monitor stock levels and feed this information to the Internet of Business things or the Internet of Marketing Things.
Semantic systems can be dynamic self-learning networks, and the Internet of Safety Things is no exception. The rule engine in a VMS is an example of a static system where actions are triggered by Boolean logic. Effectively, this is an early version of the Internet of Safety Things.
By using artificial intelligence, systems would be constantly learning to improve their performance. In the earlier example of a dialogue between the door and camera, if John Smith had been prescribed new glasses and had undergone a drastic haircut, the system would quickly learn the new look character was in fact the right person.
This could be further enhanced by the rapid increase in the processing power of edge devices. At present much data transferred between devices is binary – on/off, yes/no for a given value – but increasingly this will become smart metadata, and then rich or intelligent metadata. Decisions will be made faster as devices are able to provide only the most relevant and crucial information for a given scenario, filtering out the other data as distracting noise, but storing it for follow-up purposes.
At Milestone, we see video as the most powerful tool in the security and safety arsenal. It is the most powerful sensor we have. It is the true key to unlocking the potential of not just the Internet of Safety Things, but all the other Internets of Things as well.
Video is a cultural phenomenon. As a means of communication, it is taking over from text or still photos for consumers, businesses and governments. It’s obvious why.
Video closes the distance between an event or a situation and the person receiving it. There is no text or other symbols to translate. It is not based on audio, so there are no language issues. The interpretation is done solely by taking in the visible images, as we do in everyday life. Because video is so very rich in information, it allows us to comprehend concepts and ideas in seconds that may have taken many hours of reading or studying in the past.
Combining the information from all the available sensors with video will provide more meaning. Information from different types of sensors can be combined to create effective semantic content.
As an example, on a dark and stormy night at what is probably a creepy old building, an audio sensor detects the sound of breaking glass. A thermometer detects a lowering of the temperature, but there’s no movement detected by the camera. It can only be one thing: a ghost! In reality, it’s probably just a broken window which means a call for the caretaker. However, if movement had been detected and visually verified, the meaning of all the information gathered by the sensors would have been different. It would have added up to an alarm sent to the security team for further investigation.
All aspects of physical security and safety can be more powerful if combined with video. They can provide more information and more data which can be used to make vital decisions. The management of video will be the true connector which enables the emergence of the Internet of Safety Things.
For many in the security sector, the Internet of Things, as it stands at present, is fluffy and ill-defined. However, the IoST can cross that chasm of understanding and practicality. IoST can provide a pragmatic bridge, using the power of video as its central enabler.
The first step is in recognising how the IoT can help us to be a smarter and safer society and the IoST is the first step on that journey.