User Rating: 5 / 5

Star ActiveStar ActiveStar ActiveStar ActiveStar Active


How HALO works to reduce false alarms passed to the Police by over 90%.


Ai, Deep Learning if you have not heard these words yet or understand what they mean or how they work then where have you been?

As humans, we possess the ability to recognise what we see, analyse it and draw our conclusions; generally, we are right; we make the correct decision. But sometimes because of distraction, tiredness or the effects of outside influences (what's happening at home) we make mistakes.


This failing is never more obvious in security operations centres.




Here operators are expected to complete an eight or sometimes twelve-hour shift with total concentration. It never happens. It is only luck that stops more errors creeping in. The likelihood of a security event occurring while on shift is low.

But, just like on the roads, when mistakes happen they can be catastrophic.

With autonomous cars now appearing on our streets, we are familiar with the statistics; we know that they will, like it or not, reduce accidents. As time progresses and the technology matures then those that remain will be driven out. Technology always improves.


Star InactiveStar InactiveStar InactiveStar InactiveStar Inactive


How HALO (Halt and Look Once) enables commercial control rooms to provide compelling home security to their portfolio.


Ai, Deep Learning if you have not heard these words yet or understand what they mean or how they work then where have you been? Now it is possible to add simple CCTV cameras to a property to protect it using nothing than the cameras. Armed and disarmed by the user using Alexa, mobile app, web page or automatically by a self-adapting schedule in Merlin.


The cameras send images to the Merlin control room whenever they detect motion but before they are used to raise an alarm these images are analysed by the HALO robot. The robot outperforms a human doing the same task by a considerable margin. If an activity is detected then alert can be sent to the homeowner using Merlin's other robot Molly who will intelligently monitor the notification process. 


False alarm generation is negligible, matching or outperforming even purpose-built sensors.  Giving protection to a property before it is even possible for a bad person to break in.


The association with VSaaS is enabling Merlin to be used in VIP protection, Police safe house installations and for large-scale deployments to housing associations. 






User Rating: 3 / 5

Star ActiveStar ActiveStar ActiveStar InactiveStar Inactive

Running an Incident Control Room is never easy, it requires decisions, sometimes life and death choices. These, in the first instance, made by people thrown in at the deep end or made to follow standard operating procedures because that is all they have.


All too often control rooms are still constructed with the expectation that operators can reliably monitor hundreds of CCTV cameras without any other systems in place. But for argument's sake let's assume that your control room is modern, a proper EOC (Event Operations Centre) with sensors, a way that the public can report an incident and with some level of situation awareness.


The reported incident. People may declare it; it may arrive electronically, and it may even be amongst hundreds of false positives. 


Take, for example, an incident in the street. Your control room may be monitoring social media feeds, for instance, the Twitter Fire Hose, and you may receive reports from members of the public. None of these can be quickly and more importantly geo-positioned reliably. Is this one incident or two? What do I despatch? CCTV invariably will confuse the matter further as cameras are hardly ever exactly in the right place and often are damaged or badly operated at the moment of crisis.


In the case of recent events, the decision has been to flood the incident with responders, just in case. But what happens if another occurs at a different place? Do I release those I have despatched to the first incident? Recent history has shown that following events, sometimes worse than the first or second, very often get inadequate attention. Sometimes costing lives because responders are tied up waiting for instructions.


Here is where artificial intelligence can help. Software with the ability to learn and detect patterns and assist commanders in making the right decisions. Identifying the severity of an incident, locating it by referencing landmarks and knowing the likely numbers of people involved. 


This information assists in despatch, positioning and directing second and third waves of responders, so they are ready but not so close that they clog access. And can be redirected if other events occur.



A cognitive software has great potential if you define what it is you wish to discover. 


Recent Fire events Worldwide could have been controlled better using artificial intelligence by receiving all the information from detection systems and responders in real time. Residents could be warned in phases, electronically with automatic and mass notification and fire services alerted more quickly to a rapidly maturing event. Directions could have been given by the software automatically by it knowing the location of the fire relative to the structure and second and third alarms called knowing the spread of fire. Connection to cognitive structural software would be useful to warn of the chance of collapse.


All too often an alarm receiving centre bases its response time only on the acknowledgement of the alert. The first signal in is held pending another for verification without any thought to where in the building the triggers are happening. It is proven that Ai can detect and increase the priority of an event based on historical data. It can learn event times, where they are and what they have been marked as in the past and apply that information to a queue. Service level agreements then based on time to action pre-filtered events rather than just being based on when they arrive in the control room. The smarter software can also analyse CCTV images and show operators where there is motion or indeed if, even before the operator has processed the alarm data; again increasing the priority based on the location of the sensor, motion and threat to the business.


Data is the key, whether that is from an external source, for example, a human, by monitoring social media, by utilising historical public information and your records. Many countries have free access to news, weather and risk and you can easily add sensors virtually anywhere. 


There is no excuse for the poorly designed control room any longer. Where the expectation is on an overworked operator to look at banks of CCTV, receive calls and process that information manually or from a paper operating procedure. No excuse at all.

User Rating: 5 / 5

Star ActiveStar ActiveStar ActiveStar ActiveStar Active

As a Security Manager you know that Merlin can reliably process Fire Event warnings, what about Panic Alerts? 

Just as good? Of course....


But what about a broken door at one of the sites you protect? Or a toilet that is blocked? Or even a request for a new network or some new lights? 


Well, they are still events, right? Things still need doing in the correct order, in a timely fashion and in a controlled way.


So, FM events are just lower priority (possibly) than security ones, and now you can process them using Merlin.  And they may not need to be done in the control room. They can even be done automatically; from broken door to repair and even payment to the contractor.


LinC is a new module that plugs straight into Merlin. It brings together a simple, yet remarkably effective, set of additional tools to the things you are familiar. There are new tasks for the MRE, a new queue (or a few), along with new modules for Nimbus and MerlinPHP.

To find out more. Click the image below. LinC is now part of the Initsys advanced training module.



Fire Engine Responding

User Rating: 5 / 5

Star ActiveStar ActiveStar ActiveStar ActiveStar Active

For half a century centralised control rooms have reacted to signals from equipment installed in the field. Fire and intruder alarms have wasted millions of responder-hours with false detection. Pointless data sent, and here I use the term literally and metaphorically, we send data that just says, "there is an alarm at this or that building". Even with protocols developed in the 1980's that were meant to provide sequential feeds to control rooms we still carry on bastardising the data feed so that the control room simply gets a message that provides no further data about the 'points' of detection and their current state.


This lack of detection point sequencing makes the control room blind, unable to use the data that undoubtedly exists to enforce the responder's view of what they are about to encounter.


It's true. Disasters have gotten less and less, but boy, when they do occur they are bigger, more horrific and much more dangerous to human life. 


Intruders get away with more because the alarm panels lock out after sending just one 'point' of data (or sometimes two) and the responders, dumb to what they are responding do not have the time to investigate every possible point of entry. 


Fires get bigger quicker because we make it easier for the monster to catch hold. Responders get data that says, "there is a fire alarm activating here", the human brain says, "we have been to this place many times before with the same information, it's bound to be the same", laws are changed to allow fewer responders sent in the first response. With no thought to the consequence should this time, God forbid, it's a real one.


But why not improve this process? It's time for control rooms to step up, to be smarter, for data analysed in real time. 


Imagine a different world, one where the ideas of the 1980's are released. By now the control room should be examining every piece of data available about a building. Every detection point monitored in real time. 


Responders would know that detectors have triggered multiple times, and are continuing to detect. We have all played the game where we try and get the LED on a PIR to light as we walk across a room. How about that 'game' at the control room? How about that data from multiple detectors, in real time, imagine that. The software would 'see' the intruder and map their path. It's what SIA protocol was designed for all those years ago. Instead, it was held back by the large alarm companies and dumb software writers who often still believe that it's the responder's job to get it right. 


And what about that fire alarm? Years ago Fire Services across the World looked for systems that pre-alerted people, made decisions on fire spread way before the responders could do the same. Years ago, 1991 to be exact, software demonstrated that it was possible to call people to tell them when it was their turn to evacuate, giving firefighters clearer pathways to the seat of the fire and managing exits routes, so they do not become clogged with a mass of people all trying to escape. Floor by floor, risk by increasing threat people told when and where to leave their safe zones. The software applying modelling rules that every firefighter knows in his head. 


It's time for regulators to put less emphasis on whether or not a control room has a concrete ceiling, blast-proof windows and metal lining. That is not how a bad guy is going to knock out a control room in the 21st century, none ever have and none ever will.


It is time for regulators to look at the systems employed, firewalls, penetration tests and the speed at which data arrives at the control room. To define what the control room should be receiving and how they react to that information. It is time to check that the information a dispatcher sends to the responder gives them a better idea of what they are about to encounter and it is time for people to be told what they can do to save themselves.




Because it is for the greater good and it is long overdue. It's what the industry accepted it must do in the 1980's, forty years ago.


Here is the smart, not lost, just forgotten.