The Internet of Things (IoT) is growing at an exponential rate, with thirty per-cent of the UK population now owning five or more connected devices, but with this proliferation comes risk.
The team at Cyber Defence Service, based in Greater Manchester, has developed a Machine Learning monitoring platform called Overt AI which uses detection nodes to scan an environment for IoT devices in real-time.
The nodes search for Radio Frequency (RF) communications, such as cellular data, WiFi signals and Bluetooth signals, and classify them by IoT characteristics such as device type.
Pete Davies founded Cyber Defence Services four years ago with Ian Strachan after the demand for IoT security became clear from a previous venture the pair were involved in.
"The job of identifying signals has historically been done manually, he said.
"What we have done is develop a system using Machine Learning that can classify signals with a 97% accuracy rate and this will only improve as we develop the product."
BACKDOOR ACCESS
IoT devices, such as temperature or humidity sensors, can appear harmless and isolated in their environment, however they are often relied upon as indicators for industrial systems.
Pete said that IoT devices can often be used to facilitate backdoor access.
He said: "Many IoT devices are often made with parts from the cheapest producer and as a result may not have continued security updates.
"Take a simple temperature sensor which may be physically secure but the method that it sends its data to a central service may not be.
"We find that the data they send is often unencrypted and can be read in plain text where you can see the temperatures that are coming in from the sensor.
"These types of devices are often connected to more critical systems such as air conditioning or building management systems.
"If the data is changed in transit so that it falsely reports that the temperature is ten degrees lower, then the air conditioning unit will do less work cooling critical computer systems causing them to fail, demonstrating a cyber-physical attack."
"If the data is changed in transit so that it falsely reports that the temperature is ten degrees lower, then the air conditioning unit will do less work cooling critical computer systems causing them to fail, demonstrating a cyber-physical attack."
GROWTH OF IOT DEVICES
The need to protect against IoT attacks is growing with an estimated 20 billion IoT devices in circulation worldwide today.
It is this acceleration in the need for protection that means the Cyber Defence Service is in demand from public bodies, private companies, law enforcement and military organisations.
Pete said: "We have clients who want us to solve a specific problem for their organisation.
"Our technology could be used by emergency services to check how many people are in a building when a fire breaks out or by investigators who may need to track an individual by the wearable IoT devices that they carry."
PRACTICAL USES
Intellectual Property is becoming a key cybersecurity battleground with many manufacturers banning mobile phones from sensitive areas.
Cyber Defence Service's Overt AI platform allows real time monitoring of sensitive areas through detection nodes that can be deployed inside buildings and externally, and for the alarm to be raised if a suspect device is detected.
Pete said: "Many companies do not allow personal equipment in certain secure or sensitive areas."
"Our technology allows us to identify if someone has a device such as a mobile phone, tablet or smart watch.
"If someone has one of these devices then the risk is that the data inside the building may be exfiltrated or an external party may be able to eavesdrop into a conversation."
Cyber Defence Service is also working closely with the Defence Science and Technology Laboratory (DSTL) to find a solution to detecting explosive devices by analysing the RF signals they emit.
Pete said: "We are doing some advanced work on deep learning and signal analysis.
"We are designing a mechanism that will identify the RF components of explosives with a focus on Improvised Explosive Devices (IEDs).
"Our technology uses Artificial Intelligence which has the ability to identify if these devices are in operation nearby."
TESTBED
Cyber Defence Service has joined five other start-ups at Salford’s MediaCityUK in a 12-week programme, which is led by UP Ventures Group, to discover and test how cities can use technology to adapt to a post-Covid world in partnership with MediaCityUK and Connected Places Catapult.
UP Ventures enables large organisations and start-ups to create effective innovation and build value together, through innovation programmes, design sprints and workshops while also helping businesses get ready for investment.
Pete said: "We are looking forward to benefiting from the expertise on offer at the programme.
"This is a new market with a lot of potential and we are looking forward to getting advice on validating the project and bringing it to market.
"We are also looking forward to advice on getting ready for investment and hopefully the programme will help us present the potential of Overt AI to investors."
MEDIACITY DEPLOYMENT
Cyber Defence Service plan to deploy IoT detection nodes around MediaCityUK during the testbed project as it looks to build a picture of devices spread across the campus.
Pete said: "In MediaCityUK, we will have a testbed of nodes operating and collecting information about IoT devices and from this we should be able to generate some interesting results.
"Hopefully we will be able to identify that everything is operating as it should but what we should be able to do is identify any anomalies in any IoT device use and look into why these may be occurring.
"An example anomaly would be a burst of transmissions coming from a device in a building at 2am when historically it was seen as constant data transfer."
FUTURE
Pete says that the Overt AI will adapt and change as different types of devices come on the market and that it will be of use to a wide range of organisations and industry sectors.
He said: "As more IoT devices come on the market, they will all be using the same frequencies that we track now meaning that we will not have to build new nodes. All we will have to do is a remote software update and refine what we are looking for on the Overt AI platform.
"We are also exploring how we can tie in the information that we are collecting with other information sources including those collected by transport companies to analyse passenger trends.
"Once you have collected data over a long period of time, it is easier to identify trends and anomalies such as spikes in traffic on transport networks.""If the data is changed in transit so that it falsely reports that the temperature is ten degrees lower, then the air conditioning unit will do less work cooling critical computer systems causing them to fail, demonstrating a cyber-physical attack."
If you think your company could benefit from working with companies on an UP Ventures programme you can contact us here .