The mining industry has accelerated the automation of even the largest and most complex underground and surface machinery, but even ardent futurists admit mines of the future will require human interventions for decades to come.

Predictive biometric screening technologies, which use digital tracking and monitoring to gather vital data on an individual's health risk factors, are set to play an increasingly important role in ensuring that these humans are kept safe.

Unlike biometric solutions used for security identification such as fingerprint, iris or face recognition, these screening tools check vital signs such as breathing, heart rate and fatigue levels in real-time, as well as identify potential long-term health risk factors.

This combination of human health monitoring with digital accuracy seems a near-perfect tool. Still, similarly, the technology raises difficult questions over user privacy - is it right that a company should access a worker's personal health data or location, even to ensure their safety?

Mining Magazine EMEA Editor Craig Guthrie talked to Alex Moss, Canaria Technologies chief executive, about her company's Canaria-V, a predictive wearable designed for applications such as mine worker monitoring that is capable of predicting heat exhaustion and cognitive fatigue simultaneously. 

Moss' company has taken important steps towards wider adoption of the predictive biometrics platform this year, in particular through a trial that is underway at Macarthur Minerals' Lake Giles iron project in Western Australia, which was announced in September.

Running such an innovative company at just 27, while also featuring in Vogue Australia and making a splash in the fashion world, also makes Moss an inspirational figure for young women breaking into the mining and engineering sectors.

How does your solution use AI in tandem with biometrics? What is the difference between a medical and a human safety-based incident?

Canaria Technologies uses high accuracy biometric data in tandem with AI to predict future physiological states in our users. We do this by measuring a combination of biometrics and some information about their environment - such as temperature, barometric pressure and humidity - to understand in detail their current state. We then use their historical data to predict when they are at a high risk of a cognitive fatigue (microsleep), heat exhaustion, or man down. It is basically the same method as the 'work out the next step in this number pattern' question we all had to answer at some point in school.

To do this effectively, we have spent the last few years building and testing five generations of medical-grade accuracy wearables to ethically gather data from our users in-field on mine sites to establish a group baseline which works for most people. Our AI then works on top of this baseline to tailor its predictions for users.

This results in a system that knows that you have a six-hour threshold for extreme heat environments and that your colleague has a six-and-a-half-hour threshold for the same extreme heat environment.

Medical and human safety-based incidents are often intertwined, and two-thirds of all heavy industrial accidents are caused by cognitive fatigue. This is especially expensive and dangerous in the mining industry when someone has a microsleep whilst operating heavy machinery; or making a logistical calculation error in a control room from exhaustion which causes a chain of effects leading to an accident - heat exhaustion is a major problem for sites operating near the equator.

For ease of categorisation, think of a medical incident as a severe change in someone's vital signs which indicates a high degree of suffering. Usually, a medical incident predicates a human safety-based incident (again, a microsleep at the wheel of a vehicle is one of the most common examples of this). 

 

Can you explain how your predictive biometrics systems improve operational efficiencies to enhance productivity? 

Two-thirds of all heavy industrial accidents have a root cause of cognitive fatigue; in Australia, there are 130 recorded severe heat exhaustion incidents per summer. Predictive biometrics systems really do two things.

Firstly, they measure with high accuracy the physical states of our users and some of their surrounding environmental information to build knowledge of major health-and-safety related problems.

Secondly, they then use this to build predictive systems to detect and send out alerts about medical incidents 10 minutes before they happen.

This means that there are two categories of ways they improve operational efficiencies: firstly, by building an understanding of known and unknown issues and then by predicting and preventing known issues before they happen.  When we started work on heat exhaustion, there was almost no data about the problem from both a human safety and financial view.

By building an understanding of known and unknown issues, we help our clients understand difficult-to-quantify issues; this can lead to immediate solutions such as new training protocols. This is quite a wide definition, because we already know that the next immediate problem we need to look into is asphyxiation. The beauty of our system is that we update our hardware with new versions of software to add alarms as we discover and solve more problems.

The second part of how predictive biometrics systems improve operational efficiencies is more straight-forward: by preventing the most serious causes of accidents (heat exhaustion, cognitive fatigue, man-down), workforces have less time spent on sick leave, there is less paperwork to document safety incidents, there is less work to find replacements for staff who have suffered incidents, and there are dramatically fewer fatalities with the accompanying dire psychological knock-on effect these have for entire sites.

We do this by sending predictive alarms to individual users via our wearables, which are the size of a hearing aid and designed to fit easily with pre-existing HSSE equipment, and to managers and control centres via our dashboards. One of our software development goals is to integrate into pre-existing software on-site to make this process as seamless as possible.

It's important to note that all of the above only works if users trust our system: we adhere to EU Best Practice Data Privacy Laws, take the time to train all users with our equipment, and only measure what we need to to prevent serious safety incidents -we can't measure alcohol levels, pre-existing heart problems, when someone takes a bathroom break etc. And our users can access their own data any time. In practice, this data is actually quite boring; it reads as a lot of spreadsheets with vaguely sense-making binary data- we're still a long way away from Minority Report style technology.

Can you tell us about how your work with the Lake Giles Iron Project is progressing? How did that collaboration come about?

It's progressing well. We're currently testing the new generation of equipment that we'll be sending out to them over the coming hot season in the southern hemisphere.

The collaboration came about from a series of in-depth conversations about the future of mining, ethical technology, and the philosophy of business with Macarthur Minerals' Andrew Bruton which took place over about two years. We just had a great time having conversations with each other at industry events, and when the right place -Lake Giles - and the right time - our fifth generation of our hardware being designed- came about, we were both in a position to pull the trigger.

We are delighted to have found a like-minded industry partner with Macarthur Minerals, and I cannot give enough credit to Andrew Bruton for bringing new ways of thinking into our industry. 

 

What were the formative steps that led you to develop this system, and adapt it to mining? How do you plan to approach privacy as a topic?

The system was originally designed for NASA as a remote high-accuracy vital signs monitor for astronauts for use aboard the International Space Station. My co-founder Dr Rob Finean and I won the NASA Global Best Use of Hardware Award in 2016 for our proof-of-concept and set up our company immediately.

Once we decided that we didn't want our immediate focus to be on R&D in the space sector - the Total Addressable Market for manned space exploration is six people, not exactly the best starting market if you have business ambitions - underground mining was the next most obvious use-case.

Both use-cases require adhering to strict hardware development protocols, have inferior and erratic internet connection, and user's lives depend on the functionality of the system.

We relocated to Australia in 2017 to complete creating and testing our predictive biometrics system; working out along the way that the first set of problems our system can solve are heat exhaustion, cognitive fatigue, and man-down incidents.

I won't go too deeply into privacy, because, again, it would be its own article. Both Rob and I were data privacy activists in the UK before we met each other and teamed up: so, it's a critical topic for us. Our stance is that the current data privacy laws in existence do not cover predictive biometrics systems. So everything we do lays down the foundation for future legislation about the ethical use of predictive biometrics systems companies as more and more emerge over the next decade.

We need our users to be safe, to feel safe, and to know that the people designing their technology are thinking about the broader ramifications of the use of their data beyond day-to-day-use, ie what if there is a data breach in five years? Are there going to be personal medical insurance premiums knock-ons from our system? How do we prevent this? Should there be legal barriers to specific uses of predictive biometrics systems? Taking a stand that identifying biometrics data should never be sold to third parties etc. 

You are successful in fashion too - what can the mining industry do to attract more diversity in its innovation sectors?

Mining has not only the available funds but also some of the best minds available to be a true leader in innovation. One of the things I've seen done extremely well is where mining companies essentially establish themselves as patrons of R&D, supporting and encouraging innovation by presenting an issue and seeking solutions through hackathons, incubators, or competitions.

However, this is often only in technology or engineering fields; why not across all aspects of mining? Innovative people, regardless of their background, like to solve problems, and when presented with the opportunity, they jump at it. I think reaching out beyond the obvious to provide funding or mentorship or both, will see a flood of people from all kinds of backgrounds arrive in the mining industry.

The people are there - I just don't believe it occurs to them that mining is the place for them. That's something that is easily fixed. Like in my case, once I started to work with the industry; I've realised I love it, others will too.