If any industry stands to benefit from safety-driven automation, it’s mining. With extreme weather conditions, unknown obstacles and varying terrain, the hazard count on mining sites is unpredictable for even the sharpest of human minds – but not for machines. In an increasingly digitalised world, the mining industry is embracing emerging technologies that reduce human error to create a safer and more efficient working environment. While automation is bound to reduce the human workforce, it is the safety of humans that is driving this change. Foretellix – a platform provider automating the testing and validation of advanced driver assistance systems (ADAS) and autonomous vehicles – works with original equipment manufacturers (OEMs) and mine operators to generate an almost infinite number of tests and uses big data analytics to ensure these new technologies are safe to deploy.

The company is working with dozens of OEMs and mining companies to implement this technology, building on Foretellix’s roots in testing ADAS technology in the automotive space. In that market, ADAS has been around for some time – according to a 2021 research report from Canalys, approximately 33% of new vehicles sold in the US, Europe, China and Japan had ADAS features. By 2030, the company predicts that 50% of all automobiles on the road will be ADAS-enabled.

“Mining is a mechanical industry, it’s all about moving stuff,” explains Dan Atzmon, head of marketing. “As is happening in automotives, the stuff that’s being moved is no longer just using mechanical technology – there is a lot of software coming into the industry.”

“ADAS augments the human driver, but you still need one,” adds Erik Gulbrandsen, vice president of specialised vehicles. Drivers are faced with rocks on the roads, unexpected holes and boulders, and wildly unpredictable weather patterns, to name but a few everyday risks. At the same time, someone testing a new piece of technology on one mining site could drive around for hours without encountering a single problem – testing on-site is therefore notoriously inefficient. Running this system successfully is not easy to do – according to an American Automobile Association report, ADAS fails every eight miles. The problem is that automated systems, be it a simple ADAS function or a full-blown autonomous mine system, must have every possible scenario accounted for and tested before they can be deemed safe.

Foretellix is using its unique language, which has since become the industry standard, to specify abstract scenarios (“dump truck driving and crossing the path of another truck” is an example Atzmon offers) and then automatically generate all the possible tests variants of this scenario in the specific mine map. Different locations, actors, speeds, angles, weather, lighting and visibility conditions are all created to test ADAS and AV technology on a specific vehicle or a full mine, and ensure that it is safe and efficient, ready for commercial deployment.

Lessons from the past

Atzmon grew up watching the turbulent coal mining situation in England in the 1980s, where workers led a year-long strike in objection to mine closures and poor working conditions. While the fruits of their labour may not have been ripe as the industry became increasingly unprofitable, the exposure of the industry’s shortcomings have had a lasting impact. “We talk to different companies and OEMs but the majority of them are very safety-focused, and we have to be honest, this is a dangerous industry,” Atzmon comments. Almost all companies are interested in using ADAS and autonomous technology, particularly because of how it can improve the safety and efficiency of their operations, Gulbrandsen says, adding: “Almost all companies we’re talking to are interested in adopting collision avoidance systems.”

There are sites where it is difficult to get people to work and roles that are simply too dangerous to accept. An obvious solution in modern times, then, is to implement technology like ADAS and autonomous vehicles where many of the risks are mitigated. In these instances, the truck is equipped with a computer, an array of sensors such as cameras and often makes use of radar or LiDAR (Light Detection and Ranging) that, together, help the computer identify the environment it is in and provide the relevant input required for supporting the human driver to make decisions.

While many of these mistakes can cause minor or major injuries – both to the drivers, those involved in a potential collision or wildlife in and around the site – in worst-case scenarios they can be fatal. In September 2019, a male driver was killed at a mine in Goldfields in Western Australia when the mechanism failed while he was in the process of closing the truck cover. The rotating arm of the hydraulic motor is reported to have struck the driver and killed him. Discussing the effectiveness of safety management systems in response to this incident, mine safety campaigner Helen Fitzroy told the Australian Broadcasting Corporation (ABC) that “the effectiveness can be quite piecemeal […] people are still dying, people are still getting hurt”.

The urgent need to make the industry safer and more automated comes at a price. A study published by McKinsey in 2021 revealed that validation costs – including those related to simulation tools; data collection and storage; simulations using cloud technology; and the testing of the full software system in controlled environments and in the real world – will make up a third or more of total investment costs in all instances in autonomous vehicle deployment. When talking about fully autonomous trucks on a complete journey, the bulk of investment on validation is worth almost two-thirds. While a parallel study for the mining industry is yet to be conducted, Gulbrandsen thinks validation investment is worth at least 50%.

The bigger picture

ADAS has, in some form or another, existed since the 1950s – with the invention of the mechanical anti-lock braking system (ABS) – and has seen tremendous growth in the past ten years due to software-based solutions such as adaptive cruise control and blind spot warnings. While its value is evident, many mining companies and industry stakeholders are looking to take things even further.

As we move towards advanced autonomy features, ADAS is just one part of a larger system that aims to completely replace the driver. “That’s where things become much more interesting,” offers Gulbrandsen, “but at the same time more complicated.” The development of fully autonomous systems will mean people can be removed from mines, and in turn dangerous situations, altogether. Instead, humans will only be involved in site surveillance and in charge of monitoring the systems themselves.

“Mining operators want to achieve a fully automated site, including several suppliers of different machine types, and they have a mine management system overarching this,” Gulbrandsen adds. “They want that to work in conjunction with extreme high efficiency and extreme high levels of safety.” While this sounds somewhat utopian – or alarming, depending on which way you look at it – developing and validating fully automated systems is hugely complex, and the rollout will be gradual for a while yet.

Foretellix looks at specifying and generating all the possible situations that autonomous machinery may be in and allows them to be tested using simulation to ensure that the technology is safe. “Our system is not installed on the truck, we work with OEMs and mine operators,” Atzmon adds. “Using a map of each individual mining site, our platform is automatically able to calculate and offer a comprehensive list of potential hazards and problems via hundreds and thousands of simulated tests until we can say that this system is safe and we have tested enough.”

For a picture of how thorough these tests can be, Atzmon is a good person to ask. He describes tests where a truck is driving and a boulder shows up on the road, an animal decides to cross the road, then another truck appears out of nowhere, all while the weather is changing and the soil beneath the vehicle becomes damp. “If we find that 50% of the tests that day failed,” which they track each time on a comprehensive chart, “the developer and test engineer will sit down and figure it out and run the tests again,” he explains. This is trial and error on a gigantic scale until all relevant tests have been deemed successful.

“I want to see what happens when a tree falls and hits a rock, which then damages a digger, which is hitting me,” Atzmon adds, before explaining that this volume of tests can only be done with automation. The other side of what Foretellix does is show companies in clear terms what they have and have not tested based on the data the tests generate – this is otherwise known as ‘measurable safety’. The team can aggregate assets and factories into their platform system, which they are then able to reuse from site to site. Gulbrandsen says: “For example, we did not have tumble weeds on our system, then we encountered one and added it as an asset, now we can use it in every type of site that has tumbleweed problems.”

An automated future

As for a time when there will be no humans on-site, Gulbrandsen is reluctant to put a date on it. Some mining companies, including BHP, have released statements declaring that they will run fully automated sites within one and a half years, which the VP deems “aggressive” but serves as a clear indication of where the industry is heading. Most plan to be there in less than five years, so an autonomous future is not as far-fetched as it may sound. Atzmon is quick to add that Foretellix’s “systems are in place already. We are ready to help our clients now.”

The idea of a safe, fully automated mine without human workers is hard to imagine, but the industry has been preparing for a more digital future for years. Gulbrandsen, for his part, attended the Euro Mine Expo in Sweden in June 2022, where the entire first day was dedicated to discussing the digitalisation of mining. “This is what everybody is doing now, everybody is looking at how technology will affect the future of mining.”

This article first appeared in World Mining Frontiers magazine.