Autonomous mining: bridging the gap between man and machine

Sometimes, when it comes to applying and advancing new technologies, it can be valuable to look outside of the mining industry to sectors, and experts, with different understanding and experiences.

Which is how I found myself sat in a virtual coffee shop with Missy Cummings in April.

Currently a Professor at Duke University Pratt School of Engineering and the Duke Institute of Brain Sciences, Cummings is also the director of the Humans and Autonomy Laboratory.

She has advised many companies on autonomous systems across multiple sectors, including miners like Rio Tinto and, if that resumé isn’t impressive enough for you, she began her career as one of the first female fighter pilots in the US military.

While Cummings scrolled through various Zoom backdrop options including hell and a zombie apocalypse (perhaps a little too close to home right now?!) I asked where she thought the mining industry is currently at with its use of autonomy and AI.

“I think the news is mixed,” she replied, settling on the coffee shop setting. “There’s a lot of hype in this area, and companies like to talk a big game because it’s a very competitive environment.

“Not just in mining but across the board, there’s more hype than there should be. I do think that Rio Tinto – I worked with them a few years ago – is probably still the lead front runner in technology, but that gap is closing.

“Their vision was amazing. They accomplished so much in a short period of time in an industry that is otherwise extremely slow to change. And then BHP saw what was happening and now they’re playing a quick catch-up game.

“I think Rio really set the bar, but they seem to have stagnated a little since then. They’ve been through so many personnel changes and all the typical highs and lows that come along with mining, and they’ve lost some key people who I think were pivotal to their vision of the future.

“It’s a bit like warfare. They had a good run and now the enemy is coming up quickly behind them, and they’re losing their edge. It’s easy to get complacent if you’re the first to market. If you don’t have the vision for how to keep pushing the edge, then eventually you’ll be overtaken.

“This a business lesson that happens many times over in many different industries.”

Getting over our materialistic tendencies

Cummings believes the biggest problem that the mining industry faces when it comes to automation is our obsession with physical objects, what she calls our love of ‘The Thing’.

“The mining industry, and I’m going to be totally sexist here, are a bunch of men who just love their toys,” she said.

“Who doesn’t want a robot truck or a robot train? People want to believe that if they buy just one more automated shovel, for example, that it’s going to solve all their problems.

“Rio’s done a great job at getting ahead, but they’re not going to be able to make the next leap ahead, and neither will any other company, until they get over their love of The Thing, and start understanding that what they really need to do is use AI and maybe some other tools to improve their processes.

“Improving processes is boring. It’s not sexy. But right now, mines have The Things, and now they need to figure out how to make them work better. Automated trucks and trains have a lot of problems. Their human interaction, and their interaction with other parts of the supply chain too, both pit to port, or pit to plant, those are all governed by processes.

“You’re not going to get from pit to port or plant unless you improve the processes dramatically. I think if we got all the CEOs together and got them to admit it they’d be like, ‘yes it’s great to have all those autonomous trucks, but we’ve still got a backlog of material going on at the crusher’.

“You can have all the things you want, but if you don’t fix the whole chain and optimise it for human interaction, which really, at this point, is human supervision, then you’re going to make some advances but you’re not going to get where you really want to be.”

Humans. Can’t mine with them, can’t mine without them

Cummings has a point. Many of the iron-ore mines in the Pilbara that are pushing the boundaries with autonomy are striving to create fully autonomous sites. Or what they term fully autonomous sites.

There is a significant amount of automation already in place but there are still gaps in the supply chain, siloed processes, which are getting harder and harder to connect.

“When you go to these iron-ore plants, for example, there’s still so much inefficiency around movements in the yard. And a lot of inefficiencies that AI could be helpful in solving,” she explained.

“They’ve got a ton of data; they could figure out how to use AI to get insights into weaknesses in their supply chain. But, more importantly they need to use AI to improve their scheduling processes for humans so that if people call in sick it doesn’t disrupt the entire schedule.”

It seems odd, counterintuitive almost, that a site that is programmed to run completely autonomously, needs to be designed first and foremost with people in mind.

“Even in the perfect mine of the future, at totally automated mines, rapid response human teams will need to come in and fix contingency operations,” said Cummings. “Things break – autonomy does not work perfectly in all settings – you can’t ever get away from that.

“So instead of spending more money on ‘things’, we need to start figuring out optimisation from the human perspective: what should humans be doing, what should the autonomy be doing, and where should we start training to have a joint team?

“Because the company who will win the next wave is the one that understands that joint space between the man and the machine. Understanding how to do workforce planning in terms of the numbers and types of people you’re going to need to have, it’s going to be a huge challenge.”

Overhauling education

The problem is nobody wants to optimise team architectures; they want to optimise pit processes.

“I think one of the issues here is education,” Cummings told me. “We primarily teach people how to optimise physical processes. But there are some great universities in Australia that also look at how to truly optimise team processes.”

The future of mining will require a lot people, and robots are the perfect way to plug a gap created by a skills shortage and increasingly inhospitable environments. However, those people can’t be just any people, they need to be highly cognisable with technology.

Mining engineers of the future will still need classical training but, going forward, they will also need degrees in computer science.

“It’s a much bigger problem across all industries that are touched by automated or autonomous technologies,” said Cummings sadly.

“There’s an understanding that education now, computer science, machine learning, AI… these are core subjects that should be taught to every engineer, not just computer scientists or computer engineers. Mechanical engineers also need to have a software background now.

“It’s very difficult to hire people that have the skill sets required to work with that kind of advanced technology.”

The rise of networking

She added that another challenge the industry faces is the understanding that mines no longer exist in a vacuum on their own.

“One problem that I’ve seen across the world, in all mining companies in all types of materials, is the fiefdom that is created by the mine manager. There’s some kind of weird feudal lord system still happening in mines across the globe. I know you know what I’m talking about!”

Cummings had caught me smirking in a knowing way. I hadn’t thought of it that way before, but it’s an apt analogy.

“It’s true,” she continued. “I’m not knocking it per se. In a time when mines weren’t connected and they had to do everything by themselves, they needed a feudal lord system. But now, with all this technology that requires remote and distributed control, we need a different type of mine manager.

“We need mine managers who are a lot more technically savvy, and who understand the value of working inside of a network system. It’s a rare skill set to have somebody who understands general mine practices and is also well versed on automated and autonomous technology.

“Good luck finding those people.”

A work around for this would be the use of more network control or other systems that would allow operations to share best practice. For example, teams of specialists in the mining community who travel from mine to mine to work on various processes.

“Humans love problem solving, but the idea of having rapid reaction teams able to go from one mine to another… We’re very far from that right now,” Cummings said. “It would require significant advancements in thinking as well as the networking technology.”

Doing business differently

It is clear that, in order to meet the challenges of tomorrow, physical, technological or otherwise, and not just tread water but actively thrive in this environment, companies are going to have to rethink the way they do business and the way they design, build and operate mines.

And, if the big players can’t figure out how to pivot their operations quickly enough to get the most out of these technologies, then soon (sooner than we think) someone else will swoop in and disrupt the party.

Cummings agreed: “In an ideal world, you would clean sheet design a mine with all of this knowledge at hand.

“That being said, there’s a lot that you can do inside these legacy systems to improve them with relatively low cost overhead compared to having to buy new trucks and shovels,” she said. “There’s a lot that can be done, but it takes a huge mind shift.

“The dream is that you could throttle a mine or pull back to match product demand. In this world of COVID-19 we hear about stockpiling of resources and that happens because of the lack of ability to adapt to these accordion effects in supply and demand.

“But if you could optimise those processes and use a little bit more intelligence to help get you there, and understand how to adapt these planning tools, I think mining could get there.

“I’m very optimistic that one day that the throttle mine will happen. I don’t know if it’s going to happen in my lifetime, but we have all the puzzle pieces that we need.”

Articulate, emulate, collaborate

I asked Cummings if there are any other industries we could look to for help or inspiration in advancing our use of autonomy?

“Construction, agriculture, military,” she replied.

“Networking is big in agriculture and it’s pushing that industry ahead quite quickly. They can network crop dusters for example, or tractors in a field. The machines can all talk to each other and share information. The military would equate this to swarm technology. There’s definitely a lot that can be learnt from those.

“For example, in mining, instead of having autonomous trucks just run their cycles over and over, the trucks could actually start working out between themselves where they need to be and go somewhere else in the mine. They could start doing a little bit more intelligent planning.”

As we were on the topic of AI, I asked Cummings to quash the ‘robots are going to steal our jobs’ mentality once and for all.

“The reality is these are not sentient beings. Robots still only work well on incredibly narrow applications. I do not have any concern at all,” she said definitively. “There are some worlds I work in where there’s a legitimate argument.

“For example, if we had driverless cars en masse, which we don’t, but if we did, what would happen to all the drivers of the world who depend on that for a living wage?

“Not in mining. Mining is dangerous, we need to get people out. Even people who work in and around the dusty communities, there’s long term health implications even if we take out the short-term safety implications of working around dangerous equipment. Mining has always been the best case to demonstrate the use of robots.”

Igniting conversation

I always end interviews by asking if the interviewee has anything they want to add… Anything important that I’ve missed?

“No, I think I’ve stomped on everyone’s parade in some way shape or form,” Cummings said with a wry smile.

Aside from our shared interest in technology, that’s obviously what Cummings and I have in common; an irresistible urge to kick the hornet’s nest from time to time.

That’s what makes people talk. And talk is the first step towards progress.