Autonomous vehicle solutions are becoming a broader trend in the mining industry as mid-sized miners and service providers glimpse the potential already seen by majors to achieve safer and more efficient operations.
However, mining companies of all sizes are still intimidated by the advanced technological infrastructure and support systems that many believe are needed for autonomous deployments. There is also the threat of legal and regulatory issues related to liability, data privacy, and cybersecurity.
To discuss these quandaries and more, Mining Magazine editor Craig Guthrie sat down at ConExpo 2023 with Ben Miller, chief mining officer at Pronto, an autonomous vehicle solutions firm that has developed a retrofit solution for various mining equipment including haul trucks, loaders, dozers, and drills.
In 2021, Pronto revealed that it had been working with Bell Equipment Engineering R&D (Bell) for around 18 months to engineer and refine Pronto's hardware and software to produce and prove a self-drive Bell articulated dump truck.
Bell provided a brand new B50E as a test unit to Pronto in November 2020 and this unit has since been operating on several mine sites on the west coast, allowing Pronto and Bell the time and experience to be self-driving on a mine site through a full load and haul cycles.
The Pronto system uses cameras mounted on the truck to observe the road and job site environment. It then feeds that data into neural networks that analyse every image in conjunction with other inputs to make optimal driving decisions.
MM: Do you believe the autonomous vehicle of the future will shift from being bigger machines to smaller swarm vehicles?
BM: We are currently conducting a study with Whittle Consulting to determine whether the trend will be towards smaller machines. I believe we cannot go all the way down to the little tiny swarms due to the congestion issue.
Also, the deposit dictates the size of the loading tool. Therefore, deposit-based decisions will remain stable. Logistics and demand will also influence the selection of the loading tool. The Bell trucks, for instance, are road portable, providing a significant advantage.
However, for over a 60-ton truck, breaking it down to move it becomes necessary. We may see a push towards the 110 class. Mostly because of logistics chain behind building and maintaining, providing consumables, and 210 trucks are very robust. I think you're going to see a push down in size, but probably not too much as there are a lot of decisions that are not autonomy-based.
MM: What are the main challenges in the adoption of autonomous vehicles?
BM: Most companies are seriously considering autonomous vehicles, but there are still laggards who are resistant to change. There are some frictional elements like labour, jobs, adoption of technology, etc.
Also, there is pressure from smaller operations sometimes run less rigorously. Autonomy does not do well with constant change - we need to find a good clean process and then automate it. Autonomy is about making things less dumb, so we need to strip things away before we automate them. The construction materials and industrial minerals producers are starting to push towards autonomy. The tier one companies have already adopted autonomy because they have rigorous processes in place.
MM: Do you have any differences in terms of your applications, whether it's a diesel vehicle or a battery electric?
BM: It doesn't affect autonomy too much. Some fleets of trucks push towards battery electric, but that is only suitable for sites with heavy infrastructure in place. The sites that use ADTs often have infrastructure right there, and they are the front line in putting into development. The Bell trucks can use HvO fuel, which is one of those options to lower the carbon footprint. They are fully enabled on these Mercedes, and we'll see quite a bit of that still going on in the mobile machinery space.
MM: What are your thoughts on the backlash against battery electric vehicles and the resources taken to create massive batteries?
BM: The application of a battery paired with a trolley assist is really interesting prospect because the battery size is quite small. You're essentially on the pit floor, and then you're on the trolley, and then you're dumping and then you're back on the trolley. So the battery size can be quite small. There are still a lot of low hanging fruit before we make a wholesale transition to battery surface vehicles. Underground, the argument for underground electric is very clear, and it has a lot more to do with ventilation constraints.
MM: And in terms of change management, when you get a project and you get more and more, how do you approach that? Obviously, it can be quite frightening to some when a big vehicle is being made autonomous.
BM: From the ultra-class side, there's a lot of meetings and time spent on change management. When you are dealing with a construction site, construction materials or industrial minerals, you typically only have three to five trucks that the primary project involves.
The number of people involved is lower, and you're able to have more one-on-one time with them. You're actively training them to use a system with a more one-on-one approach. The labour pressures we see, of not having to run the machines, are much more obvious when there are only three operators and one isn't present..
Operators in smaller can feel the impact of teammates not being present more acutely. One nice story I tell them to make them comfortable is that if all of us in this room had one mule, and we loaded that mule down, and we passed it around, that would work.
But there are other ways where we can choose to put up a string of mules together and have one person in charge of that string. We can then take those other workers and use them as parts of the process that are lacking.
Operations often have a lot of gaps and are trying to fill them. So autonomy is a way to stretch the labour that we have. I think it also provides an opportunity for the economy to draw younger people back into the industry. We definitely have a demographic issue within the industry where we're quite heavy on the 50-60 year olds, and quite thin on the 20-year-olds. It provides an opportunity there. We do have to manage it correctly, but there's a path we need to follow.
MM: In terms of the machine learning that you use, do you have the datasets for it? Are they mature datasets based on the projects? Or do they start as young ones that don't factor in?
BM: We have very mature base datasets coming from on-road and other data sources. But once we're up and running, we will typically work to characterize the site specifically. It may be that the soil colour texture has some difference. It could be that the uniforms that are worn are slightly different. So, we will train additional training on that.
In the tier one deployment, we were performing a series of commissioning tests, which involve testing against mannequins to make sure that we do pickups testing against vehicles testing and spillage. Typically, those tests are run midday, but we struggled on through the day and got to the point where the sun was setting and it was straight in aligning with the camera.
We had an opportunity to collect data around that specific case where the sun is in your eyes. We had a big breakthrough on some of our ways of managing that because we had an opportunity to collect that data. So all of our sites, we gather logging desks on a regular basis from the trucks. If we notice issues around certain detections, whether they're false positives or we aren't detecting or characterising them correctly, we'll take those logs and then actually handle eight or eight logs and then feed it back into the system.
