Latest / Internet of Things with Fexingo: Connected Devices, Sensors, and Industrial IoT / How IoT Sensors Are Preventing Mining Vehicle Collisions
Transcript
- Lucas: So, last month a haul truck driver at a Rio Tinto iron ore mine in Western Australia was pulling a loaded 240-ton truck up a ramp when a light vehicle — a Toyota Land Cruiser — pulled into his blind spot. Luna: That sounds like a recipe for disaster. In mining, those blind spot collisions are one of the leading causes of fatalities. Lucas: Exactly. But in this case, the truck's onboard IoT system detected the Land Cruiser via radar and GPS, triggered an audible alert in the cabin, and when the driver didn't react fast enough, it automatically applied the brakes. The truck stopped about four feet from the Land Cruiser's roof. Luna: So the system actually intervened before the driver even realized the danger. How common is that kind of tech in mining right now? Lucas: It's becoming the standard for the biggest operators. Rio Tinto, BHP, and Anglo American are retrofitting their fleets. But it wasn't always this way. Ten years ago, the approach was mostly just training and mirrors. The problem is, when you're sitting 15 feet off the ground in a cab with massive blind spots, no amount of training overrides physics. Luna: Right. So the sensors are essentially giving the truck a 360-degree awareness that the driver can't have physically. What's the sensor stack look like? Lucas: It's a combination of short-range radar around the chassis — typically four to six units — plus a roof-mounted GPS and a mesh radio that talks to other vehicles and to a central server. The radar detects objects within about 30 meters. The GPS gives position accuracy within about 10 centimeters. And the mesh radio sends that data to every other equipped vehicle in the pit, so everyone knows where everyone is in real time. Luna: So it's vehicle to everything, or V2X, but specifically for an off-road mine environment. That's different from the V2X you'd see on public roads because the terrain is constantly changing. Lucas: Exactly. In a mine, the roads shift every day as blasting and excavation change the landscape. So the system has to be dynamic. It can't rely on a fixed map. The vehicles are constantly updating their positions relative to each other and to the current haul road geometry. Luna: I bet false alarms are a concern. If a system brakes automatically too often, drivers might start to distrust it or even disable it. Lucas: That was a real problem in early deployments. The first generation of proximity detection used magnetic fields and had a very high false-positive rate — like 20 false alarms per shift. Drivers hated it. The newer radar-based systems are down to about one false alarm per shift, which is tolerable. And the system logs all events, so if a driver disables it, that's a reportable incident. Luna: That's a smart accountability loop. So what's the cost per vehicle, and how do operators justify it? Lucas: The hardware and installation runs about $8,000 per vehicle for a light vehicle, and maybe $15,000 for a haul truck because it needs more radar units. That's not cheap when you're equipping 500 vehicles across a site. But consider this: a single haul truck collision that puts a truck out of service for two weeks costs about $2 million in lost production and repair costs. So if the system prevents just one collision every couple years, it pays for itself many times over. Luna: And that's just the direct cost. There's also the human cost — fatalities and serious injuries. Mining companies have been under pressure from regulators and unions to adopt this tech. Lucas: That's a big part of it. The Western Australian government now requires proximity detection on all new haul trucks, and existing trucks have to be retrofitted by 2027. So it's becoming mandatory, not optional. Luna: I wonder if the same approach could work in other heavy industries — like construction or agriculture, where you have big machines and blind spots. Lucas: Absolutely. In fact, some construction equipment manufacturers are already testing similar systems. Caterpillar has a product called Cat Detect that uses radar and cameras. And in agriculture, John Deere has a system for combines that prevents them from running into grain carts. The sensors are essentially the same; the software just has to adapt to the specific vehicle dynamics and environment. Luna: So the core IoT platform is becoming a commodity. The differentiation is in the integration and the algorithms that decide when to brake. Lucas: Exactly. And that's where the real engineering challenge lies. You have to balance safety with productivity. If the system brakes too conservatively, you slow down the haul cycle and reduce throughput. Too aggressively, and you risk causing a rear-end collision with the truck behind you. So the algorithms have to account for vehicle weight, road grade, weather conditions — a wet road means longer stopping distance. Luna: That's a lot more complex than I imagined. I thought it was basically 'object detected, stop immediately.' Lucas: No, if it did that, you'd have trucks skidding off ramps. The system actually predicts the trajectory of both vehicles and applies graduated braking — first an alert, then partial braking, then full braking if the driver doesn't respond. And it does this in coordination with the other vehicles via the mesh network, so the truck behind you gets a 'forward collision warning' even before you brake. Luna: That's impressive coordination. It's like a miniature air traffic control system for the mine pit. Lucas: That's a perfect analogy. And the data from all these events gets fed back to a central operations center, where they can analyze near-misses and identify high-risk zones in the mine. Over time, they can adjust the haul road layout to reduce those risks — for example, by widening a curve where multiple near-misses have occurred. Luna: So it's not just reactive safety; it's also generating proactive insights. That's the real long-term value of IoT in this context. Lucas: Right. And this is a good moment to mention — if you found this breakdown useful, the reason we can do deep dives like this without any ads or sponsors is listener support. People who get value from the show buy us a coffee — it's buy me a coffee dot com slash fexingo. Luna: Yeah, it genuinely makes a difference. Even one coffee helps keep the show independent and allows us to spend time researching topics like this. Lucas: So back to the mine — one of the more fascinating developments is that the system is now being used to detect not just vehicles, but also pedestrians. Miners on foot are given wearable tags that broadcast their position, so even if they step into a truck's blind spot, the truck knows they're there. Luna: That's a whole new layer. Are those tags the same as the vehicle sensors, or a different technology? Lucas: They're different. The pedestrian tags use ultra-wideband radio, which has much finer resolution than GPS — about 30 centimeters. The tags are worn on the helmet or belt, and they transmit a unique ID to the mesh network. So the system can tell the difference between a person and a vehicle, and it gives different alerts. For a pedestrian, the truck will apply brakes more aggressively because the risk is higher. Luna: And what about driver fatigue? That's still a factor even with all this tech. Lucas: That's actually a separate system that's often integrated. Many mines now have in-cab cameras that monitor the driver's eye movements and head position. If the system detects microsleeps or prolonged distraction, it alerts the driver and can even slow the truck down autonomously. So you have multiple layers of safety: the driver monitoring, the proximity detection, and the V2X coordination. Luna: It's almost like a safety cocoon around the vehicle. I can see why regulators are pushing for it. Lucas: And the data is compelling. At Rio Tinto's mines that have fully deployed the system, they've seen a 70 percent reduction in vehicle to vehicle collisions and a 90 percent reduction in vehicle to person incidents. Those are huge numbers. Luna: So what's the next frontier? What's the limitation of the current tech? Lucas: One big challenge is communication reliability in deep pits. The mesh network relies on line of sight between vehicles, but if a truck goes around a corner or down a steep ramp, the signal can drop. So companies are deploying fixed relay towers at key points to ensure coverage. Another challenge is interoperability — if you have a Rio Tinto truck and a contractor's truck that uses a different system, they might not talk to each other. There's a push for a common standard, but it's not there yet. Luna: That's a classic IoT problem: the hardware is cheap, but the ecosystem integration is hard. It's not just about sensors; it's about getting everyone to use the same protocol. Lucas: Exactly. And that's where the industry is headed. The goal is a completely connected mine where every machine, every person, and every piece of equipment is on a common network, and the system can coordinate movements across the entire site autonomously. Some mines in Australia are already operating fully autonomous haul trucks — no driver at all — and the proximity detection system is a stepping stone toward that. Luna: So the collision prevention IoT we talked about today is essentially the safety layer for a future where humans and autonomous machines share the same space. Lucas: Exactly. It's a critical enabler for full autonomy. Without reliable collision detection, you can't safely put autonomous trucks alongside manned light vehicles. So this sensor tech is laying the groundwork for the mine of the future. Luna: And it's saving lives today. That's a pretty good combination. Lucas: Absolutely. And the numbers prove it. So if you're in an industry with big machines and blind spots, this is definitely a tech to watch.