Mines are well-defined and highly controllable areas, making them ideal testing grounds for the early commercial use of autonomous vehicles and equipment. Mining vehicles do not encounter unexpected events like impaired drivers or a distracted pedestrian crossing the road. The insulated mining environment aids the accelerated adoption of full autonomy while the commercial sector continues to train driverless vehicles for more dynamic circumstances.
Artificial intelligence (AI) was quickly embraced by the mining sector to gain new efficiencies amid dwindling resources. EARTH AI has been using machine learning algorithms to identify patterns of ore distribution based on past discoveries. They are able to highlight ore bodies and map out hard rock with extreme detail to increase their chances of identifying promising mining areas. Leveraging AI for autonomous driving builds on these operational gains while presenting new levels of safety. Companies like Caterpillar have led the way to fully autonomous mining vehicles, testing driverless bulldozers and dump trucks across global mining operations. Advanced machine learning and AI are the future for full vehicle autonomy in the mining sector.
Efficiency
Autonomous mining vehicles can operate continuously without the need for breaks or shift changes. These driverless vehicles rely on software, rather than operators, to perform predictable tasks such as positioning drilling equipment and hauling material. Companies using autonomous vehicles produced by Caterpillar reported a 30 percent improvement in productivity over manned operations. These autonomous vehicles use machine learning and AI algorithms to inform decision making processes and efficiently.
For example, machine learning in obstacle detection software can convert sensory data streams into usable information within the control system. It then plans the safest and most efficient pathway around an obstacle. AI algorithms help companies automate their decision making through a system that selects which type of vehicle they should use for that specific job. Additionally, 5G technology allows these vehicles to quickly process data through a stable high-speed connection. This is an integral function in mining operations since excavating sites are often located far from civilization.
Maintenance
In the demanding environments of mining operations where temperatures are either extremely high or low, these hulking vehicles need to be able to operate continuously. With autonomous operations, vehicle maintenance becomes predictable since these machines are programmed to operate in a controlled manner. Vale has been using AI algorithms to predict vehicular failures, monitor engine health and economize fuel consumption. The vehicles are already equipped with radar, laser sensors and GPS to avoid obstacles and reduce accidents. This AI can also analyze data and determine when a system will fail, ensuring repairs are made prior to malfunction and eliminating wasteful downtime.
Safety
One inherent obstacle for mining vehicles is that many operate below the surface where GPS is not available. So, companies like Sandvisk have created an intelligent system guided by lasers that map out and record a path for the autonomous vehicle. The system uses high dynamic range (HDR) technology to capture detailed image frames. It functions similarly to the way human eyes gathers information. This allows vehicles to see around corners and increase the drivers’ perception in darkness through infrared imaging. The system accurately directs the vehicle through a determined route, monitoring the speed and controlling steering, brake, hauling and dumping. It also decreases the chances of an accident and prevents unnecessary danger to the operator. Additionally, fully autonomous mining vehicles allow operators to work above ground or from a safer onsite location rather than among dangerous conditions.
Lower costs
Predictable maintenance and increased efficiency trims operational costs for mining companies. The use of machine learning for predictive maintenance reduces the number of overload events that can effectively halt an entire mining operation.
AI for sorting maximizes efficiency and works much quicker than traditional sorting methods. Sorting through worthless dirt and rocks is time-consuming and inefficient. Therefore, TOMRA employs machine learning and infrared sensors to examine every piece of material moving through equipment and sort based on multiple criteria (color, size, hardness). This efficient separation technology increases overall value of the deposit and reduces energy consumed during processing.
Premio: At the Core of Mining Autonomy with Industrial PCs
Premio’s industrial rugged compute solutions provide the building blocks for mining autonomy. Powerful processing, I/O flexibility and reliable connectivity are encased in durable, purpose-built hardware.
Rugged edge computers carry rich IoT versatility into challenging or treacherous settings. Strong processing supports advanced mining applications closer to the data sources and subordinate equipment systems. Fanless, cable-free designs remove chief failure points for hardware dispatched to punishing areas. Rich I/O options smoothly integrate legacy mining technology as well as leading-edge cyber-physical systems.
Premio’s solutions augmented to support intense GPU processing carry out intricate machine vision applications and advanced inference analysis of voluminous and varied data to make realtime decisions informing autonomous systems and vehicles.
Waterproof touchscreen computers provide intuitive HMI interaction between mine operators and autonomous systems. The WIO series are reinforced with high-utility stainless steel for lightweight durability and resilience against corrosive and caustic substances. IP66-rated enclosures prevent aggressive water and dust penetration, while 7H hardness glass resists scratches from sharp objects and grit.
Premio’s rugged edge and embedded compute solutions form the hardware backbone of an autonomous mining system. From prospecting to ground-breaking to site closure, every process improved by AI decisions and independent responses demands stable, versatile delivery devices.
