WA startup maps the road to Level 5 drilling autonomy
Autonomous drilling has been one of the mining industry’s most celebrated advances. Yet, despite the enthusiasm, publicly documented progress towards truly autonomous drilling remains limited.
WA technology startup drillEQ believes the industry has mistaken remote supervision for autonomy: impressive engineering inside closed, vendor-owned systems, but still no rig that adapts to changing strata conditions the way a skilled operator would.
Founded by former drilling professional Nathan Wehi, the company is on a mission to develop truly unsupervised autonomous drilling – Level 5 on the Global Mining Guidelines Group (GMG) autonomy scale, the benchmark the industry set for itself.
For all the enthusiasm around autonomous drilling, much of it has settled on artificial intelligence as the obvious engine of progress. drillEQ is deliberately cool on that assumption.
“I think of drillEQ as a pragmatic consumer of AI, not a chaser of it,” Mr Wehi said.
“We are solving deterministic problems because the industry requires it, and that’s a positive case of putting AI technology to use, not just talking about it.”
“No one’s solved unsupervised autonomy in drilling yet – whoever gets there first will do for this industry what the tri-cone bit did a century ago.”
Rather than pursuing automation through increasingly complex artificial intelligence alone, drillEQ has spent its first year establishing the data infrastructure and deterministic decision-making needed to support higher levels of autonomy.
The company’s work began with the deceptively simple challenge of listening to the drill rig – and every tool it has built comes out of that same roadmap.
drillEQ started by auditing unlabelled values, working out what a rig’s raw data actually meant before it could be trusted. That groundwork produced NODE, a plug-and-play module that captures machine data and pushes up to COMMS, an OEM-agnostic data pipeline developed with technology partner Genolis. Together, NODE and COMMS are the data foundation everything above Level 1 depends on.
Field trials with Villcore Mining have enabled the collection of production drilling data and validation of multiple measurement-while-drilling (MWD) outputs.
By GMG definitions, these achievements represent the threshold of Level 1 – the point at which reliable machine awareness becomes possible. While modest on paper, the milestone establishes the data foundation required before higher levels of autonomous decision-making can be achieved.
GMG’s guideline sets six levels, 0 to 5, for autonomous mining equipment generally. Applied to drilling specifically, Levels 2 and 3 are preset cycles, run unattended or supervised remotely, meaning the rig does not adapt to changing ground conditions.
Level 4 is the real break in the ladder. At this point, the system reads the ground and changes what it’s doing in response and handles its own fallback within a defined ground range.
Level 5 is that same capability across the full range of ground a skilled operator would handle, with no routine human intervention.
For Mr Wehi, that operational design domain (ODD) has become as important as the autonomy level itself.
“Every claim above Level 2 needs a stated ODD – the range of ground conditions, hole geometries and consumable states a system is proven to handle,” he said.
“A level without a domain isn’t a claim, it’s an advertisement.”
Embedded in the same roadmap are CENTRAL, a platform combining engine telemetry, fuel burn and MWD data, and BLOCK, which uses drilling information to map ground hardness and ore boundaries. Like NODE and COMMS, both are built around passive, listen-only integration.
Whether deterministic systems or AI define the future of autonomous drilling remains an open question. However, what is becoming clearer is that the industry’s path to Level 5 will depend as much on robust data, transparent engineering and clearly defined operating boundaries as it will on advances in machine intelligence.
