Epiroc Mobilaris Mining Operations

by Nick Clark | Published April 25, 2026 | PDF

Epiroc's Mobilaris platform is one of the leading mining-intelligence and fleet-coordination products deployed in underground and open-pit operations worldwide. It integrates positioning, traffic management, production reporting, and increasingly autonomous loader and hauler tasking across heterogeneous equipment fleets. Yet at the level where digital intent becomes physical motion — a 60-tonne loader committing to a tramming path past personnel, a long-hole drill rig committing to a blast-pattern position, an autonomous truck committing to a ramp-grade descent — the platform expresses commands as binary issue-and-execute rather than as graduated, reversibility-evaluated, post-verified commitments. AQ's governed actuation primitive supplies precisely that missing structural layer: every actuator commitment selects from a graduated mode set (continue, defer, refuse, partial), is gated by a harm-minimization evaluation against the credentialed configuration of the operating context, and is followed by a post-actuation verification step that returns into lineage. This article describes the structural gap in Epiroc Mobilaris and the AQ primitive that fills it as a freedom-to-operate disclosure.


1. Vendor and Product Reality

Epiroc AB is a Swedish multinational and one of the two dominant suppliers of underground and surface mining equipment globally, alongside Sandvik and Caterpillar. After acquiring Mobilaris Mining Intelligence in 2021, Epiroc absorbed an established mining-IT product line whose flagship Mobilaris Mining Intelligence (MMI) and Mobilaris Onboard suites are deployed in operations such as LKAB Kiruna, Boliden, Agnico Eagle, and dozens of other underground mines across Scandinavia, Canada, Australia, and South America.

The platform consolidates ultra-wideband and Wi-Fi-derived underground positioning, vehicle telematics, production reporting, and traffic-management views into a single operational picture. On top of this real-time picture, Epiroc layers its 6th Sense automation portfolio — Scooptram Automation for autonomous loaders, Pit Viper autonomous drilling, and remote-operations centers for both underground and surface fleets. Customers increasingly run mixed fleets where Epiroc autonomous loaders share ramps and intersections with human-operated trucks, and where blast-clearance, ventilation, and refuge-chamber states are computed centrally and pushed to equipment.

Operationally, the architecture is command-flow oriented: the central Mobilaris layer computes a route, a clearance, or a production directive; the equipment executes; telemetry returns; reports aggregate. Authority for an actuation lives implicitly in the role of the operator or the dispatch user, and exceptions surface as alerts rather than as architecturally-defined refusals. This is consistent with the rest of the mining-IT industry, but it is precisely the layer at which regulators (MSHA, the Australian WHS regime, the EU Machinery Regulation, ISO 17757) are tightening expectations around demonstrable harm-minimization and audit-grade post-actuation verification.

2. The Architectural Gap

Mobilaris and the broader Epiroc autonomy stack lack a structural property that the regulatory and operational frontier increasingly demands: graduated actuation modes with credentialed reversibility evaluation. When the platform issues a tram command, a drill commitment, or a haul-cycle continuation, the underlying command model is binary — the equipment either executes or it does not. There is no architectural distinction between "execute fully," "execute partially under a degraded configuration," "defer until a credentialed observation refreshes," and "structurally refuse." Refusal, when it occurs, comes from interlocks and emergency-stop logic rather than from a first-class composite admissibility decision.

Three concrete consequences follow. First, harm-minimization in safety-critical maneuvers (tramming past a refuge bay during shift change, descending a ramp with degraded brake telemetry) is delegated to interlock libraries written by equipment OEMs rather than evaluated as a structural property of the commitment. Second, reversibility is not architecturally evaluated: a commitment to begin a long-hole drill cycle and a commitment to begin a blast-clearance ventilation purge are processed by the same command channel even though they have radically different reversal costs. Third, post-actuation verification is implicit: telemetry returns, but the platform does not architecturally close the loop by treating the verification result as a credentialed observation that re-enters the admissibility chain for downstream commitments.

This gap is structural, not implementational. No amount of additional sensors, additional dispatch UI, or additional ML on the Mobilaris side closes it without introducing the missing primitive. That primitive is what AQ supplies.

3. What the AQ Governed-Actuation Primitive Provides

AQ's governed actuation primitive specifies that every actuator commitment in a conforming system pass through a graduated mode selection from a defined set — at minimum continue, defer, refuse, and partial — with the selected mode produced by a composite admissibility evaluation rather than a binary permit/suppress gate. The mode set is technology-neutral: a tramming path, a blast-clearance directive, a refuge-chamber lockdown, and a ventilation reconfiguration are all expressed in the same architectural shape, even though the implementing actuators differ.

Each commitment carries an explicit harm-minimization evaluation parameterized by the credentialed configuration of the operating context. For a Scooptram tramming past a refuge bay, the configuration includes the credentialed presence list, the credentialed brake-system class, and the credentialed ventilation state; the actuation mode is selected to minimize the worst-case harm under that configuration rather than the expected-case throughput. For a long-hole drill rig, the configuration includes credentialed blast-pattern, credentialed personnel-clearance, and credentialed ground-control state. The harm-minimization evaluation is structurally distinct from the throughput evaluation, and both contribute to mode selection.

Reversibility is evaluated as a first-class property of the commitment. Commitments with low reversal cost (a tram-stop request, a drill-pause request) are admissible under weaker credentialing; commitments with high reversal cost (a blast initiation, a stope-fill commitment) require stronger credentialed observation and stricter mode selection. The reversibility evaluator is itself a credentialed component, and its output enters lineage.

Post-actuation verification is structurally required: every commitment produces a verification observation that re-enters the chain at the observation layer, allowing downstream commitments to admit or refuse based on the verified state rather than the commanded state. This recursive closure is what distinguishes governed actuation from a flowchart of dispatch operations and what makes the substrate auditable end-to-end against MSHA, EU Machinery Regulation, and ISO 17757 expectations.

4. Composition Pathway

Epiroc integrates the AQ primitive as a substrate beneath Mobilaris and 6th Sense without replacing either. The Mobilaris dispatch and operational-picture layer continues to compute routes, production directives, and clearance states; what changes is that each computed directive is expressed as a candidate commitment to the AQ governed-actuation layer rather than as a direct command to the equipment. The substrate then resolves the commitment into a graduated mode and emits the corresponding actuator command.

Authority credentialing maps cleanly to existing mining roles: shift bosses, blast supervisors, ventilation officers, and remote-operations operators each carry credential classes whose authority scope is published in the operation's governance configuration. Credentialed observations from underground positioning, gas monitoring, ground-control monitoring, and equipment health flow into the admissibility evaluator. Lineage records — observation, weighting, mode selection, verification — accumulate against the operation's audit-grade store and become available to MSHA inspections, internal incident review, and insurance underwriting without bespoke ETL.

For the OEM autonomy products (Scooptram, Pit Viper, Boomer), AQ integration is at the commitment-issuance boundary: the autonomous controller proposes a commitment, the substrate evaluates and selects the mode, the controller executes the resolved mode, and the verification observation closes the loop. Existing interlock libraries continue to operate as a final safety net, but the architectural locus of harm-minimization moves into the substrate, where it can be reasoned about, audited, and certified once rather than re-litigated per equipment platform.

5. Commercial and Licensing Implication

Licensing is structured as a per-commitment substrate license to Epiroc, with sub-licensing to fleet operators bundled into Mobilaris and 6th Sense subscriptions. Epiroc gains a defensible architectural moat against Sandvik and Caterpillar that does not depend on equipment differentiation: the substrate is what allows fleet operators to demonstrate audit-grade harm-minimization to regulators and insurers, and the substrate is licensed exclusively into Epiroc's mining channel for the relevant field.

Customers gain three concrete benefits. They gain a single architectural locus for safety-case demonstration that survives equipment turnover and OEM change. They gain insurance-grade lineage that materially reduces premiums on autonomous-fleet endorsements. And they gain a structural answer to the regulatory drift toward demonstrable, post-verified, reversibility-aware actuation that is otherwise impossible to retrofit into command-flow architectures. For Epiroc, the result is a substrate-level differentiator at exactly the architectural layer where the next decade of mining-autonomy regulation is converging.

Nick Clark Invented by Nick Clark Founding Investors:
Anonymous, Devin Wilkie
72 28 14 36 01