Autonomous Operations — Rail & Haulage

Context

Large-scale mining operations with autonomous haulage vehicles and SCADA-based rail monitoring infrastructure operating in safety-critical environments alongside human-controlled systems. IT/OT convergence, determinism, and failure containment were primary design constraints. Details abstracted in line with sensitivity requirements.

Role and mandate

Solution architecture responsibility across both autonomous haulage and rail monitoring programmes — IT/OT boundary definition, EDP design, security control alignment, and remote access architecture across operational and enterprise domains.

Architecture problem

Enable data exchange and remote operational visibility between OT field systems and enterprise platforms without introducing non-deterministic failure modes, unsafe control paths, or attack surface expansion across safety-critical operational boundaries.

Constraints

• Human safety and regulatory compliance — failure containment treated as a design-first requirement.
• Strict separation of control and observation planes across autonomous vehicle systems.
• Legacy field protocols and heterogeneous vendor estates across haulage and rail systems.
• Remote access requirements for operational monitoring without expanding control plane exposure.
• Low tolerance for operational disruption in continuous mining operations.

Decisions and trade-offs

• Explicit IT/OT zoning with protocol mediation via gateway patterns — unidirectional data flows for telemetry where bidirectional control was unnecessary.
• EDP architecture designed around data diode patterns for safety-critical telemetry paths.
• Identity and network boundaries treated as safety mechanisms — not convenience layers.
• Remote access architecture scoped to observation only — no remote control path into autonomous vehicle systems.
• Rail SCADA monitoring isolated from enterprise integration points with explicit boundary controls.

Outcomes

• Reduced blast radius for operational faults and cyber events across both programmes.
• Clear auditability of control boundaries and responsibilities for safety-critical systems.
• Remote operational visibility achieved without compromising control plane integrity.
• Improved confidence in autonomous system isolation, failure containment, and recovery assumptions.