Capability. The route planner converts a structured mission task into ranked, explainable route options. Tasks, including start and end points, time windows, platform, speed, endurance, depth limits and exclusion zones, are captured in machine-readable form. Environmental and spatial data are fused in a semantic knowledge graph. Candidate routes are computed over a network whose edge costs reflect real conditions, and each route is returned with its segments, warnings and summary metrics attached.

How the routing engine works. Forecast ocean currents are ingested as time-stamped snapshots at multiple depths. Environmental values are computed for each network edge and each snapshot, so that geometries are not duplicated and costs are not recomputed from scratch. At request time, the engine selects the relevant snapshot and is designed to update only the edges within the active depth and exclusion constraints, which keeps responses fast. The same pass checks chart-data confidence and battery use, and calculates a risk index that flags gaps in the underlying data.

Highlights. The capability was built and demonstrated as a vendor-agnostic prototype for an Australian Defence program in geospatial integration, where the environment is three-dimensional and changes constantly. The prototype was delivered in March 2026 and demonstrated with operational users. Planners can ask questions about the mission and its data in plain English, answered from the governed knowledge base. Every planning step is exposed as an OGC API process, so that steps can be tested, replaced or extended independently.

Explainability and control. Planners see several candidate routes side by side and can override them. A recommendation is always accompanied by its evidence: the conditions it assumed, the constraints it honoured and the data it relied on, including where that data is weak. The workflow records provenance from tasking through to review, so that outcomes can inform future planning and decisions can be reconstructed afterwards.

Client benefits.

• Plans reflect the conditions a vehicle or crew will actually encounter.
• Risk is visible before commitment, not after.
• Decision-makers can approve with confidence, because the reasoning is presented alongside the plan.
• The open, platform-agnostic architecture avoids lock-in to any single vehicle, map engine or cloud provider.

The same pattern of constraints, live environmental data, route, warnings and evidence applies equally to emergency response, infrastructure inspection, field operations and logistics.