Why it matters
Without AI, decompression plans risk misalignment with gas management, actuator timing, and SMFR safety checks, leading to potential out-of-gas events or decompression illness claims.
From design to safety
Design
AI-assisted decompression planning design
The system reads proposed deco schedules, gradient factor targets, gas-switch rules, and early-stage gas management constraints from the manifest and equipment inventory to validate ascent/descent paths and rebreather compatibility with cylinder states; cross-checks include gas purity constraints and emergency gas priorities.
- deco schedules
- gradient factors
- gas switches
Build
Validation during equipment fit-out
Data from stage cylinders, runtime tables, and gas switch manifolds are captured, cross-checked against buoyancy, inventory, redundancy requirements, gas purity logs, and oxygen clean records for safe transport; setup includes pressure vessel labeling and cylinder tethering to deck racks.
- stage cylinders
- runtime tables
- gas switches
Operation
Real-time plan monitoring
Live dive roster synchronizes with dive computer cloud logs, gas pressures, O2 sensor readings, and deco tracking feeds; AI confirms adherence to the planned profile and flags depth-time excursions, with automated alerts to crew and updated on-screen dashboard.
- deco schedules
- stage cylinders
- runtime tables
Maintenance
Plan and data integrity checks
Regularly verify updated gradient factors, re-train models against new gas mixes, audit oxygen clean records, log reagent and membrane changes, and verify calibration of gas analyzers and purge systems; maintain storage of calibration certificates and batch records.
- gradient factors
- runtime tables
- gas switches
Safety
Emergency action alignment
The AI flags deviations, recalculates decompression, updates the emergency action plan, and routes changes for sign-off by the dive supervisor and safety officer, ensuring redundancy with backup gas plans and surface support; ensure oxygen kit readiness and dive flag compliance.
- deco schedules
- gas switches
- stage cylinders
Part by part
- 1
deco schedules
- Watch for
- Decompression stops drift, incorrect stop depth, or skipped stops in the schedule; signs of inconsistent timing or mis entered depths
- AI action
- Flag deviations; trigger plan revalidation and supervisor sign-off; propose corrected stop sequence and updated dive roster
- 2
gradient factors
- Watch for
- Inconsistent gradient factor values or out-of-range GF60+ with potential under/over-buffering
- AI action
- Recalculate safety margins and propose adjusted profiles; log rationale in the plan history
- 3
gas switches
- Watch for
- Gas-switch timing misalignment or incompatible gas mixes; cross-match with PO2 targets
- AI action
- Suggest alternative gas order and verify with duty officer; update gas plan version
- 4
stage cylinders
- Watch for
- Cylinder count mismatch or end-of-service indicators not logged; gauge labels missing
- AI action
- Cross-check against dive roster; alert for reweighting or additional cylinders; enforce labeling controls
- 5
runtime tables
- Watch for
- Outdated runtime tables or missing BT (bottom time) entries; misalignment with dive computer logs
- AI action
- Update tables; notify tech supervisor for validation; archive older versions
- 6
oxygen kit cabinet
- Watch for
- O2 kits not calibrated or out-of-date SAR entries; mismatched shelf inventory
- AI action
- Flag expired kits; schedule service and oxygen clean checks; record inspection results
Worked example
Catching a safety drift in deco for a trimix charter
Illustrative scenario with sample data.
A mixed gas charter with two rebreathers requires precise deco for a 60-minute bottom at 60 msw.
Signal in
deco schedules: P02 0.21; GF38; gas switches: O2->He at 3rd stage; stage cylinders: 2x1500L E- size; runtime tables: 60 min bottom, 25 min ascent
System out
AI alert: plan deviates from safety margins; adjust GF to 44, re-run with updated dive plan; notify supervisor
Human sign-off: Diving supervisor reviews AI flag, approves updated plan, and confirms rebreather configuration before departure
How to roll it out
Phase 1
Weeks 1 to 3
Baseline data collection from current decompression plans and equipment; establish data feeds and logging conventions.
Phase 2
Weeks 4 to 8
Prototype validation with simulated dives; calibrate gradient-factor model, gas-switch logic, and deco stopping rules.
Phase 3
Weeks 9 to 12
Live pilot on small charter; monitor alerts, refine thresholds, and formalize safety sign-off workflow.
Where people stay in charge
- All safety-critical decisions remain with the dive supervisor or course director who signs off on changes to decompression plans.
- Automatic plan adjustments require human validation prior to execution and must be auditable in the dive log.
- AI outputs are advisory; records must reflect final approved plan and staff signatures for compliance.
Questions owners and crews ask
What exactly does AI check in a decompression plan for a technical dive?
AI checks deco schedules, gradient factors, gas switches, and stage cylinder allocations against safety margins, then flags deviations for supervisor sign-off. It also validates gas-millisecond timing harmonization with the dive computer inputs.
Who signs off on AI-suggested changes to decompression plans?
The dive supervisor or safety officer signs off on AI-suggested changes after reviewing the AI rationale and updated plan within the dive roster, with formal sign-off recorded.
Can AI handle mixed-gas and rebreather-specific deco rules?
Yes. The AI uses mix-specific deco parameters, rebreather control logic, and gas-switch timing to validate plan feasibility against equipment capabilities.
What records are produced by AI-assisted plan checking?
Audit-labeled plan revisions, sign-off timestamps, gas usage logs, oxygen kit status updates, and delta-logs from plan versus execution are stored for compliance.