Progress, milestone by milestone.
The project runs from December 2025 to December 2027 across nine milestones. Here's what has been delivered and what comes next.
Stage 4: AI system development.
With the design complete, this stage turns it into working models and interfaces, ready for co-design and testing with fleets.
Detect and predict fatigue
Train algorithms that pick up fatigue in real time and can explain their results.
Alert the driver
Personalised warnings that reach drivers early without getting in their way.
Fatigue Risk Advisor
Connect it to the recommender system so fleets can see risk across all their drivers.
- Lock the data dictionary, labels and data split
- Benchmark the feature extractors on public datasets
- Sync the camera, wearable and GPS on the edge device
- Build the in-drive alerts and post-trip recommender
- Add the driver and fleet dashboard views
- Collect test-drive data using the 1–5 fatigue scale
- Retrain and validate on drivers the model hasn't seen
From planning to final report.
Dates for future milestones are indicative.
Agreement signed and foundation laid
The funding agreement was signed and foundational research began: literature review, behavioural risk mapping, an AI and data scan, and early ethics and privacy work.
Stage 1: Planning
- Project working group and partnerships established
- Governance, ethics and evaluation frameworks defined
- Stakeholder engagement and communications plans
- Industry partner sessions under way
Stage 2 begins: Data collection
- Logistics pilot started with a small internal group
- Time-aligned streams tested: video, wearable, motion and GPS, and vehicle data
- Iterative data reviews with Murdoch University
- Driver onboarding and consent materials developed
Stages 2 and 3: AI system design
- Measurement framework across six evidence sources
- Common dataset structure with participant-level data splits
- Three-part fatigue label: class, score and confidence
- Three-layer design: reactive, predictive and recommender
- Standalone, vehicle-agnostic hardware concept
Stage 4: AI system development
Training explainable detection and forecasting models, building personalised driver alerts, and connecting the Fatigue Risk Advisor.
Stage 5: Co-design and first phase of testing
Dashboards co-designed with drivers and fleet managers, tested for usability, clarity and trust, then trialled with partner fleets.
Stage 6: Second phase of testing
Interventions refined from feedback, with changes in fatigue risk measured at driver and fleet level.
Stage 7: Training and toolkits
Driver and manager training modules, and practical implementation toolkits tailored for smaller operators.
Stage 8: Evaluation and reporting
Final evaluation, industry sharing of findings and the final report to the NHVR.