AustraliaPublic sector2022 – ongoingNational Heavy Vehicle Regulator
Machine learning models to target fatigue and crash risk in heavy vehicles
Australia's heavy vehicle regulator used machine learning to find higher-risk drivers and vehicles, and fatigue infringements found rose 4%
Problem
The National Heavy Vehicle Regulator oversees nearly 1 million heavy vehicles across Australia. It needed a way to focus roadside checks and enforcement on the operators and vehicles most likely to cause serious crashes.
Approach
The data science team ran machine learning experiments on the Databricks platform and reused templates to speed up delivery. It built a fatigue engine to flag drivers and operators likely to cause fatigue-related crashes. It also built a crash prediction model that combines vehicle, defect and crash data, and further engines for risks such as roadworthiness and over-mass vehicles.
Results
Takeaway
Targeting enforcement with a risk model gave the regulator a measurable lift in the infringements it found. The page does not say how the 4% was measured.
Categories
- Industry
- Public sector
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Off-the-shelf tool or platform
- Goal
- Quality and risk
- Evidence
- Vendor case study
Sources
Source summaries
Databricks describes how the regulator built machine learning engines for fatigue, crash risk, roadworthiness and over-mass vehicles. It reports a 4% increase in yield of high-risk fatigue infringements. It also reports that the crash prediction model found a cohort with a 1-in-38 daily chance of a fatal or serious incident.
Full details are in the original sources.
