Please note: The algorithm descriptions in English have been automatically translated. Errors may have been introduced in this process. For the original descriptions, go to the Dutch version of the Algorithm Register.

Predictive maintenance roads

To ensure road quality and safety, an algorithm has been developed to provide insight into road wear. The algorithm predicts required maintenance on South Holland N roads using existing historical measurements.

Last change on 8th of November 2024, at 13:32 (CET) | Publication Standard 1.0
Publication category
High-Risk AI-system
Impact assessment
Field not filled in.
Status
Out of use

General information

Theme

Space and Infrastructure

Begin date

Field not filled in.

Contact information

Voor vragen en opmerkingen kan je terecht bij: digitaalzuidholland@pzh.nl. Wil je bezwaar maken, dan kan je terecht bij de Juridische afdeling van dienst beheer organisatie. https://www.zuid-holland.nl/contact/.

Link to publication website

https://d87b57413625f4d57879c589.azurewebsites.net/

Link to source registration

https://www.zuid-holland.nl/politiek-bestuur/feiten-cijfers/algoritmeregister/algoritme-voorspellend-onderhoud-wegen/#h1a00d4f6-1721-4852-8ba1-5603649d9ba9

Responsible use

Goal and impact

The results of the algorithm will help develop multi-year maintenance planning for provincial roads. For now, it is still a pilot. The model predicts in certain years when a road falls below certain threshold values. This is a signal that maintenance needs to take place. The end goal is to carry out proactive road maintenance.
The algorithm helps to make better multi-year maintenance planning for roads. As a result, you theoretically have less chance of unnecessary maintenance. You do the maintenance only where it is needed and at the right time. When the algorithm works perfectly, there is also less unexpected maintenance. The benefit for road users is higher road quality. This can prevent accidents. Maintenance is realised on time.

Considerations

The algorithm helps make better predictions about road maintenance and bets on explainability. Province of South Holland finds it important that the algorithm is not a black box, but that it is clear how the algorithm arrived at a particular prediction. Weighting is applied to see which type of road damage leads to a particular prediction. It also explains road maintenance interventions using principal values.

Human intervention

Yes, the algorithm serves purely as a tool for multi-year maintenance planning (MYP) to roads. There will always be a check on the input of the data (including road maintenance indicators) and the output (how does an algorithm arrive at a prediction) of the algorithm before a human decision is made for in the maintenance planning. Besides, an end-of-life prediction is only one of the factors for determining the maintenance timing: the environment, traffic management, finances, capacity, strategy, etc. also play a role which means that a human consideration is always made for the MYP.

Risk management

There could potentially be unnecessary maintenance and costs if the predictions from the algorithm turn out to be wrong afterwards. These risks only apply if you assume that the algorithm alone is relied upon. However, this is not the case.

Legal basis

Road management system CROW

Links to legal bases

CROW: https://www.crow.nl/

Operations

Data

Census point data, KNMI weather data, ARAN, DBI areal data

Technical design

Random Forest

External provider

This web app was created by Geronimo.AI on behalf of the province of South Holland in a Startup in Residence.

Similar algorithm descriptions

  • Images of road sections are acquired during inspections and processed by the algorithm. This algorithm gives bumps and damages a score from 1 to 9. Based on these scores and their spatial distribution over the road sections, the algorithm provides suggestions for maintenance.

    Last change on 5th of February 2025, at 8:43 (CET) | Publication Standard 1.0
    Publication category
    Other algorithms
    Impact assessment
    Field not filled in.
    Status
    In use
  • Smart deployment (optimal allocation) of road inspectors on the main road network based on historical and current incidents.

    Last change on 17th of May 2024, at 9:11 (CET) | Publication Standard 1.0
    Publication category
    Other algorithms
    Impact assessment
    Field not filled in.
    Status
    In use
  • Prediction model providing risk scores on road accidents based on a broad dataset of road characteristics and accidents.

    Last change on 23rd of August 2024, at 15:37 (CET) | Publication Standard 1.0
    Publication category
    Impactful algorithms
    Impact assessment
    Field not filled in.
    Status
    In development
  • It is a traffic model to provide city-wide insight into the nuisance caused by activities (construction works, road works and events) on the Bereikbaarheidsregie network (the Hague's main traffic network), in order to safeguard Hague's Accessibility as much as possible.

    Last change on 23rd of August 2024, at 15:22 (CET) | Publication Standard 1.0
    Publication category
    Other algorithms
    Impact assessment
    Field not filled in.
    Status
    In use
  • Inspection of provincial roads using LiDAR or laser technology. Using these techniques, the road section can be scanned in great detail and all unevenness highlighted. From this, the algorithm selects damage images on the basis of which the province will see what maintenance is needed here.

    Last change on 5th of February 2025, at 9:11 (CET) | Publication Standard 1.0
    Publication category
    Other algorithms
    Impact assessment
    Field not filled in.
    Status
    In use