News - Workshop on BD Pave project hosted by NETIVEI Israel

FIRM 23 official logo_cropped 1.jpg


Forum of European National Highway Research Laboratories

Workshop on BD Pave project hosted by NETIVEI Israel

23rd February 2021


On February 23rd FEHRL’s Israeli member, Netivei Israel, hosted a workshop on BD Pave (Big Data for smart pavement management) where presentations on the work being undertaken in Europe and work underway in Israel were given.

The invitation only event was well attended, with 41 people online from Israel, Europe and the USA and from FEHRL institutes and National Road Administrations, including members of the Project Advisory Board for BD Pave. 

Edward Yosilevsky introduced and hosted the meeting, with Revital Levi, Head of Netivei Israel R&D department welcomed delegates. Martin Lamb, Programme Manager for BD Pave gave an introduction to FEHRL and its Strategic European Road Research Programme, from which BD Pave originated, then Dirk Jansen, BD Pave initiative leader gave an overview of the programme.

The technical presentations were started by Leif Sjögren, of  VTI, Sweden, who gave an overview of the work he has undertaken for WP1. Over 100 variables were recorded based on a survey of 15 countries. Leif discussed which sources were available, wanted and available from traditional sources, and how new sources could fill the gaps, especially as we develop digital twins. He felt we should concentrate on adding new needed/wanted information, investigate best use of sensor and data fusion (adding information from multiple sources) to develop combined indicators, determine data quality requirements and invest in work force skills for this new technical area.

Mahdi Rahimi of BASt then presented on data mining and techniques the cluster and sort data, determining what are outliers to the data. This was based on work undertaken in Germany on the Traffic Speed Deflectometer. These advanced statistical methods will help pavement engineers to make better decisions on pavement interventions and strategies.

Matteo Pettinari of the Danish Road Directorate, presented the LiRA (live road condition assessment using internal car sensors data) project, which explores how we can use car sensor data to measure road conditions. Modern vehicles have over 150 sensors that could be used to measure road conditions, with DRD operating a homogenous car fleet in and around Copenhagen, collecting CAN Bus data from the vehicles computer coupled with additional hardware for processing based on a Raspberry Pi. They hope to determine algorithms and models to determine 6 of the following 10 indicators with an accuracy of more than 80%, from vehicle data (1. Friction, 2. Cracking density, 3. Potholes, 4. Noise, 5. IRI, 6. Energy Expenditure, 7. Patched area, 8. Unevenness, 9. Rutting depth, 10. Texture depth).

Mrs. Revital Levi, presented work undertaken by Netivei Israel on a crowdsourcing approach for road marking condition assessment. The specific innovation here was a very quick competition turnaround based on a challenge statement of the problem of the difficulty of assessing road marking quality, to focus resources on where the conditions were poor. Three of the four companies who proposed a system used machine learning and AI, which is highly relevant to BD Pave.

Alex Shtein, presented the work undertaken by Netivei Israel on the use of data from multiple sources for their road maintenance management system. Netivei Israel have been collecting  various sources of pavement and bridge condition data for their Pavement Management System, Bridge Management System and  Safety Management Systems for around 13 years. They have a computer model which processes the data and enables decision support trees. The next phase of the programme will be to collect new sources of data to augment this, such as crowdsourced data. They have had a R&D trial using drones for inspection of bridges and also to create a digital twin.

Finally, Amit Arusi, presented work on the use of DL/ML methods in road maintenance applications. The goal of the work has is to develop a model using machine learning to identify maintenance tasks undertaken by contractors that deviate from what would be expected. The identification of these tasks can target inspections which generate cost savings through the prevention of fraud. Several different methods of modelling were trialled to determine one that works with a high degree of confidence.  

Martin Lamb wrapped up the workshop, outlining how organisations could get involved in the BD Pave Project, stating that organisations research or reinvestment projects can be fitted round the project, with the identified work-packages identified for BD Pave, primarily acting as a framework, rather than something set in stone.

For more info about the BDPAVE project, please contact: Programme Lead, Dirk Jansen (, or Programme Manager, Martin Lamb (


Share on Facebook Share on Twitter Share on LinkedIn Share on Google Plus