There are currently over 300,000 connected Scania lorries and buses. These vehicles generate enormous amounts of data; however, in order to provide meaningful results and concrete added value, they must be compiled and processed. This is where the D-ICE project comes in, an arena for data-driven innovation.
Funded by Vinnova since 2017, D-ICE is managed by RISE as a platform for data owners in need of large amounts of calculating capacity. As early as 2016, Scania initiated the Vinnova-funded research project FUMA (Fleet telematics big data analytics for vehicle Usage Modelling and Analysis). The goal of this project was to utilise the large amounts of data created by Scania’s connected lorries and buses to increase the understanding of heavy vehicles. FUMA is a collaboration between Scania and the Fraunhofer-Chalmers Centre for Industrial Mathematics. One prerequisite for researching large amounts of data is sufficient calculating capacity. In its role as a facilitator for collaboration and the increased use of big-data analysis in industry, D-ICE provides and maintains the necessary platform for calculation in the project.
D-ICE provides both a digital platform for storage and calculation at the SICS ICE data centre and a physical platform focused on networking and seminars. Using this external platform has provided Scania with the opportunity to collaborate with its research partners, something that would otherwise have proven difficult.
“For security reasons, external researchers are not permitted to retrieve Scania data; however, a function in the HOPS program used by D-ICE supports multiple users working with the same data,” explains Sara Sylvan, project manager at Scania.
How would you describe the delivery from RISE and the work carried out in the project?
“This is a high-performance collaboration platform with good support. For our part, it has generated know-how regarding big data from developers at RISE. In return, they have been afforded an insight into how we, as the client, actually work. So, there has been a two-way exchange of knowledge,” says Gustav Rånby, a developer at Scania.
“Through the project, we have also been getting to know Mobilaris, a company facing similar challenges to ourselves,” says Sara Sylvan.
Scania’s research project is intended to improve the mapping of HGV usage once the vehicles have rolled off of the production line. It is also hoped to achieve a clearer overall view and decision-making basis to more thoroughly answer questions such as:
- Can the causes of any faults be traced to anything in the driving pattern?
- How do we optimise the flow of vehicles from the production line to the end customer?
- How do our customers use their products, so that we can offer every customer the best possible service?
In the project, the data obtained from vehicles – which includes their position at various points in time – forms the basis for constructing a mathematical model showing where, when and for how long vehicles remain at various locations. Each stop constitutes input for a descriptive statistic that can be used to understand how and why vehicles behave as they do.
The project has also constructed the mathematical model for plotting the vehicle’s movements in a systematic manner, something that has not been done previously.
“With over 300,000 HGVs on the road, the amount of data generated is enormous. Writing programs to handle terabytes of data has its challenges,” says Gustav Rånby.
Have you begun using the structured mathematical model?
“We are discussing its use in a number of areas within Scania; for example, in the subprocess, ‘vehicle ready in factory – deliver to customer’, which has many stages and is difficult to monitor. More accurate mapping of the course of events will improve monitoring,” explains Sara Sylvan.
D-ICE was financed by from until June 2018.
“We have however decided to continue given the good results. Scania will be footing the bill until 2019,” concludes Sara.
The D-ICE project was financed by Vinnova. Project partners are Ericsson, RISE and startup Logical Clocks. Other participants include Luleå University of Technology, data-analysis specialists PreEye and SWECO Society.