STREAM - Innovations for Sustainable, Smart, and Efficient Automation

STREAM is a VINNOVA-funded challenge-driven innovation, where researchers from SICS, in collaboration with industry, will create new and competitive automation solutions.

The project started in 2012, it is led by researchers from SICS in Västerås and Kista and implemented in collaboration with ABB, Addiva, Blue Institute, Bombardier, Eduro, Mälardalen University, Mälarenergi, Prevas, Trafikverket and Volvo CE. 

STREAM will create new automation solutions through integration and further development of current technologies for automated condition monitoring, diagnosis, planning and optimization in existing systems.

Case studies done in participation with industry is an important part of the project. The results from the case studies will build up the STREAM toolbox. Eventually this toolbox will consist of a broad spectrum of innovative products and services for process automation with direct impact on future production and maintenance management.

The project results are expected to have an impact on the international market which will maintain and strengthen Sweden´s strong global position in automation. 

Markus Bohlin talks about STREAM at SICS Open House 2015:


More information

Number of items: 1.

Olsson, Tomas and Holst, Anders (2015) A Probabilistic Approach to Aggregating Anomalies for Unsupervised Anomaly Detection with Industrial Applications. In: Proceedings of the Twenty-Eighth International Florida Artificial Intelligence Research Society Conference.

This list was generated on Sat Oct 22 22:40:45 2016 CEST.
Project news

17 sept. 2013   EXJOBB OM Smartare sätt att köra tåg

ETT AV DE FÖRSTA DELRESULTATEN I STREAMPROJEKTET HAR PRESENTERATS. Det var Panagiotis Gkortzas som under ett seminarium på Bombardier presenterade det exjobb han gjort vid Mälardalens högskola: Study on optimal train movement for minimum energy consumption.

Panos studie är indelad i tre delar. Den första är ett förslag till modell för beräkning av energiförbrukning. Den andra är en presentation av en metod baserad på dynamisk programmering och Hamilton-Jacobi-Bellman ekvationen. Exjobbets tredje del är en fallstudie som inkluderar presentation av en preliminär algoritm som utvecklats inom examensarbetet.

Bland åhörarna på seminariet fanns representanter för SICS, Mälardalens högskola, Bombardier, Addiva och SJ. Samtliga gratulerade Panos till ett väl utfört arbete och efter presentationen uppstod en livlig diskussion om möjligheterna att använda och vidareutveckla resultatet från studien. Närmast kommer en doktorand vid MDH att ta vid och fortsätta arbetet.

– Panos har gjort en modell för tågsimulering från punkt a till punkt b. Nu ska vi gå vidare och se hur vi kan vidareutveckla modellen, göra den flexiblare och mer detaljerad, samt se om vi kan lägga in vilka sträckor som helst i den, säger Fredrik Wallin på MDH, som är ansvarig för fallstudien i STREAM.

Panagiotis Gkortzas

Panagiotis Gkortzas presenterade en modell för tågsimulering


Suppliers: ABB, Bombardier Transportation, Prevas, Addiva

Academy and institutes: SICS Swedish ICT, Mälardalens Högskola, Blue Institute

Customers: TTI Algeciras, Long Beach Container Terminal, SJ, Green gargo, Outokumpu

Innovation team: Automation Region, Westermo, Länsstyrelsen Västmanlands län, Motion Control, Atlas Copco, Linköpings Universitet, ICA, Volvo Contrucion Equipment

Case studies
STREAM is a cross-industry project for advanced automation. Three case studies are included in the project. The results of the case studies will be part of a toolbox of relatively generic modules that STREAM will produce. The modules of the toolbox will be applicable throughout the automation industry.


The case study will develop decision support to train drivers for energy optimal driving. Given parameters like scheduling, track profile, load or traffic lights along the way it can be energetically advantageous to pull up or slow down at different times. The results of STREAM will be used to produce energy optimal motion profiles for drivers as well as for automatic control systems. This knowledge will also increase the possibility of replacing diesel trains with battery-operated trains on non-electrified sections of track.

Contact: fredrik.wallin [at] (Fredrik Wallin)


The case study is about managing the vast amount of operational messages and errors from cranes in container terminals, in order to react quickly when something is about to go wrong. At the moment, the case study is analysing data from the container cranes. From event data we are hoping to find cranes with an abnormal frequency of events, which may indicate errors. The case study will also collect information on the movements of containers the cranes perform. From these data, one can assess whether some cranes systematically takes more time for the transfers than they should, and thus find cranes with errors that cause production losses.

Contact: Anders Holst


The case study is about forecasting the heating of elements in soaking pits. The approach of the project is to use the data obtained by studying the warming phases and using statistical methods to predict the heating process and estimate the certitude of these forecasts. With a large amount of historical data precision in the forecasts can be obtained. Used in practice, such a tool can be made self-learning by gradually making the parameters in the models more precise.

Contact: Jan Ekman