Measuring energy needs with prototypes that monitor trains in real time with a local wireless sensor network. Photograph: IMMS.
Measuring energy needs with prototypes that monitor trains in real time with a local wireless sensor network. Photograph: IMMS.

fast realtime

Low-latency communication in radio sensor networks and investigations on real-time behaviour of communication stacks

Smart transport to be monitored in real time

There are nine project partners working on ”fast realtime“, tackling the basic principles and key technologies that will facilitate distributed sensor-actuator systems that work in real-time applications. This work will be the key to new developments, supporting, for instance, industrial automation and smart public transport. The nine partners are following a variety of routes to the real-time systems and are putting them into three demonstrator scenarios as examples of the variety possible.

One of these scenarios is the monitoring of the trucks on a goods train. For this setting, IMMS is developing a wireless sensor network that enables data to be received, processed and transmitted in real-time. In parallel, the Institute is at work on a contribution to the methodology in the form of general design guidelines for a new type of system design. Further, IMMS is developing a technological component which will minimise latency, as delay in data transmission is called, so that any systems involved can fulfil tighter time restrictions. The approach being followed is optimising communication stacks in Linux real-time systems.

Project website on fast realtime (in German)

  • Communication stacks in Linux real-time systems

    Especially in industry, there is currently a trend towards the use of Linux as the OS for systems working in real-time. In this field, data is usually output or read in a protocol stack, proceeding through various layers. For the TCP/IP Internet protocol, these are application, transport, Internet access and network access.

    The buffering involved means that latency arises; latency is also caused by the ever-changing memory demands and possibly by the fact that several data streams are being processed in parallel. Such latency can impair the real-time capacity of the entire system.

    IMMS is conducting research into ways of measuring the latency, characterising it and then eliminating any that is avoidable.

  • Monitoring goods trains in real time

    The use of sensors, location methods and status monitoring will, it is hoped, set a series of innovations in railway trucking in motion, for instance by enabling trains to find out for themselves the data that will give them clearance for their next route section. Currently, the information has to be gathered from the infrastructure along the line, with all that this means in terms of maintenance.

    The networked systems integrated into the train will establish positions and journey times, check the train has all its trucks, permit automatic coupling of trucks and monitor the load and the wheels. The systems used for these purposes at present radio the data for each individual truck to an external central controller via mobile communications.

    The ”fast realtime“ approach being investigated goes beyond this. The individual trucks are to be connected together under a monitoring application by means of a local wireless sensor network that requires no mobile communications.

    It is the role of IMMS to create the prototype of the train’s smart sensor network, which will then ceaselessly exchange data (bidirectionally, in a robust fashion, with low latency) between each truck and its neighbour the length of the train. The network will monitor the trucks and check all are present in the correct order.

    For upstream use on the railway network, the train sensor network will be linked by a single mobile radio link to a location-related cloud service that has been optimised for the relevant signal transmission time, with the purpose of checking any changes and events also externally. This service can evaluate that information, if necessary triggering automatic measures.

  • Funding

    The fast-realtime project is supported by the Forschungszentrum Jülich GmbH (PtJ) with funds from the German Federal Ministry of Education and Research (BMBF) in the „Twenty20 – Partnership for Innovation“ programme under the reference 03ZZ0504J.

Duration

2016 – 2018

Reference

03ZZ0504J