Following the success of the recent SciRoc Challenge, the team at The Open University have been analysing the wealth of data that were exchanged between the robots taking part in the competition trials and the MK Data Hub over the course of the week. The MK Data Hub is a data infrastructure that supports the acquisition and management of both static and dynamic data streams in order to facilitate the rapid development of smart city systems. The system provided a vital backbone for the SciRoc competition, acting as a central point through which robots both accessed competition data and also communicated their actions and reasoning.
Specifically, communication took place through the MK Data Hub’s custom-built SciRoc API, which includes support for sharing real-time position coordinates, images and status updates. As well as forming the basis of a bespoke web application for the competition judges, this solution also fed into live screens around the competition arena, allowing the public to watch simultaneously both the live action taking place in the arena, as well as its visualisation. As a result it became easier for the audience to better understand the rationale for the events taking place in the arena and for the judges to assess the performance of the robots.
Over the course of the competition, over 200,000 data items were passed from robots into the MK Data Hub and we have now made available a small subset of this data sample in the form of a SciRoc-Replay application. This web application offers the chance to replay trials from the competition and watch them play out in real-time alongside video footage of the robots in action. Nothing beats being part of the audience during the SciRoc competition, but this is the next best thing! Watch it at https://datahub.mksmart.org/demos/sciroc-replay/.
Going beyond SciRoc, this solution provides a general-purpose technology, which can be used in any robotics competition to log all trials, to assist the judges, and to make it easier for the audience to understand the action.