The research partners will focus on developing a reference platform for vision-based system designers that defines a set of guidelines for the selection of relevant combinations of computing and communication resources (minimizing both energy resources, development costs and time-to-market).
The guidelines for the reference platform will define what a piece of hardware or software must look like in order to be Tulipp-compliant. Then the project partners will develop an instance of the Tulipp reference platform, comprising a scalable low-power board designed to meet typical embedded systems requirements of size, weight and power (SWaP), a low-power operating system and image processing libraries, and an energy-aware tool chain.
Three use case demonstrators will be developed as proof-of-concept and validation of the reference platform, covering different industrial domains with emerging complex image processing requirements. These will include a medical imaging surgical X-ray system designed to significantly reduce radiation doses by 75%; a smart automotive embedded vision system for advanced driver assistance (ADAS) that, in addition to the low-level image processing, intelligently interprets what is on the images to deliver safer driving experiences; and an embedded image processing system to create smart drones and Unmanned Aerial Vehicles (UAVs) for the intelligent search and rescue of survivors at disaster incidents.
By the end of the project in 2018, Tulipp expects its work to extend the peak performance per Watt of image processing applications by 4x and average performance per Watt by 10x. Beyond the official completion of the TULIPP project, it is expected that this will be extended to 100x and 200x by 2023.
Tulipp will work closely with various standards organisations to propose the formal adoption, on an industry-wide basis, of new standards derived from its reference platform. The project members are also seeking to establish an Advisory Board of vision-based systems stakeholders to review the work of the project on a progressive basis and help extend the reach of image processing applications into