softgate has successfully completed "ReKoSys", a project in cooperation with the Bavarian State Ministry for Commerce, Infrastructure, Transport and Technology (Bayerisches Staatsministerium für Wirtschaft, Infrastruktur, Verkehr und Technologie). This research project examined the extent to which classification methods for object recognition in camera images (cognitive functions) can be ported to embedded systems (e.g. FPGA Field Programmable Gate Array).
Two of the cooperation partners, namely Erlangen based Fraunhofer Institute for Integrated Circuits IIS and Dallmeier electronic GmbH & Co.KG from Regensburg, concentrated on automated (pre-) analyses of video data from security cameras. The Department of Computer Science 12 (Hardware-Software Co-design) of Erlangen-Nuremberg University and softgate focused on transferring these procedures into automotive network technology. To provide a multitude of network protocols, one single powerful network structure is to be used. A cognitive intelligent function is able to switch between these network protocols. Up to now, each protocol needed its own network infrastructure, significantly affecting weight and costs. Thus, a demonstrator had been developed to visualize the so called CAN+ protocol, an extension of the widespread automotive CAN bus. This extension is able not only to transfer standard CAN messages (e.g. wiper impulse) but also dynamic video data (e.g. rear view camera) via this infrastructure.