The work of the Embedded Systems and Operating Systems group (EOS) of the University of Magdeburg centers on reliable and timely cooperation in large systems composed from networks of interacting smart artefacts. Such systems will receive increased attention because technological advances allow for smart components that provide the ability for self organization, spontaneous interaction, and autonomy of behaviour. We envisage a universe of cooperating things. These artefacts may be physical objects comprising mechanical parts, electronics, networking facilities and software. Thus, they are active components and autonomous, perceive or act on aspects of their physical environment and have the ability to communicate and cooperate spontaneously driven by changes in the environment or in the network universe. These components also may be software only, however, our view on the system model is driven largely by the existence of physically materialized smart objects. Objects may be nodes in a wireless sensor network, ECUs in a car, smart transducers in an autonomous robot, a robot in a team of robots, or navigation and discovery aids in a smart environment. One of the common properties is that usually there are substantial constrains of resources affecting performance, memory space, power consumption and network bandwidth. The main objective of EOS is to develop components and adequate interaction abstractions and communication mechanisms for such systems. The work is focussed on architectures and algorithms which support predictable communication and safe coordination of low power/performance components in a mobile environment. Primary design requirements are related to the autonomy of components, adaptivity to dynamic changes in the environment, and robustness against the failures of nodes and networks including failures of sensors and actuators. We investigate in architectures which allow the seamless integration of components and networks largely varying in computational and communication abilities. Besides developing the system infrastructure, we also designed hardware for smart autonomous components. Examples rang e from smart optical sensors, navigational devices like magnetic field sensors, acceleration sensors and position sensors to environmental surveillance sensors capturing temperature, light intensity, pressure humidity and vibration. Recently, the EOS group works towards mixed reality systems for safe collaboration between humans and robots in a production environment and on safe coordination of mobile air and ground vehicles. Emphasis is on the reliable perception of the environment exploiting a rich set of on-board and remote sensors. This includes the dynamic discovery and use o f external sensors as well as a way of assessing the trustworthiness of such information sources. Checking the sensor information against the expectations derived from a simulated environment model further enforces this assessment. Research is partly carried out in the VIERforEs project funded by the Ministry of Research and Education of Germany (BMBF) and by the KARYON project (Kernel-based ARchitecture for safetY-critical cONtrol, STREP proposal ID: 288195) funded under the FP7-ICT-program of the European Commission.