Abstract
Mobility-as-a-Service (MaaS) describes a class of applications where traditional real-time control systems are enhanced by backbone services accessed via the mobile Internet. In order to implement MaaS, new architectures for multi-stage real-time systems with several layers of control loops have to be implemented. Using approaches such as analytic redundancy, hard real-time control loops are extended with software-defined sensors that deliver data with soft real-time semantics. We describe a real-time control experiment that has been implemented in our Distributed Control Lab with four stages -- an extended digital Carrera race track (D132), custom built sensor/actuator boards, a control PC, and an outer control loop established via web services -- and present a timing analysis. Our architecture allows for decoupling of hard real-time processing on embedded control units and soft real-time data acquisition on the outer layers.