RTPS middleware for real-time distributed industrial vision systems

Designing and constructing real-time distributed industrial vision systems (RT-DIVS) from scratch is very complicated task. RT-DIVS has conflicting requirements such as reasonable development cost, ease of use, reusable code and high performance. The success key in building such systems is to recognize the need for middleware software. Middleware plays a major role in developing distributed systems efficiently. Real-time publish-subscribe (RTPS) model is one of the latest developments in real-time middleware technologies. Network data distribution service (NDDS) is RTPS middleware developed by real-time innovation (RTI). NDDS is widely used in real-time distributed and embedded systems for mission critical applications. The research work presented in this paper discusses the employment of NDDS for RT-DIVS and the advantages of NDDS´s quality of service (QoS) policies in covering the requirements of RT-DIVS. An experimental test set-up is used to verify the NDDS´s performance for RT-DIVS. Tests results show that RTPS middleware (and NDDS specifically) is suitable for soft and firm timelines requirements for distributed industrial vision systems.