English (United Kingdom)Greek

Increasingly sensor nodes are implemented as complex sensor board architectures, driving the emergence of a new breed of monitoring devices. In a typical setting, they are able to host automated computational and data storing operations. The fact that processing power and low latency memory became increasingly available and affordable for many types of small scale designs, lead to significant sensor board advancements in terms of both hardware components and supporting software. Recent examples of cutting-edge wireless sensor modules are powered by a potent 32 bit processor/controller accompanied by several Megabytes of flash memory, while support for external memory can also be available. Corporate R&D utilizing these architectures, has made possible the transition of (industry) process-supporting sensors from simple logic and algorithms printed on bare metal, to complex middleware platforms sitting on top of dedicated sensor operating systems and to interfaces enabling the development of (industry) process oriented application components. These tiny systems are presently capable of serving diverse processing needs with sophisticated software functions and scalable services. These small scale agents are currently populating every supervising, monitoring, tracking, position system and they are called “Smart Sensors”.

Smart sensor behaviour may vary from simple signal amplification to advanced data modeling techniques for condition monitoring. The intelligent attributes of a smart sensor me include one or more of the following functionalities.

  • They include the processing capacity and the proper software routines to process data locally.
  • They can make efficient use of the network infrastructure through complex protocols and distributed communication patterns. Smart sensors are able to implement policies that enhance the network robustness and flexibility, and lessen the burden on centralized nodes.
  • They can support the execution of advanced distributed processes. These may include collective decisions, node task allocation and workflow management for the entire network.
  • They should be able to classify the data according to its criticality, in order to avoid unnecessary data processing during a critical stage of the monitoring item. Smart sensors can evaluate situations and configure sensing frequency enabling better monitoring performance when identifying a critical state.
  • They should be capable of self-diagnosis and self-calibration and be able to periodically prompt on coordinating sensors to collect and process network statistics. Such processing can result in network self customizations that balance topology and upgrade sensing performance. Faulty sensors can be easily identified, while the deployment of new nodes can be estimated based on sensing coverage maps and algorithms.
  • They can be re-programmed and facilitate the network to receive remote software updates. Additional processing techniques can be downloaded to a smart sensor. This feature eliminates the down-time of the network for updates.

Increasing efforts on all-optical sensors technology, power by demanding applications’ requirements have lead to the development of platforms with versatile operations characterized by unique sensitivity, compactness, reliability, electromagnetic immunity and low cost. These attributes managed to promote them to a preferable solution for real-world applications, from mechanical sensing to chemical/biochemical and pharmaceutical industry. This unified photonics platforms exhibit the additional inherent capability of fiber-optic based transmission allowing high-speed interconnection of multiple remote sensors in a single management centre.

Publications - Journals