Types of sensors: how machines see, detect and communicate smartly

Over the last 30 years, the Industry 3.0 has been based on a well-defined automation architecture consisting of five layers. Field devices such as sensors and input/output devices at the lowest layer send data to logic controllers through analogue signals. SCADA systems perform remote control tasks and Manufacturing Execution Systems (MED) allow users to carry out complex tasks such as production scheduling. Finally, at the highest layer, ERP systems allow to generate management reports and share manufacturing data, such as the state of the order, with other systems such as the Customer Relationship Management (CMR). 

In Industry 4.0, these layers become diffuse due to Cyber-Physical Systems (CPS), the direct evolution of embedded systems, which offer the possibility of decentralized data processing, Edge Computing or Cloud Computing.

Classification by types of intelligence

• Analogue sensors: information where information is obtained from sensors and read in situ by an operator. Data can be registered manually or not (Industry 3.0).
• Centralized computing: the sensor can register the information in a digital format, although it is not restricted to just one standard (different formats). Information is collected vertically and placed in a centralized management system based on a sensor considering the SCADA system, from SCADA to MES and from MES to ERP (being possible to skip intermediate steps).
• Cloud computing: In a way analogous to centralized computing for the difference that instead of deep and processed in the systems physically located in the facilities of these servers sent to remote servers. This modality is already used in the business models more typical of industry 4.0 such as payment for use or payment for service (SaaS, PaaS, iPaas among other forms), instead of paying for annual software licenses or the purchase of servers. Allowing greater flexibility when making changes to the systems and costs more adjusted to the consumed
• Fog Computing: Donde los datos se recopilan, procesan y almacenan en la red a través de una pasarela IoT o nodo de niebla. Esa información se transmite a esta pasarela desde varias fuentes de la red donde se procesa y los datos de mayor valor se transmiten de vuelta a aquellos dispositivos que se pueden aprovechar de ello. El Fog Computing tiene la ventaja de permitir que un dispositivo de procesamiento único y potente procese los datos recibidos desde múltiples puntos finales y envíe la información exactamente donde se necesita. Además, ofrece menor latencia -menos tiempo en el envío de datos- que una solución de Cloud Computing, donde los datos se enviarían y recibirían desde la nube.
• Edge-Computing (computación in situ): Son los mismos sensores los que disponen de la inteligencia suficiente como para procesar los datos, evitando la necesidad de transmitir verticalmente la información, trabajando el sistema como células independientes pero conectadas entre ellas.  Haciendo el sistema menos dependiente y más robusto frente a anomalías.

Classification by types of communication

There are a lot of models of communication between systems, here we will limit ourselves to mention the most characteristic of each category.

• Wired

  • Serial port: Basic communication, up to 128 devices interconnected and short effective distance, normally 9600 bauds, it is a deterministic communication (it can work in real time). RS-232 is vulnerable to interferences, meanwhile RS-428 is immune to them (twisted pair).
  • Ethernet/IP: Widely spread. More effective distance, more devices connected, and it is not deterministic (difficult to work in real time). 
  • EtherCat: Clearly outclasses Ethernet/IP due to its low latency, high speed, high number of devices connected and possibility of working in real time.

• Wireless​

  • Mesh network: Each node is connected to all the other nodes. With technologies of low switching frequency and low energy cost such as ZigBee or Z-wave, this technology is used in microsensors. With an effective distance of less than 70m, low bitrates (up to 250 KB/s), they have an autonomy of up to two years. Bluetooth can be also included here, but it has a higher consumption.

  • Cellular network: this network was originally created for mobile phones. Its frequency bands are higher, what allows a better bitrate, although it involves a higher energy consumption. They allow getting rid of operating ranges, since they are quite spread in urban areas. Among them, we can find the famous LTE or 4/5G and the ones intended to be used in the industrial environment such as Sigfox or NB-IoT.