The ties between RFID and identification applications have long been established. No wonder why: RFID stands for radio-frequency identification and it was developed with the idea to identify assets. As such, the technology has become widely popular for its usage in logistics and supply chain applications, providing advantages over bar codes to uniquely identify items – this allows for track & trace services without the need of line of sight, at distances up to 10m and readings of up to 700 items per second. As such, RFID is a good source of data for IoT just as an identification tool. As an example, items at retails stores are tagged with RFID devices so that companies can take inventories in a fast and reliable fashion and replenish stores to avoid out of stock losses.
Adding sensors to the mix
One of the greatest advantages of RFID – if not the biggest – is the possibility of working battery-free. Obviously, there are active RFID solutions that can communicate over kilometres of distance but being able to track RFID tags at some meters without using batteries at all is very convenient. Now, the IoT is not just about identification. There is a triad to successful IoT systems: data collection, data processing and data delivery. RFID systems are included in the data collection part but... why limit this to identification? Latest advances in RFID technology have resulted in wireless and battery-free devices. RFID tags basically harvest RF energy to power up a chip that replies back with a unique identification number. New developments use that harvested energy to power up external devices such as sensors or actuators – which is a new source of data for the IoT. Temperature sensors, pressure sensors, humidity or soil moisture sensors, voltage or current sensors... you name it. Any of these can be wireless and battery-free and provide an important set of data. Even LEDs can be flashed, mechanical relays switched or displays changed wirelessly and without batteries. As pointed out previously, IoT is also about processing all the data and delivering it correctly. RFID hardware can provide lots of data. However, the true potential of the RFID systems lies in the combination of data collection and data processing, converting data into meaningful and actionable information.

EDNAOL 2016MAY26 TA 01Fig1 Figure 1: Rotor temperature monitoring system.
As an example, RFID is a great technology to collect data of hot spots inside motor and generator rotors. In these motors, winding temperature or permanent magnet temperature are key but, as a rotating device, wiring sensors to the rotor is not a possibility. Battery powered wireless devices are technically feasible but the cost of changing batteries is too high when having to stop the motors for this purpose, sometimes affecting the production of a whole manufacturing line. RFID temperature sensors can be placed on the rotor to collect data from hot spots. These sensors will never require a battery change so the solution works fine. Still, having the data is only part of the equation. This data has no value by itself. Engineers must process this data to optimise motor design. Maintenance staff at a manufacturing line also need to process the data to enhance maintenance and optimise motor life cycle. Engineers will have to set up temperature thresholds to automate corrective actions. They will have to apply data mining techniques, digging into the historical data stored during years of operation, with hot spot monitoring sensors being used in a variety of motors. They will also recommend new sensors to be implemented, improving their designs and maintenance routine.
RFID is a helping hand, the key lies on processing
Once again, battery-free RFID sensors and actuators are a great fit for such use cases. You can use them in a myriad of applications such as rotating parts monitoring, hardly accessible area monitoring – think about soil moisture monitoring in large agricultural farms or structural health monitoring in buildings, tunnels or bridges – or high voltage areas – such as switchgear bus bar temperature monitoring.

EDNAOL 2016MAY26 TA 01Fig2*__Figure 2: __RFID system for data centre infrastructure management (DCIM). *text in italic
However, it is important to note that the key is not data but information. More data only means more processing power required which will become an issue sooner or later. It would be wise from our side to understand which data is really valuable and which is not before throwing it into the IoT/Big Data pool. Having a purpose for each data source we implement – or selecting which data to actually process and which data to discard – will be the challenge of the near future. Discussions on which data to process locally or not are also taking place – for example: temperature sensors in cold chain applications may not transfer all 10 days' temperature measurements but just critical events such as time elapsed above a temperature threshold. With 50 billion connected devices by 2020 as a market estimate, processing needs to be well thought in advance.
About the author
Mikel Choperena is Product Development Manager at Farsens.