Low-cost, pervasive antenna-based sensing
Motivation
Data capture, from sensors and edge devices, is the first step in building an IoT application. Not all applications are equal. Some may require very precise, on-demand sensor data where hardware cost is not an issue. Other applications, particularly some in fast moving consumer goods, infrastructure monitoring and agriculture, require large-scale, infrequent monitoring of objects where the data needs are simplistic, but hardware cost is often the dominant concern.
In agriculture, for example, soil moisture sensors may come under this category. It may be sufficient to know whether the soil is moist 1-2 times a day, however this needs to be done at good spatial resolution and at affordable cost. As another example, it may be necessary to scan hundreds of items at periodic locations in the supply chain (such as when items enter or leave a warehouse) and be able to assess if they meet a quality metric (such as whether they were ever exposed to temperatures higher than 40C). However, the cost per sensor must be well under $1 per sensor.
The antenna-based sensing concept developed at our labs seeks to address the latter type of application and builds upon well established, low-cost, wireless identification technologies such as UHF RFID.
Beyond the ID in chipped and chipless RFID
UHF RFID technology has improved significantly in the past 20 years. Tags have seen an average read range increase from 3-6 m, tag costs have reduced to 3-5 cents in bulk, readers have become more sensitive and more diverse reader antennas such as multilinear and phased array options have become commercially available. The Gen 2 protocol has also seen evolutions, such as the Gen 2 v.2 update, that presents solutions to emerging concerns like security and privacy. In the past 10 years, RFID has also expanded into multiple sectors such as agriculture, civil infrastructure, healthcare, oil and gas and defense. All in all, RFID technology is well positioned for low-cost, wireless identification and digitization of physical assets. There is, therefore, good potential to expand this technology beyond identification – for pervasive sensing.
Fig. 1 shows the concept of antenna-based sensing and a reader interrogating a tag. The reader decodes the RFID tag’s EPC ID and most applications of the technology stop at this step. The reader is also capable of gleaning power level (RSSI) and phase information as a function of frequency from the tag’s response; however, this information is typically ignored. The antenna-based sensing concept encodes sensing information in these signal parameters. For example, what if the crossing of a threshold temperature manifested itself as a controlled change in the power level of the tag, or an abrupt change in the range of frequencies at which the tag responded to the reader? In addition to ID, one could obtain basic information about a parameter of interest (such as temperature, cracks, light intensity, moisture etc) near the tag with no change to the data capture hardware or wireless protocol.
Figure 1: The antenna-based sensing concept: Changes in a parameter of interest are related to the amplitude, frequency or phase response of a tag.
The core research challenge is in achieving reliable signal transduction – i.e., converting change in the modality of interest to a change in the signal characteristics of the antenna.
Example of using tag power change for sensing
Case example: Soil moisture sensing
Fig. 2(A) shows an antenna-based soil moisture sensor. The moisture level in the soil near the monopole probe changes its characteristic impedance and this manifests as a controlled change in power level. Fig. 2(B) illustrates this. When the soil is saturated the power level is very low. As the moisture dries out and dissipates from the soil, the power level gradually improves and approaches that of dry soil.
Figure 2(A): an RFID antenna-based soil moisture sensor
Figure 2(B): Tag response power as a proxy for soil moisture content. Dry soil has a strong signal response of -30 dBm, while saturated soil registers a -25 dB drop. As the soil dries out, the power levels will improve.
Example using carrier frequency for sensing: Fluid level sensing
Fig 3(A) shows a fluid level sensor attached to a drinking glass. When the glass is empty the tag is broadband and is seen on all frequency channels in the 902-928 MHz band. When the glass is full as in Fig 3(B), the tag is narrow band and is only seen on a subset of channels. By looking at the frequency response, one can infer the fill status.
Figure 3(A): Frequency response of a sensor deployed on an empty glass – the tag responds in all frequencies in the 902-928 MHz band
Figure 3(B): Frequency response of a sensor deployed on a full glass – the tag responds only in the 920-928 MHz band
The following are some of the of antenna-based sensing solutions designed at our lab:
| Industry segment | Application |
| Healthcare | Vaccine vial crack detection, anemia sensing, incontinence monitoring |
| Agriculture | Soil moisture and soil salinity sensing |
| Infrastructure condition monitoring | Crack detection, expansion gap monitoring |
| Cold chain | Temperature alarm, time-temperature indicator |
| Retail | Weight and light intensity sensing |
Contact
For more information contact Dr. Rahul Bhattacharyya
Figure 4(A): Anemia sensing using paper-based lateral flow assays
Figure 4(B): Crack detection by developing an RF “smart skin”
Figure 4(C): Temperature alarm sensing using shape memory polymers