Giraldo et al (2019) have examined the potential use of nanotechnology in plants to improve crop yields and increase the stability and resilience of the world’s food supply. The underlying concept involves the application of embedded or “wearable” smart plant sensors that could provide feedback on environmental changes and pathogen-related stresses in real time. This could improve the efficiency and accuracy of corrective measures – for example, whether or not pesticides or fungicides should be applied, when, where, and what type.
The goal is to make resource allocation (water, fertilizer, pesticides, herbicides, fungicides) more efficient, reducing costs, increasing yield, and reducing potential environmental impacts.
One proposed design would be to “embed” nano-sensors inside of plant tissue. Generally, this would involve coating carbon nanotube based sensors with DNA designed to allow the sensor to coexist inside of the cell of a plant without integrating it into the plant’s DNA directly. Some of the advantages of this design is that it would be possible to report on detailed plant physiological changes – sugar and water levels, pH, salinity, among others.
Data collection from embedded sensors might be accomplished by using autonomous vehicles to drive- or fly-by to retrieve data. Another method might involve the use of strictly wireless base-station technology – that is, the plant sensors would be connected to an integrated network, much like the computers and printers in an office LAN.
The second method would consist of installing “wearable” sensors on the surfaces of plants. A great deal of work has already done with flexible wearable sensors, particularly on clothing and other manufactured goods, which could lead to faster implementation times. These sensors might be valuable for measuring the emission of gas molecules from plant leaves, or for detecting external environmental conditions in proximity to the plant.
Data collection might best be accomplished through the use of autonomous detection vehicles driving through the rows, or via wireless base-station technology (which might be more efficient and practical than embedded tech, given these are larger, external sensors).
One obvious issue is how would all of these devices be powered? One obvious solution might be RF. Devices would have to have some ability to collect and store data, but would not be in a constant state of autonomous transmission. Instead, powerful RF signals would be beamed to activate the sensors and collect data. This is something that has been done for years for loss prevention purposes in retail stores, or to tailor customer experiences while shopping.
Regardless of the specific methods used to activate and collect data from these sensors, there remains the question of how bombarding fields with potentially powerful EMF signals might impact either the health of the plants themselves, farm workers, or local residents. What sort of interference might be created that could interfere with other air broadcast transmissions like cellphones, TV or radio?
Another potential issue (at least in centralized data collection situations) would be bandwidth. Streaming real-time data on multiple plant physiological parameters from thousands of plants would represent a substantial data stream. One could instead revert to autonomous vehicle collection, but centralized collection has its appeal, in that autonomous vehicle capital and operational costs could be completely avoided. Perhaps the only way to make a centralized wireless system work would be through the implementation of 5G, a technology that has its own security and potential health-related issues.
Finally there is the matter of “big data” – collecting dozens of parameters from each plant, potentially in real-time, from thousands of plants – that’s a LOT of data to process and analyze. There are some vague proposals for “machine learning” forms of analysis, which at present, I don’t feel is even close to being ready for “prime time”.
Another idea would be to examine the data in aggregate to get general ideas about overall field conditions and plant health. One might also select a statistical method for distributing these sensors to gauge a “representative” notion of plant growing conditions – that is, we don’t implant or affix sensors to every plant, we choose a representative number distributed across the fields, and use this more limited data set to make decisions.
While there are many technical issues to be addressed, I still believe this is promising technology that will eventually see its way into the agricultural mainstream.