A potent combo

Enough time has passed that an update on our collective efforts to use RFID tags for inventory in nurseries and garden centers seems warranted. Before we delve into our current efforts, we should reflect on how we got to this point.

(Editor’s note: RFID technology is used to track and manage assets and products. A microchip and antenna make up the “tag,” which can be read by an RFID scanning device. The scanner turns the radio waves into digital data. While similar to the more familiar bar code system, RFID offers a major advantage over bar coding in that the scanning device does not need direct line of sight to read the tag. Learn more in this article: bit.ly/NMRFID2021)

Roll back the calendar to October 2017 and the AmericanHort Production Technology conference in Dallas, Texas. Tom Fernandez was speaking in one room about “RFID for the Green Industry,” and James Robbins and Joe Mari Maja were speaking in a concurrent session in another room on using drones in the nursery industry. Part of Fernandez’s talk was reporting on his 2013 HRI grant-funded research at J. Frank Schmidt Nursery and related efforts at Flowers by Bauers and Countryside Garden Center. After our respective talks, the three of us talked about combining efforts to focus on using a drone to collect plant inventory in nurseries. A plan was set in place to submit a grant to HRI. In late 2019, the team was notified the project would be funded and work could start in 2020, but COVID had other plans for us.

Experimenting with tag types

Over the next two years, Maja, although isolated in Blackville, South Carolina, conducted some basic research investigating tag types and orientation on tag counts. Two papers were published from this effort. In 2022, once travel restrictions were lifted on the research team, we started a large experiment at R.A. Dudley Nurseries in Thomson, Georgia. Beginning in February and bi-monthly through December (a total of six trips), the team evaluated tag counts using a reader and antenna(s) mounted on a drone. In February, the team started with four tags attached to four different plant types (4,000 total tags). Maja had constructed a custom RFID reader and a single antenna panel that would be attached to the underside of the drone. Initial results were disappointing, so the team decided to make several changes before the next data collection trip in April. A key to our success in 2022 was bringing in experts from Avery Dennison and William Frick & Co. By the April trip, the tag experts had designed five additional tags (now three tag types: loop-lock, peel-n-stick and stake) and six antenna types, and Maja had modified the platform to test up to three antennas.

Each of these modifications yielded vastly improved tag counts. Over the next few months, the team continued to discover other factors that could improve tag counts. In June, we also started including hand-held counters in our work, which led to other ideas for the future. We consider the hand-held reader as low hanging fruit with the earliest path to adoption. At our wrap-up meeting in December, the team was very optimistic about the direction we were headed, but we lacked funding to continue efforts.

In 2023, we had no external funding, but thanks to the generosity of the vice president of research at Clemson, Maja and Robbins continued the effort with the addition of Van Patiluna. As we did in 2022, research was conducted at R.A. Dudley Nurseries in Georgia. With a totally redesigned RFID reader and antenna, the team evaluated two tag types (loop-lock and stake) and two tag antennas. In addition to those factors, we were also able to evaluate the effect of reader power level and drone flight altitude (distance from the RFID tag). The newly designed payload (reader+battery+antenna) weighs less than 1 pound. Results demonstrated that we can achieve 100% tag counts.

The next frontier

What are our thoughts on the next steps? Clearly, we are closer than ever to a commercially viable solution to improve the inventory process for nurseries. Feedback so far from growers seems to favor a peel-n-stick tag for containers, but we are also investigating a loop-lock tag. We would love to get industry input on their preferences. A key consideration for the team is to produce a system with a user-friendly interface to download plant counts that are linked to their existing production database.

We are also excited about the potential for the same RFID tag used by the grower that will be transferred to the retail customer for their use. Retailers tell us they are equally interested in RFID tags to help improve checkout and reduce shrinkage. The RFID tag is owned by the initial purchaser, so any data encoded onto the tag can remain confidential or be shared along the value chain as the owner desires. A password can be required for reading the data on most tags, which can also remain confidential or be shared. Many tags also have a “kill code” that will permanently disable the tag if sensitive information has been encoded onto the tag. An RFID tag can have a similar code (electronic product code or EPC) to a bar code, making it easily integrated into the retail system. The best way to integrate RFID tags into other operational uses at a nursery is to link the code with the nursery’s database system. Tag data can easily be transferred from the RFID reading system to almost any database system. Transfer of data from the reader to a database can be rapidly accomplished by electronic means so that the information can be used in real time. The database remains separate and confidential so that sharing of the EPC does not transfer any production information to downstream recipients other than item identification as long as no other information has been encoded onto the tag. Even so, the protections mentioned above can be used to ensure confidentiality of data.

Crack the code

Growers and retailers have many options for coding tags. Users can purchase pre-coded tags, or they can configure on-site. For on-site coding, users can use a hand-held reader/writer (make sure to purchase a hand-held with encoding capability) to “blast” encode all tags within the hand-held’s field. When blast encoding, you must be sure that only the labels you want to encode are in the hand-held’s field of influence, or you could overwrite existing RFID labels. Another way around this is to lock RFID tags, an option for many tags, after encoding so that they won’t be overwritten without first unlocking. There are also printer/encoders that are identical to current label thermal printers but with the added ability to encode RFID tags. Make sure the printer/encoder can handle the dimensions of the labels you want to use. These types of printers can print a visible label only, encode the RFID tag only or print a visible label and encode the RFID tag. The label must have an RFID inlay; there is no printer/encoder that will apply an RFID inlay to an existing label. Each label can have different information printed and/or encoded onto it with these printer/encoders. The label writing/encoding software is very similar to those used for existing thermal label printers.

Options for garden centers include RFID portal antennas at exit points and/or handheld readers.

Acknowledgments: 2022: Sincere thanks to Bennett Dudley and R.A. Dudley Nurseries, Inc.; Shane Hutton, DJ Lee, Gary Stegall with Avery Dennison; Brent Howell, formerly with William Frick & Co.; Jim Owen, USDA-Wooster; Jerry Han with Clemson. The 2022 project was partially funded by a grant from the Horticulture Research Institute (HRI) of the AmericanHort Foundation. 2023: Sincere thanks to Tanju Karanfil, vice president of research, Clemson University.

Joe Mari Maja is senior researcher and director for the Center of Applied Artificial Intelligence for Sustainable Agriculture at SC State 1890 Research & Extension at South Carolina State University. Tom Fernandez is a professor in the Department of Horticulture at Michigan State University. James Robbins is a retired professor of commercial ornamentals at the University of Arkansas System Division of Agriculture.