What about the RFID asset software? Functional coverage and hardware match
Last year we came to a factory that did auto parts, and they went into the RFID asset inventory system, and they found a strange thing: most of the pallets in the warehouse were cleaned up, but there were always several lined pallet labels that couldn't read. Two weeks of tattering, it was discovered that the reading device antenna and tag were not exactly right. The only way to buy was to look at software, ignore hardware fit, and then get exposed when it was deployed.
The software function is important in the selection of RFID asset software, but hardware suitability is key to determining whether or not the system is on the ground.
The functional coverage is not "the better" but to target your actual process.
Many teams choose software with a list of functionality, and then compare it by item. But the function covers this matter not by whose house is more functional but by whether or not you can cover your core process.
A standard RFID asset management system covers at least a few links: asset entry, tagging, routine inventory, location queries, asset transfer, and end-of-life storage. The functional modules around these links consist broadly of:
- Initialization and coding of labels: RFID tags and indexing of the EPC (Electronic Production Code, understood as the only "electronic ID" number for each asset)
- Batch count: Handheld terminals or fixed-readers scan in the area, automatically aggregate label data
- Positioning and trajectory: Some systems support the regional positioning of assets through multi- antenna deployments, recording asset movement paths
- Data interface: interface with ERP, finance system, MES to synchronize asset data
- Early warning and statement: automatic alarms for assets not captured, inventory not counted over time, label damage etc.
In our experience, smaller and midsize warehouses should focus on bulk counts and data integration. Large groups with distributed assets also need multilevel permissions and cross-region transfer workflows. If you lend assets or return them to the manufacturer for repair, a dedicated module is more useful than ten unrelated features.
The hardware fit is a lot more complicated than I thought.
RFID hardware is not as standardized as bar codes. Market labels and readers are branded in a variety of ways, with different frequencies, protocols, antenna polarization, and it can easily be a problem to match them.
We need to align the frequency first.
Common are UHF UHF (860-960 MHz, reading distance, suitable for warehouses and logistics), HF (13.56 MHz, reading distance short, appropriate for door-barreling and close-range identification), low frequency LF (about 125kHz, strong penetrating power, suitable for metal environment). The different frequencies cannot be interoperable, the first step in choosing a type is to confirm the frequency range you need.
The protocol must be compatible.
The RFID labels have different standards such as ISO 18000-6C, EPC Class1 Gen2 (which is the most used protocol standard in the UHF field at present). If a reader only supports an old agreement, the new label cannot be read, and vice versa.
The labels match the material of the object.
Metal surfaces, water, and other liquids interfere with RFID signals. A customer that stores chemical raw materials told us ordinary tags are nearly impossible to read on metal drums. They need purpose-built on-metal tags, which have a different internal structure and cost 30%–50% more than standard tags.
We need to get the interface and software through.
Readers typically provide SDK (software development toolbox, equivalent to the access interface that a reader has opened to the software) or intermediate, and the API of the software is able to successfully connect, deciding whether data can be returned in real time. We have encountered a project where a brand-name reading writer file is incomplete, and it took twice as long to develop the interface than expected.
The field test is not a chance to save.
Software features are more complete, hardware parameters are much better written and do not count without field testing.
Several key indicators to be tested should focus on:
- Read accuracy: how much can be read in bulk, 100 labels, and more than 99 per cent is generally required by industry
- Distance of reading: Effective reading distance in the physical environment, with a significant reduction in metal and liquid environments
- Compositing capability: how many labels can be identified in a fixed-reader within one second, and the warehouse peak is enough
- Leakage and string reading: the omission is read to be unread, and the string reads " A" misreading the label as B, both indicators directly impacting on inventory accuracy
Run a small pilot in a real operating environment. Select 5-10 shelving units and several dozen tags, and test continuously for 3-5 days. Include normal operations, irregular stacking, and people moving through the area to obtain convincing data. When designing solutions for customers, Shanghai Zejie Technology typically arranges an on-site proof of concept (POC), a small trial to verify performance, and measures these indicators before deciding how to proceed.
Once deployed, the real test is long-term stability.
The system is on line only at the beginning.
- Ageing or defunct labels, especially outdoors or high temperature environments
- The position of the reader antenna is blocked, and changes in the way the storeroom is modified or stacked are affected.
- Compatibility of software version upgrades with existing hardware
- The system slows down when the data is up.
These problems will not be fully exposed at the selection stage, but they can be avoided in a number of ways: selecting suppliers with continuous capacity; confirming software to support subsequent hardware expansions, such as a new reader for high frequency segments; and databases and deployment structures that sustain business growth. In actual terms, we find that these border conditions are clearly stated in advance, and that the cost of post-shipment maintenance is reduced by a large amount.
The selection thing, back to the three core issues.
To close the above, the RFID asset software selection is ultimately answerable to three questions.
First, what are your assets, where they are placed and what processes need to be covered? Functional modules are more important than functional quantities for these actual needs.
Second, can you choose a reader, label, protocol and interface? Hardware fit is not a single parameter match, but a system-wide synergy.
Third, do suppliers have real project experience and ongoing service capacity? This is often more important than the product itself, as the RFID project is largely at site.
Two things are suggested before you choose the type: one is to draw your own current asset management process, and then point out a pain; the other is to list hardware that already has hardware or budget to cover. Send both materials to suppliers, so they can plan on your own facts, instead of using a generic template. The supplier that does this, the selection works half way.

Ms. Lei
Mr. Pang