How does RFID Asset Management make sense layers and business synergies?
A lot of people think it's easy to manage RFID.
Many people think that RFID asset management is a two-step "labelling + reading" approach. Indeed, the hardware selection and placement of sensory layers (labels, readers, antennas) determine data quality, and business synergies determine whether or not these data are able to re-feasible. The two tiers are not designed -- not for sense layer quality, not for the next report, not for the data; not for the work team, only for the labels to be buried in the database.
What the hell is "sensorship" in the senses?
The sensory layer is not just two pieces of hardware, a reader and a label. It contains at least four links, each of which affects what eventually reads.
Selection of frequencies and protocols. The UHF (UHF) frequency bands in the country are concentrated in windows 840-845 MHz and 920-925 MHz, which is commonly known as the EPC C1G2 (also known as ISO 18000-63) protocol, with a reading distance usually between 1 and 10 metres, suitable for a mass inventory. HF (HF, 13.56 MHz) has short reading distances of not more than 1 metre, and allows for recognition, and bulk asset management is inadequate. The selection frequency is essentially a group of parameters that "wanted to read as far as, how fast, how many can read at a time."
The label is not as cheap as possible.The lowest cost of paper-free glue labels, but attached to metal surfaces or liquid packagings, reading distance shrinks and can not even read at all. When you encounter a metal asset (e.g. a cabinet, a mold), you need to isolate it with suction; when you encounter a plastics turnbox, you have to see the polarization and the angle of the label. The most common problem we encounter in the project is that we have purchased a set of generic labels that want to "eat" all assets, and the tags on the metal shelf are completely destroyed.
Readers and antennas need to "see" labels.. Readers themselves are signal processing units, and the real decision is on antennas. A four-caption fixed reader can connect to four antennas with each of their own radiation orientations. It is common in the warehouse aisle to hold the antennas down on both sides of the channel, covering tags on the tray when the forklift passes; if it is a dispersed asset, the antenna will be directed towards the shelf, and the distance between the shelves cannot be too close, otherwise the signal decay will eat up to half a reading distance. The door-band discs need two sets of antennae cross-tripping to avoid one-way leakage.
Collision and read-out rates. RFID allows multiple labels to return the signal at the same time, relying on a collision-proof algorithm (which gives the labels line up and avoids interference). But it is theoretically possible to read 100 labels, which actually are in the scene because of their cover, angles, multi-range reflection (the signal is offset or reinforced by a wall or metal reflector), and over 85% full reading. The remaining part that cannot be read is usually supplemented by "reread + data cleansing" -- that's why we go down the data link.
The link between the sensory level and the business system is not as simple as "the interface."
The tag data is placed in a reader and not directly thrown to an ERP or WMS (warehouse management) system, with at least one intermediate missing.
Data filtering to middle. The same label may be read several times in the same second because of antenna switching, and the time stamp is not consistent. The middle part has to do a few things: go heavy, merge, judge where the label is most likely to be, and record signal strength (RSSI) support positioning. These things can also be done locally by reading and writing, but more firmly on edges or servers, so that subsequent adjustments are made.
The event should be sealed in the business languageThe bottom device only reports "label XXX" when it's read by antenna Y at T-time, signal strength Z, and business systems want "assets XXX enter/leave A area at T-time," which is not the same thing. It's common to add an event resolutionr to front service, translate "reading/not read" into "entering/exit/inventory counting/unusually lost" standard events, and WMS and ERP are only connected to this level.
Separate in real time and volume. A second-level response is required for a restricted business (assets are immediately alerted), and this type of walk-through is sent to the operating system. Inventory count is a time-bound task, which is then pushed to ERP after reading a round of data stabilization. It cannot be done in one mode, otherwise the system either gets flooded by information or it can't work.
The business co-design determines if the system is really working.
Hardware completion is only the beginning, and business synergies are the key to running.
Business events match label actionsSo, for example, one of the equipment moved back to the warehouse from the maintenance workshop, and the business process was "maintenance completed + inventory," and could the RFID system automatically trigger both when the device passed through the workshop? The key is whether the tag ID code (EPC code) on the device or the ERP's main data are a single-to-one map. Today, using asset numbers, tomorrow with serial numbers, then the automation will be largely by hand.
The count went from "time-on-time" to "state-on-state."The traditional inventory is a yearly centralized count, and the real advantage of RFID is to keep an ongoing count. In our project for a warehouse factory, the antenna above the storage position is open, the system is in real time compared with the assets on the disk and the actual assets on the bank, and the difference over the threshold automatically generates review of the one-to-one tube. The inventory becomes a back-office permanent presence, and the pressure on the front yard is down.
We need to make an agreement on the anomaly.. Missing, unreadable, or obstructed tags are all common when a system first goes live. Agree in advance whether each exception should generate a work order or be grouped into a daily report, and who will verify the exceptions. RFID data is only reliable once those issues are resolved. We have seen projects still reconciling old discrepancies three months after launch, usually because this process was not in place.
From planning to landing, a few indiscretionary nodes.
First, we'll do a small-scale pilot.Do not put thousands of labels on the top. Pick 50 to 100 typical assets, cover 2 or 3 typical scenarios, such as metal shelves, plastics and outdoors, to see if the readout rate can be stabilized at more than 95%.
Field survey of the provinceAn antenna position is influenced by shelf, metal door, elevator well, and looks at the right place on the drawings, which may not actually be read. We usually walk around the warehouse with a cart + hand-reader, and we measure the actual reading distance of each antenna installation point, and then we come back to re-engineer the program.
The sticker job to write SOPSo, if you want to get a good look at the label, who will put it, how you'll check it. The same asset is posted on the top left corner today and on the bottom right corner tomorrow, the count will slip.
KPI doesn't just read "How much have you read."And this is the number of indicators that Shanghai-based Chettech has been most concerned with at project acceptance -- deciding whether the system can really access day-to-day management processes, rather than leaving a collection of useless labels and incomplete reading data.
Technology is only the starting point for RFID asset management. The real test is how well the sensing layer works with business operations. Tags must be applied consistently, antennas positioned correctly, data filtered according to clear rules, and business events defined properly. Bringing these pieces together makes the system useful for asset management instead of producing a pile of unreviewed alerts and records that do not match.

Ms. Lei
Mr. Pang