How does RFID take stock of the goods? Goods identification and recording
Many people think that RFID counts are "cleaning." It's not that simple.
Swing the reader onto the shelf, the backstage automatically comes out of the inventory -- a first impression of RFID counts. The RFID does read in bulk and far from, but it can read, depending on the range of prefixes: labels, frequency, labelling and reading. Skipping these preparations is hand-to-hand, often with incomplete reading, misreading or re-cleaning. The gap in the efficiency of the count is not always at the device itself, but at the level of detail of pre-selection and on-site testing.
We'll get to know the RFID readings first.
The RFID system consists of three parts: labels (on the goods), readers (handheld or fixed channels), antennas (to send and receive radio frequency signals). On a stocktaking, the antenna continues to fire an RF signal, the tag receives a "awakening" signal, re-reflects stored coded information, and the reader takes over and decodes it into the background database. There are two key points to understand this process:
- Frequency determines the distance to read and apply the scene. Low frequency (LF, 125-134 kHz) has a short read range and strong resistance to interference, making it suitable for metal components or animal tags. High frequency (HF, 13.56 MHz) has a read range of 10 cm to 1 m and is often used for libraries and item-level tagging. Ultra-high frequency (UHF, 860-960 MHz) has a read range of 1-10 m and is widely used for warehouse stocktaking. In China, the commonly used UHF band is 920-925 MHz.
- Batch reading of impact-reliance algorithms. When multiple labels appear in the reader range, one by one need to be identified in order to avoid reading or conflict. ISO 18000-63 (also known as EPC Class 1 Gen 2, or EPC, Electronic Production Code electronic product code) is the most commonly used protocol standard in the current storage logistics field and supports reading hundreds of tags per second.
Standard operating step for RFID counts
The physical inventory is normally carried out in the following order:
1. Label selection and encoding
Choose tags to suit the material. Adhesive flexible tags are common on paper packaging. Metal goods require on-metal tags with a wave-absorbing layer. For liquids, use special tags made of high-dielectric-constant material, a measure of its ability to store charge. During tag issuance, write an EPC to each tag and link it to a SKU (stock keeping unit) in the back-end WMS (warehouse management system) or ERP (enterprise resource planning) system. This gives each item its own ID number.
2. Label pasting
The label is usually requested to be in the direction of the passage or reading, and to be flat-faced without a mask. The tag on the metal shelf leaves the surface of the metal at least 5 mm, so that the signal is not blocked. When you have it, a check is recommended to confirm that the label is read properly by the handheld.
3. Live reading
There are three main types of inventory:
- Handheld machine inspection: The operator holds the terminal at a speed synchronised along the shelf channel, and the antenna automatically identifies labels within 3-5 metres of the perimeter, suitable for over SKU-distributed warehouses.
- Fixed Channel ReadInstallation of fixed reading and writing machines and antennas at the warehouse entrance or forklift corridor, where goods are automatically recorded as they pass through and suitable for frequent access to the garage.
- RFID TunnelReading channels in a packaged or packaged form, 360 degrees of no-mortem scanning at the time of passage, often used in clothing, electrician sorting centres.
4. Data upload and reconciliation
Handheld read data is uploaded to the backstage in real time via Wi-Fi or 4G, and the system compares the EPC code actually read with the book inventory, generating a discrepancy report. The missing, reread, misreading goods enter an abnormal process, which is manually reset.
5. Abnormal treatment
Common anomalies include: damage to the label is not readable, changes in the location of the goods are not in the reading range, and two SKUs are coded over time. The process involves re-routing the location with a backup handheld, adjusting the label or position, and eventually uploading the data.
The easiest pits to step on.
The RFID inventory is not complicated, but the site details determine most of the success rates. The following problems are very frequent in the actual project:
- Metal and liquid interference. Metals diminish the signal of a frequency, and liquids absorb signals, leading to lower reading rates. The solution is to select metal labels or increase their distance from the surface of the metal.
- Tabs mislabelledThe label is pressed to the bottom of the goods, covered by tape or attached to plastic foams, which reduces the distance to read and even fails to read.
- Reader power and antenna are not right.. The power is not enough to read the labels of the shelf next door.
- No batch compatibility test performedThere are differences in the compatibility of labels and readers between different brands, especially when domestic production and imported equipment are mixed.
Preparation of several pre-implementation
If it is intended to introduce RFID inventories in the warehouse, it is proposed that before the official roll-out the following work be done:
- Inventory scene assessment: a single-grade inventory of the whole tray, box or single class, whether the goods are metal, plastics or liquid packaging, and the length of the shelf and the width of the warehouse.
- Small batch pilotSelection of a class, a storage area for piloting, running through the signposts, reading, uploading and checking the entire process, statistically successful reading (over 99 per cent is generally required) and time-consuming single-take count.
- Device Selection: Determination of label specifications, type of reader and antenna polarization based on pilot results (roof polarization is not sensitive to direction and suitable for handheld scenes; linear polarization reads far away but is highly oriented).
- System docking: RFID intermediate (data-processing layers between readers and business systems, for data filtering, heavy, format conversion) interfaces are to match existing WMS or ERPs, avoiding "reading but not being found in the system".
Shanghai-based Cheeky Way Technologies, when doing RFID inventory projects for clients, usually spend a day or two on the scene to see how the goods enter and leave the warehouse, how the shelf is arranged, how the operator moves around in time, and then backs up the tag program and read the process. The benefits of this are that the program fits into reality, the efficiency of the count increases can be stabilized several times more, and the error rate can be significantly reduced.
Summary
The core of the RFID inventory is "Preselectional preselection, labelling, reading-matching." Skipping any link directly affects the accuracy of the count. Operationally following the "Selecting code -- standard label -- field read-checking -- data checking" process, with special treatment and field touchup for metal/liquid environment, generally achieves a success rate of over 99 per cent.
If it is intended to introduce RFID inventories in the warehouse, it is recommended that the two indicators of reading success and time-consuming single-discretion counts be piloted from one storage area and one class before deciding whether to extend them to the whole warehouse.

Ms. Lei
Mr. Pang