How does a rfid-based storage management system work, label modelling and an inventory radius
RFID Knowledge

How does a rfid-based storage management system work, label modelling and an inventory radius

PublishedShanghai Zejie Smart Technology Co., Ltd.

The RFID-based storage management system addresses two different issues: which one of the objects is read, and what the reading proves. The former relies on the data model of the label and object, the latter on the area of understanding, the task and the inventory rules. Reading the labels in their entirety does not mean that they already know where each piece is; or making the reading distance more accurate. Design should first determine the size and identity of the management, then use the real goods validation area, and then turn the collection into a reversible stock event.

Label modelling first distinguishes between “goods, singles, packaging and load”

SKU is used to distinguish a commodity specification, the same SKU may be expected to be hundreds of thousands of objects, so the SKU code cannot normally assume the unique identity of a single item. When it is necessary to trace each case, a differentiated serial identity should be assigned to each example; only when the whole box is handed over, you can manage the box number and the inside details; when tracking only the turn disk, you need to distinguish between the tray itself and the goods in this load. The size of the management particle is retrospective, the transfer and the breakout requirements are determined, and the number of labels cannot be calculated directly on the basis of the amount of SKU.

For example, a reusable tray carries the A shipment today and loads the B shipment tomorrow. The tray label is sustainable to mark this load, but the system must end old loading relationships and create new ones. If read to the P001 permanently and read to the A batch, the next turn-off may bring the wrong batch into stock. The same problem is between the box and the individual: once the whole box is unloaded, the original list of boxes cannot continue to represent the actual number of boxes.

Object Identity and Master Data The relationship changes must be recorded.
Commodity Specifications SKU, name, specifications, unit of measure unit conversion, version and start-up; not single-sequence identity
Single or fixed packaging Object ID, label identification, batch, owner Postmarks, change of logo, unboxing, boxing, scrapping
Swingbox or Tray Load ID and current loading relationship Load, unload, hand over and relationship time effective
The library and the reading point. Library position ID, reading area ID, equipment and antenna configuration Layout and configuration version; reading points do not directly equal the repository of inventory

EPC code separated from business fields and bound relationship preserved history

The EPC is the tag identification data whose code is consistent with the selected scheme, the label capacity and external system requirements. When the GS1 system is introduced, the appropriate identifier should be selected according to the specific object: SSGtin identifies the commodity example, SSCC is used for the logistics unit, and a set of internal codings with random digits cannot be called standard EPC. The company prefixes, serial number ranges and binary encodings should be validated according to the GS1 rules; if only using an enterprise internal identifier, the name space must also be clearly named, the exclusive scope and the external transfer rules.[1]

The database needs to manage at least the object list, label binding form, packaging or loading relationship table, reading event sheet, inventory schedule and inventory business record. The binder sheet should contain tag marking, object ID, time of commissioning, time of expiry and cause of change; the inventory record should keep material, batch, owner, store position, number and operational status. The variable inventory information should be maintained by the system and cannot rely on a fixed section of the label.

Label writing, read checksum binding should form a traceable process. After the operator changes the label, the old sign is to be removed from the valid bind; failed writing or binding should enter the queue for processing. The tag can be written by RFID printing equipment that meets project requirements, writing spaces or other suitable reading and writing devices, not all of which must be configured with RFID printers. Whether lock storage should be determined in conjunction with subsequent re-use needs, avoiding difficult recovery operations before the encoding is validated.

The " stock count radius" should be changed to a verifiable reading area

The field of on-site reading is not usually a antenna, but rather a clear circle or ball. Antenna orientation map, polarization approach, label direction, material to be marked, reflector reflection, feedback loss and movement paths affect readings. The distance that you can read at the farest and occasionally is different from the range that you can find within the specified time frame; the maximum reading distance in the equipment specification cannot be considered as the distance between shelf or installation.[2]

For example, what layers and what reservoirs are to be read in A-A-A-A-A-A-B-N-A-B-A-A-A-T. The increase may expand the coverage and may introduce more non-target labels; a simple reduction of power may also be lost together with the less readable labels within the target area. Compare tagging should be installed, antenna position, polarization, segregation of reading areas, combined effects of operating paths and equipment parameters rather than only one "common radius" [2] [3].

RSSI, read times and antenna sources can assist in screening candidates for the number of readings, but RSSI cannot be directly equal to distance, let alone prove that a label is a given library position with a fixed threshold value. Software rules should be verified in relation to mission scope and live evidence, and not cover up the serial reading problems caused by antenna layout. Channels also need to be accessed in a direction that triggers a signal, time series of multiple reading areas or other validated decision basis, and single antenna reading does not automatically prove that a library has been entered.

The number of acceptances and inspections with “expected collection” and “real collection” is not counted only.

First, a manual target label set E for a controlled test is prepared, and then the delabeled group O collected from this window is recorded. The valid target is the intersection of E to O, the missing goal is E less O, and the non-target read in the group is O less E. Here the count is different identities, not original readings; the same label is reported repeatedly, and cannot be counted as multiple items.

Example used only to describe the method of calculation, not the result of a client test:The target area has 100 confirmed labels, with 98 target labels and 3 adjacent area labels. Target reading rate is 98/100, or 98%; the target labels in the collection result are 98/101, or 97.03%. If only one report reads "to 101", it hides two missing readings and three non-targets. The indicator denominator, the statistical window and the determination range must be described together.

When further acceptance is performed, a distinction should also be made between the first reading and the after-retrospection results, the original device read and the final system adoption result, and the single performance and repeated test performance. It is recommended that a test matrix be established to keep the original log and configuration version item by item:

The receiving and inspection threshold should be based on an agreement of the cost of misleasing, operational nodes and manual reading.

Five: How do we get the results to WMS and decide if it's gonna get worse or worse?

The inventory should start with a list of tasks, and an expected inventory or label list and its version. The device capture results will form a candidate reading, the middle layer is decomposed to re-link the mission, WMS calculates differences and arranges for review. Unreaded labels may be physically missing, damaged, blocked or misscoped, cannot be deleted directly; read multiple tags can also come from a neighbourhood, and should not automatically execute the transfer library.

If you allow the receipt and dispatch to continue during an inventory, you should specify how to handle and move. You can freeze the selection for a short time, or record a baseline snapshot, and align the valid moving event during the inventory to the agreed check point. You cannot use the first book count as a denominator, but then compare the number of sites after the inventory is completed without checking the operations that occurred during the period.

The project can use the company SDK, LLR or other interfaces depending on the actual support of the device; using information to be sent or put in question, does not naturally decide whether to discard data, sustain it, confirm, retry, Zen and backlog monitoring are key to resilience.

The event should retain the object, task, point of understanding, time to collect, time to receive and process the results. The incident modelling can be used when sharing across the system; this is how business visibility information is organized, not to mean that a project that does only inventory in its warehouse must fully deploy the EPCIS platform.[4]

VI. Pre-delivery inspection: can it be explained where each difference comes from?

The project should deliver the object and coding rules, label binding and re-marking process, loading relationship maintenance rules, reading areas and antenna points bitmaps, controlled test logs, interface field lists, roundup and distribution rules and discrepancy approval processes. Reader solids, antenna positions, label type or packaging changes should be revalidated to the affected reading area instead of following the original test conclusion.

If the pilot is still different, it can be sorted in order: whether the labels are tied to the physical, if the list is expected to be accurate, if the target tags appear in the original log, if the software is assigned to the correct task, whether WMS is confirmed at the right time. Such a sort of sorting chain distinguishes between physical reading, data models and business rules, so that each count does not necessarily adjust power. The basis for determining whether RFID storage systems are changed is that each inventory is identified, time, place and business basis, rather than how far it is readable on the device panel.

Technical information and instructions on application

The test method and the data sheet here do not represent field measurements for the implementation of the design recommendations. The relevant standards and equipment information: [1] GS1 EPC Tag Data Standard, used to check identification objects and code rules; [2] Impinj " Choosing the Right RIAN RFID Reader Antenna for Your Resolution " ; [3] Impinj, IOT Data Interface API - Express Investment Consultants, for reference to the interpretation areas and parameter adjustments of different applications;[4] GS1 EPCIS and CBV Implication Guideline, for reference to operational event modelling. The specific frequency and power settings need to meet deployment requirements, and uniform frequency parameters are not provided here.