6  RFID Standards and Protocols

6.1 Start With the Story

A working RFID system has several contracts at once. The tag and reader need an air-interface contract, the reader and software need a control contract, and the application needs event data that still means the same thing after filtering.

This chapter separates those contracts so design reviews do not mix them together. Start with the boundary that must interoperate, then choose the standard evidence that proves the right layer is doing the right job.

6.2 In 60 Seconds

RFID standards are easier to use when you keep the layers separate. ISO 14443, ISO 15693, ISO/IEC 18000, EPC Gen2, and NFC Forum specifications describe tag-to-reader behavior. LLRP describes a reader-control interface. EPC, EPCIS, and CBV describe identifiers and business events. Start with the workflow, choose the standard family that fits the read pattern, pilot it with real tags and materials, and record the settings and retest triggers.

Phoebe’s Field Notes: Why “Gain” Means Something Different at 13.56 MHz

Phoebe’s Why

Antenna gain, EIRP, and the familiar \(1/d^2\) Friis law are far-field ideas: they describe a radiated wave whose power density has already settled into an inverse-square spread. Close to the antenna – inside the reactive near field, roughly \(d < \lambda/2\pi\) – the fields have not detached from the antenna yet, and a small loop couples to a nearby loop the way two transformer windings couple: through mutual inductance, not radiation. That is the real reason ISO 14443 tap cards and ISO 15693 vicinity tags behave so differently from EPC Gen2 portals. At 13.56 MHz the whole HF interaction distance sits deep inside the near field, so “antenna gain in dBi” and “EIRP” are not the operating physics there; coupling coefficient and a steep near-field falloff are. At 860-960 MHz the tag sits comfortably in the far field of a portal antenna, so gain, EIRP, and the \(1/d^2\) backscatter link budget this chapter’s own dock-door pilot depends on apply cleanly.

The Derivation

Reactive near-field boundary:

\[d_{nf} \approx \frac{\lambda}{2\pi}\]

Far field, forward link (reader to tag), aperture form:

\[A_{tag} = G_{tag}\,\frac{\lambda^2}{4\pi}\]

\[P_{tag} = \mathrm{EIRP}\cdot G_{tag}\left(\frac{\lambda}{4\pi d}\right)^2\]

A passive tag only wakes up once \(P_{tag}\) clears its chip threshold \(P_{th}\), so the maximum power-up range is:

\[d_{max} = \frac{\lambda}{4\pi}\sqrt{\frac{\mathrm{EIRP}\cdot G_{tag}}{P_{th}}}\]

Inside the near field this aperture reasoning does not apply. Loop-to-loop magnetic coupling instead falls off close to \(1/d^3\) in field amplitude – close to \(1/d^6\) in power – far steeper than the far-field square law.

Worked Numbers: This Chapter’s HF And UHF Bands

• HF near field (13.56 MHz): \(\lambda = c/f = 22.1\) m, so \(d_{nf}=\lambda/2\pi=3.52\) m. An ISO 14443 tap (~2 cm) and an ISO 15693 vicinity read (~10 cm) are both far inside that 3.52 m boundary – neither is a far-field antenna problem. Comparing the two ranges with the near-field power law: \((10/2)^6=15{,}625\times\), a 41.9 dB drop, versus only \((10/2)^2=25\times\) (14.0 dB) if a far-field square law wrongly applied. The near field costs roughly 28 dB more than a naive \(1/d^2\) guess over the same 5x distance change – which is why HF range is measured in centimeters by physics, not by regulation.
• UHF far field (860-960 MHz, using 915 MHz): \(\lambda = 0.328\) m. A portal panel antenna of aperture \(D\approx0.3\) m has a far-field boundary \(2D^2/\lambda = 0.549\) m – well inside the chapter’s 2.4 m read zone, confirming the whole dock-door zone is a legitimate far-field, gain-and-EIRP problem.
• Forward-link range check: with a catalog-typical FCC-max EIRP \(=36\) dBm \(=3.98\) W, tag antenna gain \(G_{tag}=2\) dBi \(=1.58\times\), and a catalog-typical Gen2 chip threshold \(P_{th}=-18\) dBm \(=15.8\,\mu\)W, \(d_{max}=16.5\) m – about \(6.86\times\) the chapter’s stated 2.4 m dwell zone.
• What that margin means: because \(d_{max}\) so comfortably exceeds the pilot’s 2.4 m zone, the dock-door read is link-budget-rich, not power-starved – which is exactly why the chapter’s own review questions focus on Q-tuning, dwell time, and material shielding rather than raw range. HF’s near-field ceiling, by contrast, cannot be bought back with more EIRP; the coupling law itself sets the short range.

6.3 Learning Objectives

By the end of this chapter, you will be able to:

• Separate air-interface standards from reader-control protocols and event-data standards.
• Choose between ISO 14443, ISO 15693, EPC Gen2, NFC Forum, and active RFID families based on workflow evidence.
• Explain why EPC Gen2 anti-collision needs a tuned inventory round instead of only higher reader power.
• Describe how EPC identifiers become useful business events through middleware, EPCIS, and common vocabularies.
• Review a proposed RFID standards choice for interoperability, operations, privacy, and retest risk.

6.4 Quick Check: RFID Standards

6.5 Why Standards Matter

An RFID project usually fails at one of three boundaries:

• The tag and reader cannot communicate reliably.
• The reader cannot be managed consistently across sites or vendors.
• The raw tag reads do not become shared, auditable business events.

Standards reduce those risks, but they do not all solve the same problem. ISO 14443, ISO 15693, ISO/IEC 18000, EPC Gen2, and NFC Forum specifications describe how devices communicate over the air. LLRP describes a network interface for controlling compatible readers. EPC, EPCIS, and the GS1 Core Business Vocabulary describe how identifiers and read events are represented for supply-chain visibility.

Figure 6.1: RFID standards grouped by responsibility: air interface, NFC behavior, reader control, and event sharing

6.6 Standards By Responsibility

6.6.1 Air-Interface Standards

Air-interface standards define how the reader and tag communicate: frequency family, modulation, command sequence, timing, and anti-collision behavior.

• ISO 14443: HF proximity cards and devices at 13.56 MHz. Use it when intentional close-range interaction matters, such as tap access, identity cards, and contactless card workflows.
• ISO 15693: HF vicinity cards at 13.56 MHz. Use it when the item can be read farther than a tap but still benefits from HF behavior, such as library, document, and item-level asset workflows.
• ISO/IEC 18000 family: RFID air interfaces across frequency families for item management. The UHF branch is the relevant family for many passive supply-chain and inventory systems.
• EPC Gen2 / ISO 18000-63: The common UHF passive item-identification air interface used in many supply-chain RFID deployments. Older material may describe this lineage as ISO 18000-6C; current procurement and compliance language should be checked against the exact specification and region.

6.6.2 NFC Forum Specifications

NFC builds on HF proximity technology and defines predictable behavior for phones, tags, and readers. In practice, NFC decisions often combine two questions:

• Which RF technology does the tag use, such as NFC-A, NFC-B, NFC-F, or NFC-V?
• Which data behavior is required, such as NDEF records, reader/writer mode, card emulation, or peer interaction?

Do not treat every 13.56 MHz tag as an NFC tag. A phone-compatible workflow needs both compatible RF behavior and the right NFC data format.

6.6.3 Reader-Control Protocols

Reader-control protocols sit above the air interface. LLRP is used to control compatible RFID readers from client software. It can standardize inventory operations, reporting, and reader-client communication, but it does not erase every vendor-specific configuration detail.

In a review, ask which layer the team is standardizing:

• Air interface between tag and reader.
• Reader control between middleware and reader.
• Event schema between business systems.
• Operational process for commissioning, monitoring, and retesting.

6.6.4 Event-Data Standards

The Electronic Product Code gives physical things a structured identifier. EPCIS and the Core Business Vocabulary help turn observations into shareable events such as what was observed, where it was observed, when it happened, and why it matters to the business process.

This distinction is important. Reading a tag is not the same as proving that a shipment was received, a medicine was dispensed, or an asset changed custody. The event layer needs business context and reviewable evidence.

6.7 EPC Gen2 Dominates UHF Supply-Chain RFID

When people say “RFID” in supply-chain and retail contexts, they often mean EPC Gen2: the EPCglobal UHF Class-1 Generation-2 air interface, standardized internationally as ISO/IEC 18000-63 and operating in the UHF 860-960 MHz band. It defines radio behavior between reader and tag plus the tag data conventions that let a Gen2 reader from one vendor inventory Gen2 tags from another.

The other standards fill in the interaction patterns Gen2 does not cover. ISO/IEC 14443 supports proximity behavior at 13.56 MHz HF, ISO/IEC 15693 supports vicinity behavior at 13.56 MHz HF, and ISO 11784/11785 covers LF animal identification around 125-134 kHz. Picking a standard is really picking a band, range class, interaction pattern, and evidence model together.

The important design habit is to keep the standards family tied to the interaction that must be proven. A medicine cabinet badge tap is not the same problem as a dock-door carton portal. In the cabinet case, the review evidence may say: staff presents a badge within a few centimeters, the cabinet logs one deliberate access event, and a missed read sends the nurse to a manual identity check. That points toward ISO 14443 or NFC-style proximity behavior because short range is part of the control.

In the dock-door case, the evidence may say: a pallet spends 1.6 seconds in a 2.4 m read zone, cartons can be stacked, and the system must identify many passive labels without line of sight. That points toward EPC Gen2 UHF because bulk inventory and anti-collision are the central mechanisms. If a pilot reads 47 of 50 tagged cartons on dry cardboard but only 39 of 50 when foil packaging is added, the standards decision is not “Gen2 failed.” It is that Gen2 remains only a candidate until tag placement, antenna geometry, shielding, or exception handling closes the evidence gap.

Standards make interoperability possible; they do not remove the need to test the actual workflow. A reviewable standards packet records the object class, range expectation, material exposure, tag family, reader-control method, event-data owner, privacy exposure, and retest trigger.

6.8 Choosing A Standards Family

Start with the workflow, not with the technology name.

Figure 6.2: RFID standard selection lanes for tap, vicinity, bulk portal, and active asset workflows

Use this sequence during design review:

1. State the object being identified: person token, consumable, case, pallet, reusable asset, tool, document, animal, vehicle, or container.
2. State the interaction pattern: deliberate tap, short vicinity read, portal pass, handheld inventory sweep, gate crossing, or periodic active beacon.
3. Check the environment: metal, liquids, body proximity, stacked items, speed through the read zone, neighboring readers, privacy exposure, and expected exceptions.
4. Select the candidate standards family.
5. Pilot with representative tags, readers, mounting positions, and process events.
6. Record the standard, regional settings, reader-control method, data schema, and retest trigger.

6.8.1 Common Selection Patterns

• Use ISO 14443 and NFC Forum behavior when the user must intentionally present a card, phone, or tag at close range.
• Use ISO 15693 or NFC-V behavior when HF vicinity reads are useful and phone or specialized-reader support is part of the workflow.
• Use EPC Gen2 UHF when many passive items must be identified without line of sight and the environment supports reliable UHF reads.
• Use active RFID or RTLS-specific systems when the asset must announce itself over a larger area or when battery-powered location behavior is required.

6.9 Folded Standards Selection Record

A standards decision should link the air interface to the workflow, data model, and pilot evidence:

Review field	What to write down
Object and event	Person token, animal, document, tool, case, pallet, vehicle, or container, plus what a read should prove.
Frequency and coupling	LF/HF inductive coupling, UHF radiative backscatter, active beaconing, or NFC tap behavior.
Standards family	ISO 11784/11785, ISO/IEC 14443, ISO/IEC 15693, ISO/IEC 18000 family, EPC Gen2, NFC Forum, or event-data standards such as EPCIS.
Reader and tag construction	Reader placement, antenna geometry, tag package, on-metal behavior, durability, and regional settings.
Backend event handling	Duplicate filtering, missed-read exceptions, wrong-read handling, event schema, and business-system ownership.
Pilot evidence	Representative objects, mounting, motion path, read rate, false reads, duplicate reads, and retest trigger.

Do not collapse all of this into “RFID standard selected.” The radio standard, reader-control protocol, and event-data standard solve different problems and usually have different owners.

6.10 EPC Gen2 Anti-Collision

Bulk UHF RFID works because the reader manages an inventory round. When many passive tags are energized at the same time, they cannot all answer at once. EPC Gen2 uses a slotted process where the reader announces a Q value, tags choose slots, and collided tags retry in later rounds.

Figure 6.3: EPC Gen2 inventory round showing reader Q value, tag slot choice, successful slots, empty slots, and collision retry

The practical review question is not “what is the maximum read rate?” It is:

• How many tags are expected in the field at the same time?
• How long does the object stay inside the useful read zone?
• How many antennas are switching during the inventory cycle?
• How much tag orientation and material variation exists?
• Does the reader adjust Q automatically, or is a fixed setting being forced?
• What read evidence proves the process works on the real package mix?

6.10.1 Q Value As A Planning Heuristic

The slot count is 2^Q. A useful starting estimate is:

\[ Q \approx \lceil \log_2(\text{expected tags in the field}) \rceil \]

That estimate is only a starting point. A larger Q reduces collisions but lengthens a round. A smaller Q shortens a round but increases collisions. The best setting depends on dwell time, tag count, antenna sequence, reader firmware, and the materials in the read zone.

Use these review statements instead of brittle promises:

• “The reader configuration was piloted with the real tag population and dwell time.”
• “The system records missed-read exceptions and retests after package, antenna, or reader changes.”
• “Dense reader behavior was reviewed when neighboring readers could interfere.”
• “The event layer does not assume that a single read is proof of business custody.”

6.11 The Select, Query, And ACK Handshake

A Gen2 reader does not just scan. It runs a defined command sequence to single out one tag at a time from a crowd:

1. Select narrows the population to tags matching a filter, such as a company prefix, so the reader does not inventory the whole warehouse.
2. Query opens an inventory round and sets a slot count from the parameter Q; each tag picks a random slot in 0..2^Q-1.
3. When a tag’s slot counter reaches zero, it backscatters a 16-bit random number called RN16.
4. The reader echoes that RN16 in an ACK; only the tag that sent it replies with its PC bits and EPC. Echoing the RN16 is how the reader proves it is talking to that one tag.
5. For reading or writing memory, Req_RN returns a handle used by the Read, Write, Lock, and Kill access commands.

This is the anti-collision mechanism: random slotting plus the RN16 handshake lets many tags share the channel and be read one by one. Getting Q close to the population size is what separates a fast inventory from a stalled one.

Use the Q value as a planning estimate before the pilot. If a dock-door reader expects about 48 tags in the field, ceil(log2(48)) = 6, so Q = 6 gives 2^6 = 64 slots. That is a plausible starting point because it gives the population room to spread out. If the reader starts at Q = 4, there are only 2^4 = 16 slots, so many of the 48 tags will choose the same slots and collide. If it starts at Q = 9, there are 512 slots, so the round can waste time stepping through mostly empty slots. Both extremes can lose reads when the pallet is moving.

Now add dwell time. A pallet moving through a 2.4 m useful read zone at 1.2 m/s is available for 2.4 / 1.2 = 2.0 seconds. If the reader cycles four antennas evenly, each antenna has roughly 2.0 / 4 = 0.5 seconds of dwell time before the package geometry changes. That does not produce a universal maximum tag count because real timing depends on modulation settings, tag sensitivity, reader firmware, collisions, retries, and materials. It does give the design review a concrete question: did the pilot prove that the selected Q behavior, antenna sequence, and filter rules complete enough inventory rounds inside the actual dwell time?

6.12 From Identifier To Business Event

A standards-compliant read is still just an observation. The system becomes useful when the observation is filtered, associated with business context, and shared in a form that other systems can interpret.

Figure 6.4: RFID data path from tag identifier to reader, middleware filtering, EPCIS event, and business application

Review each layer separately:

• Tag memory and identifier: Is the identifier encoded consistently? Who owns the numbering scheme? How are duplicates prevented?
• Air-interface behavior: Which standard family is used? Which regional settings, channels, power constraints, and tag classes apply?
• Reader control: How are readers configured, monitored, and updated? Which settings are standardized and which remain vendor-specific?
• Read filtering: How are duplicate reads, ghost reads, missed reads, and antenna zones handled?
• Event representation: What business step, location, disposition, and timestamp are recorded?
• Sharing boundary: Which partners or systems can query events, and what privacy or access rules apply?

6.13 Gen2 Memory Banks Are Separate Evidence Fields

Reading a Gen2 tag well means knowing where its data lives. The standard defines four memory banks that are addressed independently:

Bank	Name	Contents
00	Reserved	32-bit kill password and 32-bit access password.
01	EPC	CRC-16, Protocol-Control bits, and the EPC identifier itself.
10	TID	Tag identifier: chip maker and model; many tags add a permanent unique serial.
11	User	Optional application data, not present on the cheapest tags.

Figure 6.5: Gen2 tag memory banks showing reserved controls, EPC, TID, and user memory

Two practical consequences follow. First, the EPC in Bank 01 is writable and clonable, so anti-counterfeiting work should read the factory-locked unique serial in the TID bank beside the EPC. Second, the passwords in Bank 00 are only 32 bits and are sent with light cover-coding, so classic Gen2 security is weak by design. The security chapter builds on that point.

Use memory banks to write acceptance tests, not just documentation. Suppose a reusable tool is issued with a 96-bit EPC. That is 96 / 8 = 12 bytes of identifier payload, but the reader report will usually include more than those 12 bytes because the EPC bank also carries protocol-control information and a CRC. The business system should parse the EPC payload according to the encoding scheme the organization actually owns, then store the reader, antenna, timestamp, and event reason beside the observation.

A stronger commissioning test reads both EPC and TID. The record might say: expected EPC prefix matched, TID manufacturer and model matched the approved inlay, 48 of 48 pilot labels produced the same EPC/TID pair across three inventory passes, and 0 duplicate EPCs appeared with different TIDs. That catches a common weak design: accepting only the writable EPC as truth. If an operator finds two tags that both report the same EPC but different factory-locked TIDs, the event layer should flag a clone or encoding error rather than silently merging the observations.

6.14 Check TID Versus EPC Evidence

6.15 Matching Standards To Roles

6.16 Worked Review: Receiving Doorway

A manufacturer wants to scan incoming cartons as pallets pass through a dock doorway. Each pallet may contain a mixed number of tagged cartons, and neighboring dock doors may have their own readers.

6.16.1 Candidate Choice

EPC Gen2 UHF is a reasonable candidate because the workflow needs passive, non-line-of-sight, bulk identification. The review should not stop there.

6.16.2 Evidence To Collect

• Pallet speed and read-zone depth, used to estimate dwell time.
• Expected tag count per pallet, including unusually dense loads.
• Tag placement on cartons, especially near metal, liquid, foil, or stacked material.
• Reader antenna layout, antenna switching behavior, and neighboring reader schedule.
• Inventory-round settings, including whether Q is automatic or fixed.
• Middleware filter rules for duplicate reads and missed-read exceptions.
• Event mapping from reads to receiving, custody, and exception states.

6.16.3 Acceptance Record

A defensible record should include:

• Candidate standards and exact product conformance claims from vendors.
• Regional RF settings and deployment constraints.
• Pilot evidence with representative cartons and pallet speeds.
• Missed-read handling and manual exception path.
• Retest trigger after tag, packaging, reader firmware, antenna layout, or process changes.

6.17 Common Mistakes

6.17.1 Treating Frequency As The Standard

“13.56 MHz” and “UHF” are not enough. The review needs the actual standard family, tag behavior, data format, and reader-control assumptions.

6.17.2 Using One Standard For Every Workflow

A hospital, warehouse, or campus can legitimately use several RFID families. Force-fitting every workflow into one standard often creates either privacy risk, missed reads, or poor user intent.

6.17.3 Confusing Tag Reads With Business Events

An antenna read does not automatically mean receiving, shipping, dispensing, or custody transfer. EPCIS-style events require context and business vocabulary.

6.17.4 Assuming LLRP Removes All Vendor Differences

LLRP can standardize an important reader-control interface, but antenna mapping, diagnostic data, firmware features, and device management may still differ.

6.17.5 Ignoring Retest Triggers

Standards conformance is not a one-time proof. Packaging, mounting, regional settings, reader firmware, tag supplier, neighboring readers, and workflow speed can all require retesting.

6.18 Review Checklist

Before approving an RFID standards choice, confirm that the team can answer these questions:

• Which workflow is being supported, and what evidence shows the read pattern is correct?
• Which standards layer is being selected: air interface, NFC behavior, reader control, identifier syntax, or event sharing?
• Which exact tag class, memory behavior, and regional radio settings apply?
• How are multiple tags, collisions, missed reads, and duplicate reads handled?
• What event data is produced, and which business meaning does it carry?
• How will privacy, access control, and partner sharing be governed?
• What changes require the pilot evidence to be repeated?

6.19 Reference Notes

Use official standards bodies and current vendor conformance statements when writing procurement language:

• GS1 EPC/RFID standards for EPC, EPC Gen2 lineage, and related supply-chain RFID guidance.
• GS1 EPCIS and CBV standards for event-sharing and vocabulary guidance.
• GS1 LLRP material for reader-client interface guidance.
• NFC Forum specifications for NFC device and tag behavior.
• ISO/IEC 18000-6 overview material for UHF RFID item-management air-interface context.

6.20 Summary

RFID standards are most useful when they are assigned to the right layer. ISO 14443, ISO 15693, ISO/IEC 18000, EPC Gen2, and NFC Forum specifications shape tag-to-reader behavior. LLRP helps control compatible readers. EPC, EPCIS, and CBV help transform observations into shared business events. Good design review starts with the workflow, chooses the candidate standards family, pilots the real environment, and records the retest triggers.

6.21 Key Takeaway

RFID standards choices should document the air interface, data model, reader integration, certification need, and interoperability boundary.

6.22 Concept Relationships

• Frequency-band knowledge explains why LF, HF, UHF, and active systems behave differently.
• Tag-type knowledge explains why passive, battery-assisted, and active tags fit different workflows.
• RFID security knowledge explains why range, identity, and event sharing create privacy and misuse risks.
• Deployment knowledge explains how antenna placement, reader scheduling, and exception handling affect the final outcome.

6.23 What’s Next

• RFID Standards Summary: use the consolidated reference after this conceptual standards chapter.
• RFID System Components: connect standards choices to tags, readers, antennas, middleware, and applications.
• RFID Tag Types: review passive, battery-assisted, active, and specialty tag choices.
• RFID Design and Deployment: turn the standards decision into a site survey and rollout plan.

6.24 Sequence Review