Patent Teardown: A 5G Beamforming Invention — What Does the Claim Actually Protect?

1. Why Should Engineers Read Patents?

Most engineers treat patents as impenetrable legal documents — walls of dense claim language wrapped in legalese. But here is the truth: patents are the most detailed public documentation of how industry solves real engineering problems. Every granted patent lays bare a specific technical problem, the prior art that failed to solve it, and a novel approach that the inventor believes is new and non-obvious.

As a telecom engineer who has been diving deeper into patent analysis, I decided to do something practical: take a real, granted US patent in 5G beamforming, read it cover-to-cover, and write down exactly what I learned. This article is the result of that exercise.

The patent I chose is US 10911128 B2 — titled "Techniques for Beam Discovery and Beamforming in Wireless Communications." It was filed by Qualcomm and addresses a very specific problem in 5G NR millimeter wave (mmWave) systems: how does a UE know that the base station has changed its beam sweep configuration?

2. Patent Identity Card

2.1 Decoding the Patent Number: What Does "B2" Mean?

When you see a patent number like US 10911128 B2, each part carries specific meaning. The suffix letter-number code is called the WIPO Kind Code — it tells you exactly what stage of the patent lifecycle you are looking at:

3. Anatomy of This Patent

Before dissecting the claims, it helps to understand the building blocks of any US patent. Here is how this patent is structured:

3.1 The Specification (Description)

The specification is the narrative portion — paragraphs [0001] through [0082]. It describes the technical background, the problem being solved, and the detailed implementation. Key sections include:

• Background [0002]-[0006]: Sets the stage — 5G NR, mmWave path loss, the need for beamforming, and the challenges of beam management overhead.
• Summary [0007]-[0012]: High-level overview of the invention — a method for signaling CSI-RS beam changes via a change indication message.
• Detailed Description [0018]-[0082]: The deep technical implementation — change indication bit embedded in CSI-RS waveform, scrambling sequences, hierarchical beam search, restore mode, and timing protections.

3.2 The Drawings (FIGS. 1-4)

The patent includes four figures:

• FIG. 1: System block diagram — UE and network entity (gNB) with beam management components, transceiver, RF front end, processor, and memory.
• FIG. 2: Multi-UE wireless system — multiple UEs communicating with multiple gNBs, showing coverage areas, backhaul links, and core network.
• FIG. 3: Beam discovery flow chart — the complete beam search state machine including CSI-RS burst reception, change indication bit extraction, restore mode logic, and hierarchical search.
• FIG. 4: Method flow diagram — the step-by-step method of blocks 402-410 covering protection interval determination, CSI-RS reception, change indication extraction, beam search interruption, and restore mode entry.

3.3 The Claims (30 total)

Claims are the legally operative part — they define what the patent owner can exclude others from doing. This patent has:

• 3 Independent Claims: Claim 1 (method), Claim 11 (apparatus), Claim 21 (non-transitory computer-readable medium) — these are essentially the same invention described in three different statutory categories.
• 27 Dependent Claims: Claims 2-10 depend on Claim 1; Claims 12-20 depend on Claim 11; Claims 22-30 depend on Claim 21. Each dependent claim narrows the parent claim by adding more specific limitations.

4. The Engineering Problem: Beam Sweep Changes in mmWave

To understand what this patent actually protects, we must first understand the problem it solves. Here is the scenario:

4.1 Beam Sweeping in 5G NR

In mmWave 5G (frequencies above 24 GHz), the path loss is extreme. To compensate, both the gNB and UE use beamforming — focusing radio energy into narrow directional beams. The gNB periodically transmits a CSI-RS beam sweep: a burst of reference signals where each OFDM symbol carries beams pointing in different spatial directions, either frequency division multiplexed (FDM) or time division multiplexed (TDM).

The UE uses these CSI-RS sweeps to discover and maintain suitable beam pairs. This is critical for both initial access and ongoing link maintenance.

4.2 The Specific Problem: Dynamic Beam Sets

Here is where the challenge lies. The gNB does not keep its beam set static. As UEs enter or leave the cell, the gNB may:

• Add new beams to cover newly arrived UEs.
• Remove beams to save power when UEs depart.
• Reshuffle existing beams to different time-frequency resources for consolidation.

The problem: the UE has no way of knowing that these changes occurred. If the UE is in the middle of a hierarchical beam search (which spans multiple CSI-RS bursts), a silent beam set change can invalidate the entire search. The UE keeps searching for beams that no longer exist, or misses newly added beams. The hierarchical search — which relies on consistent beam mapping across bursts — breaks down.

4.3 Why Not Just Broadcast the Change?

A natural question: why not just have the gNB broadcast a message saying "I changed the beam set"? The patent addresses this directly in paragraph [0028]: in mmWave, a broadcast message is prohibitively expensive because it must itself be transmitted in a beam-swept manner — you would need to send the change notification on every beam, which defeats the purpose of saving resources.

5. The Invention: Change Indication Signaling via CSI-RS

The patent's solution is elegant in its simplicity: embed a single-bit change indication message directly into the CSI-RS waveform itself.

5.1 How It Works

The mechanism described in the specification (paragraphs [0029]-[0031]):

1. Each CSI-RS beam's waveform carries a change indication bit.
2. The gNB toggles this bit whenever it modifies the set of beams in the sweep (add, remove, or reshuffle).
3. The bit is conveyed through the choice of scrambling sequence: when bit=0, waveform uses scrambling sequence A; when bit=1, it uses scrambling sequence B. The sequences are cell-specific (depend on cell ID and symbol number).
4. The UE performs hypothesis testing: upon receiving CSI-RS, it descrambles with both sequence A and sequence B, performs channel estimation, and picks the hypothesis with higher likelihood. This yields the detected change indication bit.
5. The UE can jointly extract the change indication from all identified beams using maximal ratio combining (MRC) for robustness.

5.2 What Happens When a Change Is Detected

The patent also describes the UE's behavior after detecting a beam set change (FIG. 3):

• If no change: Continue the ongoing hierarchical beam search normally.
• If change detected during a search: Interrupt the hierarchical search and enter a restore mode — systematically search for the time-frequency locations of previously identified beams to salvage as much prior search work as possible.
• Protection intervals: The gNB may be restricted to only changing beam sets at specific times (e.g., start of every fourth radio frame), giving the UE a guaranteed window where it does not need to monitor the change indication bit.

6. Claim 1 Dissection: Element-by-Element

Now let's do what patent analysts actually do — break Claim 1 into its constituent elements and analyze what each one means technically.

6.1 Analyzing Each Element

Preamble — "A method of wireless communications"

This establishes the statutory category (method) and the general field. The preamble is often not limiting in US patent law unless it gives "life, meaning, and vitality" to the claim. Here, it is fairly generic.

Element 1A — Receiving a CSI-RS beam with a change indication message

This is the foundational element. It requires:

• A UE (not a base station — this is a UE-side method).
• Reception of a specific CSI-RS beam that is part of a set of CSI-RS beams (i.e., a beam sweep).
• The CSI-RS beam must include a change indication message — this is what makes it novel. Existing CSI-RS beams in the prior art did not embed change indication signaling.

Element 1B — Extracting the change indication message

The extraction step is explicitly claimed. This is significant — the claim does not just require the message to be present; the UE must actively extract it. The specification describes this as hypothesis testing between two scrambling sequences.

Element 1C — Determining whether the beam set has changed

The determination is based on the value of the change indication message. The claim does not specify the value encoding (that is left to dependent claims) — it only requires that the UE use the value to decide if a change occurred.

7. Claim Tree: Independent vs Dependent Claims

Understanding the claim hierarchy reveals how the patent builds layers of protection. Here is the claim tree for all 30 claims:

7.1 Method Claims (Claims 1-10)

Claim 1 [Independent — Method]
  ├── Claim 2: Change indication = 1-bit (add/remove beams)
  ├── Claim 3: Hierarchical beam search, all symbols, all sub-arrays, 1st iteration
  │     ├── Claim 4: Subsequent burst search using one sub-array (2nd iteration)
  │     └── Claim 5: Max RSRP threshold → best sub-array → different directivity beams
  ├── Claim 6: No change → check restore mode → finish if threshold not reached
  │     └── Claim 7: Restore = search for previously identified beams
  │           └── Claim 8: Restore includes finding highest RSRP symbol
  ├── Claim 9: No change + not in restore → hierarchical search on all symbols
  └── Claim 10: Joint extraction via maximal ratio combining (MRC)
        

7.2 Apparatus Claims (Claims 11-20)

Claims 11-20 mirror Claims 1-10 exactly but in apparatus form (receiver + memory + processor). This covers the physical device — any 5G UE chipset implementing this method.

7.3 Computer-Readable Medium Claims (Claims 21-30)

Claims 21-30 again mirror Claims 1-10 but cover the software/firmware stored on a non-transitory medium. This covers the firmware running on a modem processor.

7.4 Key Dependent Claim Analysis

Claim 2 (One-bit indication) — Narrows the change indication to exactly one bit. First value = change; second value = no change. Also specifies that "change" means beams added or removed. This is the most specific encoding limitation.

Claims 3-5 (Hierarchical beam search) — These claims protect the multi-iteration beam discovery process: first iteration uses all sub-arrays with omni-directional patterns; second iteration uses the best sub-array; RSRP threshold triggers narrower beams. This maps directly to the specification's description of the three-round search process.

Claims 6-8 (Restore mode) — These protect the novel "restore mode" where the UE salvages previously found beam locations after a beam set change. Claim 8 adds that the restore process uses RSRP comparison across symbols to relocate beams.

Claim 10 (MRC extraction) — Protects the specific technique of jointly extracting the change indication bit from multiple beams using maximal ratio combining, improving detection reliability.

8. Claim Mapping to 5G NR Architecture

Let's map the key claim elements to real-world 5G NR components:

9. Commercial Significance & SEP Context

9.1 Why This Patent Matters Commercially

Beam management is one of the most critical differentiators in 5G NR performance, especially at mmWave frequencies. The techniques covered in this patent address a fundamental operational issue — beam sweep configuration changes — that every mmWave gNB and UE must handle.

Key commercial implications:

• Qualcomm's modem dominance: As the assignee, Qualcomm can embed this technology into Snapdragon modem chipsets that power a majority of 5G smartphones globally.
• Potential SEP status: If the change indication mechanism described here becomes part of the 3GPP NR specification (or is necessary to implement the CSI-RS beam management framework), this patent could be declared a Standard Essential Patent (SEP), requiring FRAND licensing.
• Licensing leverage: Even if not an SEP, this patent covers a practical optimization that competitors would want to implement for beam search reliability in mmWave deployments.
• Priority date advantage: The March 2017 priority date places this patent early in the 5G NR standardization timeline (3GPP Rel-15 was frozen in June 2018), meaning the invention was conceived during the standards-formation period.

9.2 Freedom-to-Operate Considerations

For any company implementing 5G NR mmWave beam management, this patent poses an FTO question: does your beam search implementation use a change indication mechanism embedded in CSI-RS? If so, you may need a license from Qualcomm — either through a direct license or through a patent pool like the one administered by MPEG LA or Avanci for cellular SEPs.

10. My Perspective as a Patent Analyst

After spending considerable time reading this patent end-to-end, here are my key observations and learnings:

10.1 What Impressed Me

• Elegance of the solution: Using the existing CSI-RS scrambling sequence as the carrier for a change indication bit is brilliant. It requires zero additional signaling overhead — the information piggybacks on a reference signal that the gNB already transmits. No new messages, no new channels, no broadcast overhead.
• Claim breadth vs. specification depth: Independent Claim 1 is intentionally vague about the encoding mechanism, while the specification goes deep into scrambling sequences, hypothesis testing, and MRC. This is textbook patent drafting — the broad claim captures the concept; the specification provides enablement.
• The restore mode concept: The patent doesn't just describe how to detect a change — it describes what to do about it. The restore mode (salvaging prior beam discovery work rather than discarding it) shows real implementation thinking, not just theoretical hand-waving.
• Protection interval timing: Restricting when the gNB can change beam sets (e.g., every fourth radio frame) is a practical constraint that balances gNB flexibility with UE search reliability. This kind of detail reveals deep system-level thinking.

10.2 What I Learned About Patent Reading

• Claims are the patent. Everything else is context. The specification can describe dozens of embodiments, but only what is in the claims is legally protected. Always start and end with the claims.
• Dependent claims reveal the real invention. While Claim 1 is broad, Claims 3-5 (hierarchical search) and Claims 6-8 (restore mode) tell you what the inventors really thought was novel and worth protecting at a granular level.
• The three-pillar structure is strategic. Method/apparatus/CRM claims cover different potential infringement scenarios. If you manufacture and sell the device, the apparatus claim applies. If you only provide firmware updates, the CRM claim applies. If you operate a network, the method claim may apply to the UE's operation.
• Priority date matters more than grant date. This patent's priority date of March 2017 places it squarely in the 5G NR development window. Patents filed during standards development carry extra weight in SEP licensing negotiations.

10.3 Potential Weaknesses I Noticed

• UE-side only claims: All 30 claims are written from the UE perspective. The gNB's role (inserting the change indication, toggling the scrambling sequence) is described in the specification but not independently claimed. A competitor could potentially implement a different gNB-side mechanism that triggers similar UE-side behavior.
• No explicit frequency limitation: The claims are not limited to mmWave. While the specification motivates the invention with mmWave path loss, the claims are broad enough to cover sub-6 GHz CSI-RS beamforming as well — which could be an advantage or could face prior art challenges.
• "Change indication message" is undefined in claims: Claim 1 does not specify that the change indication is a bit, or that it's embedded in the scrambling sequence. Only Claim 2 narrows it to a one-bit indication. This means Claim 1 technically covers any mechanism for embedding change information in CSI-RS, which is very broad.

11. Key Takeaways for Engineers

12. Further Reading & Patent Search Resources

• US 10911128 B2 on Google Patents
• 3GPP 5G System Overview — Official 5G NR specifications
• WIPO PatentScope — International patent search
• ETSI IPR Policy — SEP declaration framework for telecom standards
• 3GPP TS 38.214 — NR Physical Layer Procedures for Data (CSI-RS configuration and beam management)
• 3GPP TS 38.331 — NR RRC Protocol Specification (beam reporting, measurement configuration)