Positioning

Bluetooth 6.0 vs UWB in the Warehouse: One Radio per Asset Class

Forklifts and pallets do not need the same radio. How to split a warehouse tag fleet between Bluetooth 6.0 channel sounding and UWB, with real numbers.

Hayat Amin, President of IP, Position Imaging Hayat AminPresident of IP, Position Imaging 4 min read
The short answer

Deploy UWB for assets that need 10 to 30 cm accuracy and fast updates: forklifts, AMRs, and automated handoffs. Deploy Bluetooth 6.0 channel sounding for the long tail of pallets, totes, and containers, where 1 m accuracy is enough and a $3 to $8 tag beats a $10 to $25 UWB tag at five-figure fleet sizes. Most warehouses should run both radios under one location engine, split by asset class rather than by zone.

Key takeaways

  • UWB holds 10 to 30 cm accuracy in steel racking; Bluetooth 6.0 channel sounding manages 0.3 to 0.5 m in line of sight and drifts to 1 to 2 m in dense aisles.
  • A finished channel sounding tag lands around $3 to $8 because ranging runs on the main SoC; a UWB tag runs $10 to $25 with its dedicated transceiver.
  • Give UWB to the 50 forklifts that need 10 Hz updates; give channel sounding to the 10,000 pallets that need an aisle, not a centimeter.
  • Channel sounding needs CS-capable silicon on both ends; a BLE gateway bought before 2025 can hear tags but cannot range with them.
  • Hybrid deployments live or die on the fusion layer between the two radios, and that layer is where the granted patents sit.

How accurate is Bluetooth 6.0 channel sounding on a warehouse floor?

Bluetooth 6.0, published by the Bluetooth SIG in September 2024, added channel sounding: an initiator and reflector sweep phase measurements across up to 72 channels in the 2.4 GHz band, with round-trip timing as a cross-check against smeared or spoofed phase data. In open line of sight, early silicon in the class of Nordic's nRF54L series measures distance to roughly 0.3 to 0.5 m. A warehouse is not open line of sight. Steel racking, metal totes, and forklift traffic at 2.4 GHz create reflections that corrupt the phase sweep, and error grows to 1 to 2 m in dense aisles, sometimes worse when the direct path is fully blocked.

UWB under IEEE 802.15.4z works differently: roughly 500 MHz of bandwidth at 6.5 or 8 GHz, with pulses around 2 ns long. The receiver can separate the first-arriving path from a reflection only a meter longer, which is exactly the skill racking demands. Expect 10 to 30 cm, and expect it to hold as the building fills up.

Racking decides the number, not the datasheet.

What does each tag cost, and where does the rest of the money go?

Channel sounding's cost story is simple: it runs on the tag's main Bluetooth SoC, so there is no second radio to buy. CS-capable parts price close to ordinary BLE silicon at volume, and a finished tag with battery, enclosure, and antenna lands around $3 to $8. A UWB tag carries a dedicated transceiver from the Qorvo DW3000 class plus a BLE side channel for wake and configuration, and finished tags run $10 to $25.

Tags are only half the bill:

  • UWB anchors cost roughly $100 to $500 each installed, and TDoA setups need wired backhaul or tight wireless sync between them.
  • Channel sounding needs CS-capable fixed nodes too. A BLE gateway bought in 2023 can hear the tag's advertisements but cannot range with it.

The deciding variable is fleet shape. Tracking 50 forklifts, the UWB tag premium is under $1,000 and irrelevant next to the vehicles. Tracking 10,000 pallets, the same premium is $100,000 or more, repeated every hardware refresh.

Multiply by your tag count first.

Which warehouse assets justify UWB?

Reserve UWB for assets where centimeters and update rate carry money or safety:

  • Forklifts and AMRs. Collision slow-down zones and traffic management need sub-30 cm positions at 10 Hz or faster. UWB two-way ranging delivers tens of fixes per second. A channel sounding procedure takes tens of milliseconds per peer and is scheduled point to point, so it starves at vehicle rates.
  • High-value tooling and calibrated equipment. When one misplaced torque tool stops a production line, a $20 tag is rounding error.
  • Automated handoffs. Crane-to-conveyor and robot-to-rack transfers need two machines to agree on position, which is exactly a two-device time-of-flight problem.

Power follows the update rate. A UWB tag ranging at 1 Hz typically lives about a year on its cell; drop to a fix every few minutes and multi-year life is realistic. Vehicle tags sidestep the issue by drawing power from the truck.

Fast-moving metal earns the expensive radio.

Where is Bluetooth 6.0 channel sounding the right call?

Pallets, totes, roll cages, and returnable containers ask a different question: which aisle, which staging lane, which dock door. One-meter accuracy answers it. Channel sounding at a ranging every two to five minutes runs for years on a coin cell, and at $3 to $8 per tag you can afford to tag everything instead of sampling the fleet.

Two cautions before committing a 2026 deployment:

  • Silicon maturity. The spec is from September 2024, the first CS-capable SoCs shipped through 2025, and cross-vendor interop is still settling. Pilot with one silicon vendor on both ends.
  • Accuracy claims. Vendor figures of 0.3 to 0.5 m assume line of sight. Measure in your own racking before sizing the rollout, because a 2 m error can still misreport the aisle.

If 1 m locates the asset and the fleet is five figures, channel sounding is the default.

At pallet scale, tag price wins.

How do you run both radios without building two systems?

Most real warehouses end up hybrid: UWB on vehicles and critical assets, channel sounding on the long tail, one map and one API on top. The radios are the easy part to buy. The hard part is the software between them: filtering noisy ranges, rejecting multipath outliers, fusing two radio types plus inertial data into one coherent track, and doing it in real time across thousands of tags. That layer took the industry years to get right, and it is where the granted patents sit.

Position Imaging licenses hundreds of granted US patents across radio-frequency ranging, real-time positioning, computer vision, and machine learning, IP that Apple, Bosch, and other major firms cite in their own filings. Licensing that layer means your team builds the product, the workflows, and the integrations instead of re-deriving ranging math, and ships in months with freedom to operate.

License the hard parts, build the product.

Patents referenced
US 11,774,249US 12,079,006US 12,066,561US 12,000,947

Frequently asked questions

Is Bluetooth 6.0 channel sounding accurate enough for warehouse asset tracking?

For aisle-level and zone-level tracking, yes. Expect 0.3 to 0.5 m in line of sight and 1 to 2 m in dense steel racking, which reliably locates pallets, totes, and containers. It is not enough for forklift safety zones or automated handoffs, which need UWB's 10 to 30 cm.

How much does a UWB tag cost compared to a Bluetooth channel sounding tag?

A finished UWB tag typically runs $10 to $25 because it carries a dedicated transceiver alongside a BLE radio. A channel sounding tag lands around $3 to $8 since ranging runs on the same SoC that handles everything else. At 10,000 tags, that gap is $100,000 or more per refresh cycle.

Can my existing BLE gateways do channel sounding?

No. Channel sounding requires CS-capable Bluetooth 6.0 silicon on both ends of the measurement. Gateways deployed before 2025 can still receive tag advertisements for presence detection, but they cannot perform the phase-based ranging exchange. Budget for new fixed nodes alongside the tags.

Should I deploy channel sounding and UWB together in one warehouse?

Usually yes, split by asset class. Put UWB on forklifts, AMRs, and high-value equipment that need 10 to 30 cm at 10 Hz, and channel sounding on the pallet and container fleet where 1 m is enough. The key is a single location engine that fuses both into one map and one API.

What update rate can Bluetooth channel sounding support?

Each ranging procedure takes tens of milliseconds and runs point to point between two devices, so it suits fixes every few seconds to minutes, not continuous vehicle tracking. For assets that move on forklifts rather than under their own power, a fix every two to five minutes is typically enough and preserves coin cell battery life for years.

Talk to the IP team

Send us your tag fleet mix and accuracy targets, and we will map your warehouse tracking product to the specific patents it needs.

Tell us the product. We map the exact scope, what a license covers, and how fast you can ship, all in a 20-minute call.

Book a 20-minute call