Best Practices & Pitfalls

BLE (RSSI/AoA) checklist

• Do a mini site survey: metal shelving, elevators, glass walls = multipath hotspots.
• Use smoothing intentionally: Kalman/particle filters + map constraints (walls matter).
• Control transmit power + advertising rate: don’t “max power everything.”
• AoA needs calibration: antenna orientation, locator placement geometry, and periodic verification.

Common pitfalls

• Expecting RSSI to behave like distance in every room (it won’t).
• Treating AoA like “install and forget” (arrays drift; environment changes).

UWB checklist

• Decide TWR vs TDoA early: impacts sync, wiring, and latency.
• Plan anchor geometry: poor geometry ruins accuracy even with perfect radios.
• NLoS strategy: detect and down-weight NLoS ranges; validate with ground truth.
• Security mode: if you’re doing access/proximity, use modern secure ranging (802.15.4z STS concepts).

Common pitfalls

• Underestimating installation effort (power, mounting, network).
• “Accuracy disappointment” from anchor geometry + NLoS, not from UWB itself.

mmWave radar checklist

• Define what “position” means: presence, count, tracks, zones—don’t promise identity without an ID layer.
• Mounting height/angle matters: ceiling vs wall changes occlusion.
• Fusion plan: combine with BLE/UWB/badge events for identity.
• Privacy story: radar feels less invasive than cameras, but still needs governance.

Performance, Cost & Security Considerations (the decision matrix)

Performance trade-offs

• BLE RSSI: easiest, but accuracy varies; can be “couple meters” in good setups.
• BLE AoA: improved precision with more infrastructure and calibration; ~0.5 m class claims exist.
• UWB: strong precision; 5–10 cm class ranging cited widely, with modern security improvements in 802.15.4z.
• mmWave: excellent sensing; radar fundamentals enable ~3.75 cm range resolution with 4 GHz bandwidth, but practical tracking errors depend heavily on scenario.

Security (when it actually matters)

If you’re doing access control, “secure distance” becomes a requirement—not a nice-to-have.

• UWB 802.15.4z adds mechanisms (like STS) designed to harden ToF measurements.
• BLE can be secured at the protocol/app layer, but RSSI-based distance is easier to spoof than time-based ranging in many threat models (especially for “am I within 1 meter?” decisions).
• mmWave is not a “ranging handshake” tech by default; it’s a sensor—so threat models look different (spoofing, replayed tracks, sensor blinding, etc.).

‍

A Practical “Choose-Your-Stack” Framework (HowTo)

Step 1: Define the minimum viable location truth

Pick one:

• Zone certainty (room/area)
• Path tracking (movement)
• Precise coordinates (sub-meter)
• Secure distance (cm-level)

Step 2: Choose the architecture pattern

• Phone-first → start with BLE (RSSI/AoA)
• Tag-first precision RTLS → UWB
• Tagless sensing (counting, safety zones) → mmWave, then add identity layer if needed

Step 3: Plan infrastructure and operations

• Power + backhaul (PoE/Wi-Fi/LTE)
• Calibration workflow
• Health monitoring (dead anchors = bad data)
• Data pipeline (location events → alerts → analytics)

Step 4: Run a 2–4 week pilot with ground truth

• Measure median error, 95th percentile, and zone accuracy
• Test NLoS corners, crowds, and “worst rooms”
• Decide go/no-go with numbers, not anecdotes

Real-World Use Cases (mini case study patterns)

Pattern A: Warehouse asset tracking (forklifts + pallets)

• UWB for forklift-level precise tracking (fast movement, tight aisles)
• BLE beacons on low-value pallets for zone presence
• mmWave at choke points for people counting + safety zone alerts (no tags required for detection)

Why this works: you spend UWB where precision matters, and keep BLE for broad coverage. The radar layer improves safety analytics without tagging every human visitor.

Pattern B: Smart office space utilization

• mmWave for occupancy and footfall (privacy-friendly vs cameras)
• BLE for room-level “who/which device entered” events (where needed)

‍

FAQs

1) Is UWB more accurate than BLE for indoor positioning?
Usually, yes—because UWB is time-based ranging. UWB is often cited in the 5–10 cm precision class, while BLE RSSI tends to be meter-level and environment-dependent.

2) What’s the difference between RSSI and AoA in BLE?
RSSI infers distance from signal strength. AoA estimates signal direction using antenna arrays, often enabling better precision (with more infrastructure and calibration).

3) Can mmWave do indoor positioning without tags?
It can do tagless sensing and tracking (targets/paths/zones). But identity usually needs another layer (badges, BLE, UWB, etc.).

4) How accurate is UWB indoors in real buildings?
Accuracy depends on anchor geometry and NLoS. In good geometry and LOS, cm-level is achievable; poor geometry/NLoS can push errors higher even with UWB.

5) How many anchors/beacons do I need?
It depends on floorplan and required accuracy. Some enterprise guidance suggests planning on the order of one anchor per ~150–300 m² in certain deployments, then validating with a survey/pilot.

6) Does UWB work on smartphones?
Some modern phones support UWB, but not universally. If “works on any phone” is a must, BLE is the safer default.

7) Which is cheaper: BLE or UWB?
BLE tags are commonly cheaper (often $5–$25 per beacon-class device), while UWB tends to cost more due to anchors + sync + higher-precision tags.

8) Does mmWave work through walls?
mmWave can penetrate some materials depending on frequency and thickness, but indoor radar performance is heavily affected by reflection/attenuation. In practice, treat walls as significant obstacles unless tested.

9) What’s best for secure proximity (like access control)?
UWB with modern secure ranging concepts (802.15.4z STS) is commonly chosen for security-sensitive “distance matters” applications.

10) Can I combine BLE, UWB, and mmWave?
Yes—and many production systems do. A common approach: UWB for precision, BLE for coverage/phone UX, mmWave for tagless sensing, fused into one event stream.

‍