Multipath — why RTLS accuracy degrades in real environments.

What multipath actually is.

A radio signal from a tag travels by line-of-sight to the anchor. It also bounces off any nearby metal or hard surface and arrives a few nanoseconds later.

The anchor receives both; the timing math assumes line-of-sight; the position estimate is wrong by the path length of the reflection.

UWB is more multipath-tolerant than BLE-AoA, which is more tolerant than RAIN RFID phase-based location, which is more tolerant than GPS indoors (which simply does not work).

No RTLS solves multipath in algorithm alone. Every production-grade RTLS solves it in placement geometry — anchor positions chosen so the strongest reflected path is geometrically distinguishable from the direct path.

The five most common multipath environments.

Metal racking warehouses: parallel vertical metal creates predictable reflection corridors. UWB and BLE-AoA need geometric offset.

Hospital corridors with glass partitions: glass is RF-translucent at BLE but reflects at higher UWB frequencies.

Pharma cleanrooms with stainless panels: stainless walls and ceiling create resonant cavities.

Automotive body shops: large flat metal surfaces plus moving metal create dynamic multipath that varies with production cycle.

HVAC-dense ceilings: ducting and trays at anchor mount height create line-of-sight gaps and reflective surfaces.

How a properly-designed survey identifies multipath risk.

Predictive RF modelling against the as-built environment, not the architectural drawing.

Material reflection map: every dominant flat metal or glass surface within 10 m of any proposed anchor is identified.

Anchor geometric offset calculation: anchors positioned so the strongest reflection is distinguishable from the direct path.

Multi-time-of-day validation: designs that work at 09:00 may fail at 14:00.

Multipath in locating systems: why reflections move the dot

Multipath means the receiver hears not only the direct radio path but also reflections from steel, concrete, vehicles, racking and bodies. Locating algorithms that assume the first or strongest path is the true path will bias position when that assumption fails. Narrowband RSSI systems are especially fragile; UWB’s wide bandwidth improves time resolution so first-path detection is more feasible — but blocked first paths still create error.

Where multipath hurts most. Metal-heavy manufacturing cells; dense pallet racking; hospitals with cable trays and HVAC at ceiling level; yards with stacked containers; and any aisle where forklifts intermittently occlude anchors. Static surveys miss the worst cases because the reflectors move.

Design responses. Raise anchors/locators; keep clear view corridors; avoid mounting behind tanks or rack backs; add redundancy so one NLOS path does not dominate; commission with traffic; and classify zones by expected error band instead of promising uniform centimetres. For Passive RFID, multipath and nulls show up as unstable read zones — fix with antenna aiming, power control and shielding rather than endless extra antennas.

Survey practices that catch risk early. Walk and drive tests at peak; compare empty vs occupied states; log bearing stability for AoA; inspect channel impulse response / first-path metrics for UWB where the platform exposes them; and mark “always-reflected” mount points as redesigns, not as problems to drown in denser hardware.

Operator takeaway. If accuracy collapses only when the line is running, you likely have a multipath/NLOS design issue, not a mysterious software bug. Fix geometry and process conditions before you approve a larger bill of materials.

Multipath myths and what to ask in an RF review

Myth: “UWB is immune to multipath.” Reality: UWB resolves paths better, but blocked first paths still bias fixes. Myth: “More anchors always fix reflections.” Reality: more poorly placed anchors can inject more bad measurements. Myth: “Empty-floor commissioning proves production accuracy.” Reality: moving metal and bodies are the dominant NLOS generators.

Questions for an RF review: Where are always-reflected mounts? Which zones lack four-anchor visibility? What is the plan for peak-hour validation? How are AoA arrays oriented relative to cable trays? For portals, where will adjacent-door energy leak? Written answers belong in the survey pack before purchase orders.

Link multipath findings to density and hybrid decisions

When a survey shows chronic multipath in a zone, options include redesign mounts, change radio layer, accept a coarser accuracy band, or reserve UWB only for the critical path inside that zone. Linking multipath findings to those choices is how locating architecture stays honest. Density alone is the expensive non-answer.

Multipath in locating systems: why reflections move the dot

Multipath means the receiver hears not only the direct radio path but also reflections from steel, concrete, vehicles, racking and bodies. Locating algorithms that assume the first or strongest path is the true path will bias position when that assumption fails. Narrowband RSSI systems are especially fragile; UWB’s wide bandwidth improves time resolution so first-path detection is more feasible — but blocked first paths still create error.

Where multipath hurts most. Metal-heavy manufacturing cells; dense pallet racking; hospitals with cable trays and HVAC at ceiling level; yards with stacked containers; and any aisle where forklifts intermittently occlude anchors. Static surveys miss the worst cases because the reflectors move.

Design responses. Raise anchors/locators; keep clear view corridors; avoid mounting behind tanks or rack backs; add redundancy so one NLOS path does not dominate; commission with traffic; and classify zones by expected error band instead of promising uniform centimetres. For Passive RFID, multipath and nulls show up as unstable read zones — fix with antenna aiming, power control and shielding rather than endless extra antennas.

Survey practices that catch risk early. Walk and drive tests at peak; compare empty vs occupied states; log bearing stability for AoA; inspect channel impulse response / first-path metrics for UWB where the platform exposes them; and mark “always-reflected” mount points as redesigns, not as problems to drown in denser hardware.

Operator takeaway. If accuracy collapses only when the line is running, you likely have a multipath/NLOS design issue, not a mysterious software bug. Fix geometry and process conditions before you approve a larger bill of materials.

Multipath myths and what to ask in an RF review

Myth: “UWB is immune to multipath.” Reality: UWB resolves paths better, but blocked first paths still bias fixes. Myth: “More anchors always fix reflections.” Reality: more poorly placed anchors can inject more bad measurements. Myth: “Empty-floor commissioning proves production accuracy.” Reality: moving metal and bodies are the dominant NLOS generators.

Questions for an RF review: Where are always-reflected mounts? Which zones lack four-anchor visibility? What is the plan for peak-hour validation? How are AoA arrays oriented relative to cable trays? For portals, where will adjacent-door energy leak? Written answers belong in the survey pack before purchase orders.

Link multipath findings to density and hybrid decisions

When a survey shows chronic multipath in a zone, options include redesign mounts, change radio layer, accept a coarser accuracy band, or reserve UWB only for the critical path inside that zone. Linking multipath findings to those choices is how locating architecture stays honest. Density alone is the expensive non-answer.