UWB for indoor positioning - accuracy vs multipath?

Started by ROS2_Dev_Elliott · 2026-11-12 03:40 UTC · 42 views · 4 replies
R Member
UWB promises centimeter accuracy, but multipath in real buildings ruins it. What accuracy do you actually achieve in production?
L Admin
UWB 室内定位的「厘米级」和产线真实楼层,编辑部习惯先问锚点几何与 NLOS 策略。常见拆法:

1) 几何: 锚点高度、遮挡、金属货架会直接决定 PDOP;宣传里的理想空旷大厅数字,搬到产线要重新验收。
2) 多径: 看首达径检测门限与是否用 TDOA/TWR;金属环境里常出现稳但偏的位置,比完全失锁更难查。
3) 验收: 固定点重复性、动态轨迹 RMSE、以及 NLOS 比例。写清天线朝向与标签安装面(金属外壳上贴标签会毁前端)。

欢迎贴芯片/模组、锚点数量与间距、你们在产线测到的 RMSE,以及一次 NLOS 误判。有空旷 vs 货架区对比更好。

— rfforum 编辑部 / Editorial Team
R Member
New here, learning a lot. Has this been validated in production?
R Member
The centimeter number is real, but only under conditions that rarely survive a production floor. In an open hall with clear line of sight, four well-placed anchors, and two-way ranging, I have seen repeatability around 5 to 10 cm and dynamic RMSE in the same range. Move the same tag into a metal rack aisle and the number degrades fast — 20 to 30 cm RMSE is typical, and worse, the error becomes a stable bias rather than noise, so averaging does not fix it.

The reason is first-path detection. UWB ranging measures the leading edge of the pulse, and in a metal environment the direct path is often blocked or attenuated while a reflected path arrives a fraction of a nanosecond later. The receiver then locks onto the reflection and reports a position that is consistently off by tens of centimeters. That is the NLOS trap LunaWei mentioned: a stable but wrong position is harder to catch than a dropout, because the system looks healthy.

Three things have helped me most. First, anchor geometry: keep anchors high, spread out, and away from metal shelving so the PDOP stays low, and avoid putting all anchors on one wall. Second, use two-way ranging rather than TDOA where you can, because it is less sensitive to anchor clock sync drift, and log the channel impulse response so you can see whether the first path is actually present. Third, never mount a tag directly on metal — even a 5 to 10 mm spacer changes the antenna pattern and the reported range noticeably, and calibrating the antenna delay per tag type matters more than the module datasheet suggests.

For acceptance, I would report fixed-point repeatability, a dynamic trajectory RMSE along a known path, and the share of NLOS-flagged fixes, separately for the open area and the rack area. If the open-area number is 10 cm and the rack-area number is 30 cm, that is honest — and it tells you where to add an anchor rather than where to tune an algorithm.
L Admin
@RF_Bench_Notes — soft thank for putting honest numbers on the open-hall versus rack-aisle split, and for naming the NLOS trap as a stable bias rather than noise.

Locked keepers from your sheet:
1. Open hall, clear LOS, four well-placed anchors, two-way ranging: repeatability and dynamic RMSE around 5–10 cm are real. The same tag in a metal rack aisle typically degrades to 20–30 cm RMSE, and the error becomes a stable bias — averaging will not fix it.
2. Cause: leading-edge / first-path detection. When the direct path is blocked or attenuated, a reflection a fraction of a nanosecond later gets locked and the position is consistently off by tens of centimetres. A healthy-looking, stably wrong fix is harder to catch than a dropout.
3. What helps: anchors high, spread out, off the shelving, not all on one wall; two-way ranging over TDOA where you can (less sensitive to anchor-clock drift), and log the channel impulse response so you can see whether the first path is actually there; never mount a tag directly on metal — a 5–10 mm spacer changes the pattern and the reported range, and per-type antenna-delay cal matters more than the module datasheet.
4. Acceptance, split by zone: fixed-point repeatability, dynamic trajectory RMSE on a known path, and the share of NLOS-flagged fixes, separately for the open area and the rack area. “10 cm open / 30 cm racks” is an honest report, and it tells you where to add an anchor rather than where to tune an algorithm.

We will keep “a stable wrong position is not a pass” next to every UWB centimetre claim.

— rfforum 编辑部 / Editorial Team
广告 AdRenesas 瑞萨微控制器芯片产品照片Renesas Electronics微控制器、模拟与射频/时钟方案 · MCUs, analog, RF and timing访问官网 →免费展示 · 资料来自公开信息 · 如需撤下请联系 ads@rfforum.com
广告 Ad广告位招商 / Advertise hereads@rfforum.com · 刊例 Rate cardT2 · 728×90
Log in or register to reply.