Proprietary downlink Reverse-TDoA architecture: fixed battery-powered beacons broadcast, the receiver inside the helmet listens.
Satellite navigation stops at the building perimeter. Bluetooth and Wi-Fi only indicate which room a worker is in — far too coarse for dangerous industrial zones. A person standing safely near a barrier and a person stepping inside an exclusion zone are fundamentally different operational events. Beyond a certain accuracy threshold, coordinates transform from simple map markers into verified events: zone entry, dwell time, and exact transit trajectories.
UWB has offered high-precision tracking for over a decade, but broad industrial adoption has stalled due to commercial realities rather than technical limitations. In conventional uplink-TDoA systems, fixed anchors must receive packets from every tag, maintain sub-nanosecond wired time synchronization, and backhaul measurements over Ethernet. Industry benchmarks show that roughly 85% of total deployment costs stem from cabling infrastructure alone: conduit runs, PoE switches, industrial lifts, and certified electrical labor.
Capacity is the second fatal bottleneck. Marketed tag counts are often theoretical arithmetic: with random ALOHA channel access at just 18% load, roughly a third of transmissions collide. The highest data-rate radio profiles marketed for capacity are the least resilient to structural steel and multipath reflections. In industrial safety, a dropped packet is an unmonitored worker.
Technical report TR-XENOM-2025-01, static testing on the DW1000 platform with four beacons.
We inverted the measurement pipeline. Fixed, low-power beacons broadcast proprietary UWB packets, and the receiver embedded in the helmet simply listens. Beacons operate autonomously without managing backhaul connections or listening for tags.
No timing distribution networks and no data cabling backhaul. At a standard 300 ms broadcast interval, a beacon runs for approximately six days on a standard 10,000 mAh external pack — field-tested and validated across continuous operations. Beacons are placed exactly where visibility is needed, not where network cables dictate.
Beacons and receiver RF paths require no complex per-device manual calibration. Deployment is pure radio planning: selecting anchor placements, logging survey coordinates, and validating coverage zones.
Helmets operate in purely passive listening mode. Adding 10,000 workers creates zero additional UWB transmissions and zero radio collisions. Infrastructure scaling depends strictly on site footprint and geometry — never on headcount.
The UWB receiver is built into the safety helmet. There is no secondary tag to inventory, recharge, retrieve, or replace, and zero chance of a worker leaving their positioning beacon behind in a locker.
We intentionally chose minimum data rates with long preambles, prioritizing RF link budget and multipath resilience around heavy steelwork over theoretical high-throughput figures that collapse in real plants.
The figures come from a furnished room, not an empty laboratory: sustained non-line-of-sight, radio paths crossing two plasterboard walls, two of the four beacons standing beside resin printers with metal post-curing chambers, and a mirror directly opposite a third. No Kalman filter, moving average or trajectory smoothing of any kind was applied.
Testing used four beacons and a stationary receiver on the DW1000 platform. The architecture is also implemented on DW3000, but these quantitative results do not transfer to it — nor to an industrial site at scale with moving personnel — without a separate measurement campaign.
A deliberately unfavorable stress series is part of the published data: coordinate availability fell to 15.79% and P95 radial error rose to 93.94 cm. Removing wired synchronization simplifies installation. It does not remove the need for coverage engineering and an acceptance test.
Presence in a hazardous zone, detected at the moment of entry
Time spent in work, hazard, service and waiting zones
Automatic radio channel switching by area
Access control for specific rooms
Personnel location during evacuation
Route and incident analysis, and event generation for the control room
The same architecture would suit indoor navigation on consumer phones — a passive listener computing its own position, with beacons that by design do not know who is listening. That is precisely the scenario FiRa calls Untracked Indoor Navigation. Apple introduced a DL-TDoA API in iOS 26 and Google added FiRa 4.0 support in Android 17, but our packet format is proprietary and phone system APIs do not understand it. Switching the beacons to standard FiRa mode would mean returning to a network of anchors that must constantly coordinate — losing the independence the architecture was built for. The direct path requires Apple and Android manufacturers to open developer access to received UWB frames and their hardware timestamps. Until that happens this is a research direction, not a shipped feature, and we keep the distinction deliberately.