LoRa MLAT (L5)

Introduction

LoRa multilateration (MLAT) resolves a device’s position from a single connectionless LoRa broadcast heard by multiple gateways. The device transmits once, with no gateway connection required; Amazon Sidewalk estimates the distance to each receiving gateway from signal strength and multilaterates a position.

How the Cloud Resolves the Fix

Because one broadcast is heard by many gateways, their reports are accumulated and resolved together:

Each receiving gateway’s known position is looked up, and its reported signal strength is used to estimate the gateway-to-device distance.
Gateways whose positions are outliers relative to the rest are discarded, so a stray or mislocated gateway does not pull the fix.
The device position is estimated from the remaining gateways and their signal-derived distances, and an accuracy estimate is produced with the fix.

The resolved position is validated against the device’s recent location history before it is published, so a result that is physically implausible for the device’s movement is not published.

Accuracy and Constraints

Requires a network-synced LoRa link.
How well a broadcast resolves depends on the gateways that hear it: more gateways, spread around the device rather than clustered on one side, produce a better fix. A broadcast heard by only one or two gateways, or by gateways in poor geometry (collinear or clustered), resolves coarsely or may not resolve at all.
Because distance is estimated from received signal strength, signal quality and obstructions between the device and each gateway also affect accuracy.