A Beacon deployment is only as accurate as the surface it is mounted on and the calibration behind it. Two identical Beacon units can report wildly different distances if one sits behind a metal panel and the other is calibrated against real walls. Getting installation and calibration right is the difference between a system that merely blinks and one that locates.
Mounting Height and Orientation
Most indoor Beacon units perform best mounted 2 to 4 meters above the floor, tilted slightly downward toward the coverage area. Mounting too low invites occlusion by furniture, carts, and people; mounting too high shortens the useful near-field. Keep the antenna face clear of direct contact with metal — a metal backing plate can shift the radiation pattern and add 5 to 15 dB of loss. Aim the Beacon so its broadside faces the zone you care about rather than a wall.
Material Attenuation
Radio waves do not treat all matter equally. A single drywall partition may cost 3 to 6 dB, while a concrete wall can eat 10 to 20 dB. Glass with low-emissivity coating and metal meshes are surprisingly lossy; a filled warehouse rack behaves like a partial Faraday cage. Even the human body absorbs around 3 to 8 dB at 2.4 GHz. Build an attenuation map during the site survey so the positioning engine knows where signals will dip.
The RSSI-to-Distance Model
The workhorse is the log-distance path-loss model:
RSSI(d) = A – 10*n*log10(d) + Xo
where A is the measured power at one meter, n is the path-loss exponent (typically 2.0 in free space, 2.5 to 4.0 indoors), d is distance, and Xo is Gaussian shadowing. A and n must be calibrated per site, not taken from the datasheet, because walls and ceilings dominate n. Calibrate by placing a reference device at known distances and fitting A and n with a least-squares pass.
Multipath and Error Compensation
Indoor reflections create multipath: the same packet arrives via several paths, and the receiver reports a fluctuating RSSI. Combat it with exponential smoothing or a Kalman filter to stabilize the estimate, then prefer fingerprinting — recording RSSI from multiple Beacon anchors at surveyed points — over pure trilateration when the geometry is poor. Averaging over a short window also rejects transient nulls.
Calibration Workflow
A repeatable calibration runs in five steps: (1) import the floor plan, (2) mark anchor positions, (3) walk a grid of reference points logging RSSI, (4) fit A and n and build a correction table, (5) validate by checking that test points land within tolerance. Recalibrate after structural changes — a new shelving run can silently break last quarter’s model.
Conclusion
A Beacon is a precision instrument only after it is placed with intent and tuned to its room. Spend the calibration hour up front and the location accuracy pays you back on every fix.