Installing the hardware

Written for whoever is actually on site with a drill. The theme throughout: what you install decides what the numbers can mean, and no amount of configuration afterwards changes that.

Decide what each space needs to answer

Work backwards from the question, not forwards from the catalogue.

The question	What it needs
Is this room in use right now?	A radar, or a fixed counting sensor
How many people are in it?	A radar, or a counting sensor
Where is this asset?	Tags, and three or more gateways spread around the space
Is this person in the building?	Tags, and one gateway per area is enough
How busy was this entrance?	A people counter on the door
How many people are in the building?	People counters on every door, netted up
How warm is it?	An environmental sensor

The row that catches people out is the second. Tags cannot answer it. A room with gateways and no radar reports how many tags are in it, which in an office where nobody wears one is zero, all day, correctly. Occupancy: which number is which is the ten minutes that prevents this conversation.

What goes where

• Gateways — they hear tags and anchor positioning.
• Radars — they count people, tagged or not.
• People counters and tags — counting people through a doorway, and following things that move.

Then, before you pack up: the checks worth doing on site, because every one of them costs five minutes now and a return visit later.

Two rules that apply to everything

Record where you actually put it, not where the plan said. Positioning is computed from where Omaya believes the hardware is. A gateway recorded three metres from where it hangs pulls every estimate near it towards a place nothing is. If you moved something during the install, move it on the plan too.

Power and mounting outlive the electronics. A sensor on a battery in a ceiling void is a return visit in eighteen months. Where mains is available, use it; where it is not, note the install date so the battery forecast has somewhere to start from.