Busy sites trip more often
Commercial alarms can go off falsely up to three times as often as home alarms. More people share the codes, and the systems are more complex.
ASU POP Center · 2007Concept in development
When the alarm goes off, know what it is.
Most alarms turn out to be nothing, but someone still has to drive out and check. Pathfinder is being designed to go to the spot first, take clear photos and send them to you, so you can decide whether it needs a person.
- AlarmSensor trips at the east fence
- CheckGoes to look, inside approved zones
- YouPhotos arrive, you make the call
PATHFINDERExample Yard ▾
LiveEventsRoutesZones
4 units online1 alarm openT+00:00:00
Fleet · 4 units02:14:07
Unit 1ETA 01:38 · 112 m to go · 1.1 m/s
En route78%Unit 2Route 2 · 418 m covered · next WP-6
Patrolling64%Unit 3Dock · full in 38 min
Charging41%Unit 4Gate A · ready
Standby92%EventMotion · S-07 East fence
Unit en route
- Sensor triggeredS-07, perimeter PIR
- Unit dispatchedRoute 3 · 112 m · inside allowed zones
- Arrived on sceneHolding 3 m from fence line
- Photos captured3 angles · Camera on
- Sorted for reviewMotion region marked · clearest first
- Waiting for operatorUnit holds position until told
Illustration of the planned operator console. Not real data.
94–98%
of burglar alarm calls that police respond to turn out to be false. In some cities it’s higher still.
ASU Center for Problem-Oriented Policing · 2007
The problem
Most alarms are nothing. Someone still has to go.
A sensor can’t tell a person from a raccoon, a branch or a loose tarp, so every alarm gets treated as real. For the business paying for it, that means guard call-outs, police fees and owners getting out of bed to check an empty lot.
Every false call costs time
Each false alarm takes about 20 minutes of police time, usually for two officers. That’s time they aren’t spending on real calls.
ASU POP Center · 2007Repeat alarms get expensive
In Vancouver, three or more false alarms in a year can get a permit suspended. Reinstating a large commercial permit costs $250, and double after a second suspension.
City of Vancouver · 2026A big bill, added up
US police responded to about 36 million alarm activations in 2002, at an estimated cost of $1.8 billion for the year.
ASU POP Center · 2007Statistics describe the false-alarm problem in general. They are not results for Pathfinder.
How it works
Three steps. Nothing for you to do until it matters.
The idea is simple: check before anyone drives out. Here is how one alarm would play out in the planned design.
01 · Patrol
Go to the alarm
When a sensor trips, it heads to that spot on a route you’ve already approved. It keeps to the zones you’ve marked as allowed and stays under the speed limit you set.
02 · Verify
Take a closer look
Once there, it holds back from the fence line and takes photos from a few angles. The clearest ones are sorted to the top, so you aren’t scrolling through a dozen blurry frames.
03 · Alert
Hand it to a person
You get an alert with the photos attached. If it’s a raccoon, mark it as nothing and go back to sleep. If it looks real, escalate it to your guard company or the police.
Safety
Built so a person stays in charge.
These are the rules we’re designing around. None of them have been tested on hardware yet, and we’ll say so plainly until they have.
Hard limits, kept separate
Speed caps and no-go zones live in a separate safety layer, apart from the software that plans routes. If a command breaks a limit, it doesn’t get through.
A person makes the call
It can look and report, but deciding what happens next is always up to you. Nothing gets escalated on its own.
Every action logged
Where it went, what it saw and who decided what, kept in one timeline you can look back through.
System layers
Three layers. Safety in the middle.
Every command from the dashboard passes through the safety layer before it reaches anything on site. Photos and status travel back the other way.
DashboardWhat you see
Where you see alerts and photos, and where every decision gets recorded.
Safety layerChecks every command
Blocks commands outside the limits, whatever sent them.
On siteSensors and the unit
Designed to stop and wait if the connection drops.
Where it could fit
For big sites with nobody on hand.
Places where a false alarm usually means a long drive or a paid call-out, and where a few photos would settle it.
Vehicle lots
Rows of stock spread over a big lot, where one alarm could be anywhere along the fence.
Construction sites
Tools and materials left out between shifts, on a site that changes every week.
Self-storage
Long aisles of units that fixed cameras can only partly see.
Monitoring teams
Teams that need a clearer picture before they send someone out.
Roadmap
Where things stand.
Concept in development. Nothing built or for sale yet.
- NowCurrent step
Plan and simulate
Working out routes, zones and limits in software, and running them in simulation before any hardware exists.
Proves the plan works - Next
Prove the basics on a bench
Testing the limits and the lost-connection stop on a bench, away from any real site.
Checks the limits hold - Later
Try it on a real site
A small, supervised trial on a real site, with people watching the whole time.
Proves it helps
Know first. Then decide.
We’re still early. We’ll share more as each step gets tested.