VisionDesign rev A, stage 2 ready (29 Oct 2026): smart gondola pair that runs the aisle itself (9 calls, 6 alone, 3 assist), canopy call lights and tone, ATECC608 and tamper switch inside; 28 camera views checked in the CAD; parts $704 per aisle at qty 1; not built

AisleOne

IEMI's one product for a store aisle: the aisle's own shelving, made smart

The two gondola runs that face each other across a store aisle, at the same size and place, watch each other's shelves and the floor between them from cameras inside their top canopies: they raise restock tasks with a photo and check the restock, flag shelf-edge labels that do not match the price file, call spills with red lights and a tone, check the wet-floor sign, call a blocked aisle or a person down, and send the store one task list. No face recognition and no live video leaves the store.

Who buys it: Owners and ops managers of independent grocers and convenience stores (1 to 10 stores), one aisle at a time; then regional chains.
Why it pays: $920 + $15/month per aisle (stage 2, was $690) instead of about $938 + $91/month for two shelf cameras, a share of a spill-camera system and a plain lit canopy pair bought separately (est.): about $1,100 against $2,030 in the first year.
Target price: $920 per aisle (two smart canopies on the store's own gondolas) + $15/month staff app (target)
First-minute wow: A jar breaks in aisle 4 at 11:42. Before anyone hears it, the canopy over the spot turns red, the aisle sounds a tone, and the nearest phone shows a floor crop; at 11:51 the lights go off on their own because the floor matches the clean picture and the sign is gone.

See the CAD design Concept page

AisleOne concept: the idea and where each part goes — The concept. The CAD design is below, under Printable CAD.

What the research found

And how this design answers it

Rev A (28 Oct 2026): the retail flagship is the aisle's own gondola pair. In each top canopy: a Raspberry Pi 5 (brain in run A), two Camera Module 3 Wide aisle cameras looking across, four call lights, a tone, ATECC608, tamper switch and a potted IRM-20-5 on the canopy light feed
Makes the aisle's calls itself: shelf gap, restock done, possible spill, wet-floor sign, blocked aisle and aisle traffic alone; label price, weekly price accuracy and person down assist; unsure means 'go and look'
Privacy first: no face recognition (FTC Rite Aid order 2023), people masked out of stills and crops, no live video leaves the store
Legal in the aisle: 1,830 mm clear (ADA 403.5.1 minimum 915), canopy never past the base deck (ADA 307.2), same runs and shelves
Folds StoreEye, ShelfScan and SpillGuard; door counts, till queues, cold rooms and the back door stay with their own zones
Stage 2 (29 Oct 2026): buyer check (Vusion Captana with Carrefour, Simbe Tally 4.0 robots at regional grocers, icetana spill alerts on CCTV), parts prices checked ($704 per aisle at qty 1), both brains now the same-size Raspberry Pi 5 2 GB, price $690 to $920 + $15/month for a margin (est.), top-three FAQs

Use case

A 6-aisle independent grocer on a Saturday with two staff

Store ownerShop-floor staffDuty manager

01Trigger: the store opens; each canopy takes a clean-floor picture and a first shelf still
02Detect: four cameras per aisle score the shelves opposite and check the floor twice a second
03Act: restock tasks with photos, spill lights and tone, sign check, blocked-aisle and person-down alerts
04Log: one signed task list and a weekly aisle report

Outcome: Gaps are filled in minutes instead of at the next shelf walk, and every spill has a signed call, sign and clear time

How it works

Architecture

01
Sense: two cameras in each run's top canopy, looking across at the other run and the floor
02
Decide: run A's Raspberry Pi 5 merges both runs and runs the gap, label, spill, sign, blocked and person-down rules; 'go and look' below the line
03
Show: amber and red call lights along the canopy, with a tone
04
Tell: one signed task list over the store Wi-Fi to the staff app

Hardware

Real parts, buyable today

Bill of materials: about $704 per aisle at qty 1 for the two smart canopies (prices checked 29 Oct 2026; canopy, cover and wiring est.); the store's gondola runs are kept.

Block	Part	Why this part
Brain	Raspberry Pi 5 2 GB (x2, one per canopy)	Run A's merges the aisle and makes the calls; run B's runs its own cameras
Aisle	4 x Raspberry Pi Camera Module 3 Wide	Each run sees the other's shelves and the floor up to it
Calls	8 x Adafruit NeoPixel Stick 8 + Adafruit 160 piezo	Amber and red lights along the canopy, with a tone
Security	Adafruit 4314 ATECC608 + Omron D2F-01	Signed calls, tamper log
Power	Mean Well IRM-20-5 (x2)	5 V from the canopy's own light feed

Mechanical

Printable CAD

4 parts, generated from code and checked as closed solids. Pick a part to inspect it, or download the STL files to print.

Loading the CAD…

Real CAD · drag to rotate · scroll to zoomReal CAD · swipe sideways

Open risks we are working on

Spill vision on glossy, sunlit floors is unproven; it stays 'possible spill, go and look'
Price labels at 2 m across the aisle need good stills; the first build tests which label sizes read
A camera cable of 1 m from the Pi 5 to each camera needs a bench test (est.)
Shoppers need clear camera notices; some stores will want the cameras to stop at shelf level
UL 962 and radio tests are still to be done; no safety claim: it does not prevent falls

Design stage only: nothing on this page is for sale. Specifications and prices are targets.