Platform
One platform, three layers
Client-side applications, the Logidot Cloud, the physical hardware and the WMS / ERP
Pick a layer to read it, or jump to its section.
Our Technology
Our Technology
To empower the digital transformation of your operations, you need to know the precise location and status of the relevant assets on the shop floor. In addition, you need to connect these in real-time and enhance them with computational power, so that they interact in the fastest and smartest way possible. Logidot helps you achieve this through three core technologies: RTLS, IoT platform and simulator.
RTLS
Technology powering our RTLS
UWB, GPS, IMU, RFID, QR and sensor fusion on one floor; TWR and TDoA
Logidot combines a handful of technologies to reliably track assets with industry-leading accuracy.
UWB
Ultra-wideband (UWB) is a short-range wireless communication protocol - like Wi-Fi or Bluetooth - which uses radio waves of short pulses over a spectrum of frequencies ranging from 3.1 to 10.5 GHz. Its unique combination of signal characteristics - a fine time resolution and resistance to interference and reflections - makes it a secure, reliable and accurate choice for asset tracking.
Find out more.GPS
Using the latest GPS chips, we can reliably track assets outdoors with submeter accuracy and in some cases also indoors by combining signals from multiple satellites.
IMU
IMUs can measure forces, angular rates and orientation using a combination of accelerometers and gyroscopes. It allows for accurate location tracking over short trajectories in case of occlusions or GPS or UWB signal is lost.
QR and Barcodes
A barcode, consisting of bars and spaces, is a machine-readable representation of numerals and characters. Today, stripes as shown below on packages of products sold at supermarkets, convenience stores and other stores are ubiquitous.
RFID
Uses electromagnetic fields to automatically identify and track tags attached to objects. An RFID system consists of a tiny radio transponder, a radio receiver and transmitter. When triggered by an electromagnetic interrogation pulse from a nearby RFID reader device, the tag transmits digital data, usually an identifying inventory number, back to the reader.
Sensor Fusion
By combining the data from multiple physical sensors via fusion algorithms, an accurate and robust tracking performance can be achieved.
Algorithms- 3D accuracy [R95] (cm)
- < 30
- Max range (m)
- 65
- Sensor fusion
- 6-axis IMU (accelerometer + gyroscope), GPS
- GNSS (optional)
- Multiband GNSS
The floor itself
State-of-the art sensors for all use cases
Anchor, vehicle tag and wearable tag
-
Anchor
A fixed UWB sensor that gives moving tags their position. Runs for months on its built-in battery, or from mains power.
-
Vehicle Tag
Interfaces with our gateway, integrates UWB, a 6-axis IMU (accelerometer and gyroscope) and GPS. Battery life of one week with the option to be attached directly to the vehicle power source.
-
Wearable Tag
Interfaces with our gateway, integrates UWB and a 3-axis accelerometer. Battery life of one day.
The layer the other two stand on: tags on the vehicles and the people, anchors on the racking, a gateway off the floor. What they measure leaves the building once and arrives in the cloud as position, identity and status.
- Tags
- Anchors
- Gateway
- 3D accuracy [R95] (cm)
- < 30
- Max range (m)
- 65
- Operating frequency
- 6.5 to 8.5 GHz
- Protection level
- IP67 (outdoors) and IP54 (indoors)
IoT Platform
Our real-time IoT platform gives location, ID, and status data for various assets and translates into measurable processes. The platform, therefore, makes it possible to monitor, optimize, and automate processes in production and logistics. The open and highly elastic architecture enables easy integration and a rapid ROI.
- Anchor Fixed on the racking
- Gateway The link off the floor
- Wearable Tag Worn by the people
- Vehicle Tag Fitted to the fleet
- Dashboard Read in the office
- 3rd Party Sensors What is already on site
Anchor
Anchors are the building’s reference points. Screwed, magneted or strapped to racking, columns and walls, each one ranges every tag in view over UWB — months on the internal battery, or wired into site power.
- UWB
- Up to 65 m range
- Battery or wired
Gateway
The gateway is the one box between the floor and the platform: it collects what the anchors hear and forwards it over whatever the site already has — Ethernet, Wi-Fi or a mobile connection.
- 3G/4G/5G, Wi-Fi, Ethernet
- PoE, DC or AC
- Encrypted end to end
Wearable Tag
A wearable tag puts the people on the floor on the same live map as the vehicles — UWB for position and a 3-axis accelerometer for what they are doing. A day on a charge, and nothing to operate.
- UWB
- 3-axis accelerometer
- A day per charge
Vehicle Tag
Forklifts, pallet trucks and tugs carry a vehicle tag. UWB indoors, a 6-axis IMU for motion and GPS when the work goes outside — a week on battery, or taken straight off the vehicle’s own power.
- UWB · IMU · GPS
- 3D to 10 cm
- A week, or vehicle power
Dashboard
Positions are shown on your own floor plan, with zones, dwell times and routes. Your WMS, ERP or MES can read the data over the API.
- Live view in 3s
- Analytics in 5 min
- Open API
3rd Party Sensors
The architecture is open at the edge too. Sensors you already have — presence, environment, gates, scanners — stream in beside Logidot’s own, so one platform carries the whole picture.
- Existing hardware
- One stream
- Rapid ROI
RTLS
How We Compare
| Technology | Accuracy | Range | Positioning | Use cases | Anchor battery | Tag battery | Infrastructure cost | Deployment |
|---|---|---|---|---|---|---|---|---|
| Wi-Fi | <20 m | <150 m | 2D | Medium | Medium | Low | $$ | Hard |
| Bluetooth (BLE) | <6 m | <50 m | 2D | Medium | Medium | High | $ | Easy |
| RFID/NFC | <10 cm | <1 cm | Spot | Low | Low | Passive | $$$ | Hard |
| Existing UWB | <20 cm | <100 m | 2D/3D | Low | High with TDoA | Low | $$$ | Hard |
| Logidot | <20 cm | 65 m | 2D/3D | High | High with TDoA | High | $ | Easy |
RTLS
Benefits
High-performance
up to 10cm accurate in 3D and low latency
Adjustable update rate
up to 16Hz
Long range
up to 50m intersensor distance
Robust
against reflections, noise, vibrations and shocks
Scalable
simply add more beacons
Self-calibrating
can be deployed in a matter of minutes
Where the work is done
Cloud
The Logidot Cloud: the RTLS platform, the insight pipeline on it, the WMS across the top and the three surfaces
Hosted for you on Google Cloud, so there is nothing to install or keep running on site. The RTLS platform fuses every reading into one position per asset; on it, the analytics layer turns tracks into measures, the intelligence layer reads those measures against how the operation is meant to run, and the automation layer acts on what it finds — rules, alerts and the reports that go out on their own. The history is stored, backed up daily and encrypted end to end.
WMS (3rd party)
The system of record you already run
Stays where it is. The platform exchanges orders, stock, master data and positions with it in both directions, over the REST API.
Client-side applicationsSimulator service
The operation, run forward
A model of your floor, run on this same pipeline before anything moves in the building.
SimulatorAnalytics layer
Positions become measures
Where tracks turn into the numbers an operation is run on: time by zone, travelled time against order fulfilment, spaghetti maps, utilisation and productivity. Analysed within five minutes of the movement itself.
- Zone analytics
- Spaghetti maps
- Analysed in 5 min
Intelligence layer
Measures become judgements
What the analytics measure is read against how the operation is meant to run, so that a busy hour can be told from an unusual one and a pattern worth acting on from noise. It is what turns the numbers below it into the judgements the automation layer acts on — the three together are one pipeline, from position to insight to action.
- Patterns
- Anomalies
- What matters now
Automation layer
Supervision, automated
Rules you set watch the live picture for you and flag the anomalies — a vehicle in a restricted zone, an asset that has not moved, a check that was skipped — while the reporting tool sends the regular numbers out without anyone assembling them.
- Custom rules
- Alerts
- Automated reporting
RTLS Platform
What everything above stands on
The real-time location layer. UWB ranging from the anchors, motion from the tags and GPS where the work goes outdoors, fused into one position per asset against your own floor plan — 3D to 10 cm, up to 16 Hz, on screen within three seconds. Vehicles, goods, maintenance and people all arrive here first.
- 3D to 10 cm
- Up to 16 Hz
- Live in 3s
REST API(s)
For the systems: WMS, ERP, MES
Everything the platform knows, readable by the systems that run the work, and their data coming the other way.
Client-side applicationsDesktop app
For the office
The control tower on any desk: the live map, analytics, alerts and reports.
Client-side applicationsMobile apps
For the floor
Logidot Operative for picking, and the checks and logs that belong on the floor.
Client-side applications- Database
- Google Cloud Platform
- Latency
- 5 min. to analytics
3s for live view - Security
- Standard enterprise grade encryption with end-to-end web security based on SSL
- Safety
- Daily backups
Where the operation is read
Client-side applications
Desktop app, mobile apps and the REST API to the WMS / ERP / MES
Desktop app for the office, mobile apps for the floor, and a REST API for your WMS, ERP or MES.
- 1 Desktop app For the office
- 2 Mobile apps For the floor
- 3 REST API(s) For the systems: WMS, ERP, MES
Desktop app
For the office
The control tower on any desk, in Chrome, Edge, Firefox, Safari or Opera, with nothing to install: the live map, the analytics behind it, the rules and alerts, and the reports. Sign in from anywhere and the floor is on screen as it is now, with its history behind it.
- Live view
- Alerts & reports
- Any browser
Mobile apps
For the floor
The platform in a hand on the floor. Logidot Operative guides picking on Android handhelds — batched orders, planned routes, every scan checked — and the mobile app carries the checks before a vehicle is used and the log of who used what, and when.
- Android
- Picking
- Pre-use checks
REST API(s)
For the systems: WMS, ERP, MES
Everything the platform knows is available over REST, and the WMS, ERP or MES you already run exchanges data with it in both directions. Orders, stock and master data come in, so a position on the map is a pallet with an order behind it; position, identity and status go back, so the manual, error-prone steps around the WMS can be retired instead of adding another screen. Plugins for the most common providers, or an integration built for yours.
- WMS / ERP / MES
- Both directions
- Plugins or on demand
The arithmetic between a tag and a screen
Algorithms
Positioning, picking routes, safety rules and flow analytics on one aisle, to scale
Four pieces of the platform turn raw readings into what you act on. Each one below is what runs today: what goes in, how it works, what comes out, and one number from the code.
Particle filter
How it works
Each asset carries 250 guesses of where it is and how fast it is moving. Every fix re-weights them, trusted by its quality — an expected error from 2 m down to 1 cm — on a heavy-tailed curve, so one wild fix is out-voted rather than followed. Fixes well below the recent quality are dropped; a gap over 3 s or a jump over 5 m restarts the track; an estimate more than 100 m from the measurements is rebuilt.
- In
- UWB fixes from the anchors, each with a quality score; multiband GNSS when the work goes outside.
- Out
- One smoothed position, velocity and confidence per asset, for the live map and everything below.
250 position guesses per asset, re-weighted on every fix
Walking graph, batches, 2-opt
How it works
The free floor becomes a walking graph on a 1 m grid, kept 0.5 m clear of racking. Orders are batched into trips that fit the cart, with heuristics from our work with the University of Bath; each trip’s stops are ordered nearest-first, then improved by 2-opt swaps until none shortens the walk. A new order joins a trip under way only if it adds at most 30% to the walk that is left (the default; each site sets its own).
- In
- Open orders and the bins they pick from, the picker’s capacity in units, kilograms or litres, and your floor plan.
- Out
- A pick list in walking order on the handheld, re-planned as orders arrive.
25% less walking, simulated on Northgate, our modelled reference DC
Rules on the position stream
How it works
Six kinds of rule: entering a zone, leaving one, a zone speed limit, a site-wide speed limit, idle too long, and near collision. Near collision compares every pair of assets seen within half a second of each other. Zone edges carry a margin — 0.5 m on UWB, 5 m on GPS — so an asset on the line does not flicker in and out, and the same pair does not alert twice within 3 s.
- In
- Every processed position, the zones on your floor plan, and the rules you set.
- Out
- An alert on the live map, graded info, warning or critical. With notifications on, critical alerts go by email and to the phone, warnings by email.
Critical Asset Forklift 3 in close proximity with asset Picker 7. Distance: 1.6m (allowed distance: 2.0m)
2 m near-collision distance every new site starts with, beside a 13 km/h speed limit
Activity model, minute summaries
How it works
A neural network reads each asset’s track and motion readings and labels it stopped, travelling or picking, committing only at 60% probability or more. Every minute, positions are binned into 0.5 m cells for the heat maps, distance is counted only while moving, and paths are thinned to 0.5, 1 or 2 m steps so a week of spaghetti map still draws.
- In
- The positions, and the motion readings from each tag.
- Out
- Time by zone, idle events and pick counts, activity breakdowns, heat, traffic and spaghetti maps, and the moves between zones.
0.5 m heat-map cells, summarised every minute
Your operation, run forward
Simulator
A model of your floor you can ask questions of before anything changes on it. Every picker and vehicle is an agent walking your real aisles to your orders, and what comes out has the shape of live data — so it lands on the same live map and analytics as a real site.
- 1 How many people does this order profile need? Run the same day with a different team, and compare lead times and how busy everyone is. Worked through below.
- 2 What does a busier day do? Turn the order rate up and see where lead times start to climb.
- 3 What does a new layout change? Change the floor plan — a rack run, a drop zone — and run the same orders through it again.
- Reads
- Your floor plan — racks, pick locations, drop zones — and orders, generated at a rate you set or loaded from a file
- Fleet
- Pickers, forklifts, pallet trucks and roll cages, each with a capacity, a speed and shifts
- Runs
- Every agent stepped twice a second over the floor’s walking graph
- Gives back
- Positions, activity and order events in the shape of live data
- Speed
- A nine-hour day in about 11 seconds in the run below; live runs play in real time
Worked example · a real run
Would a third picker help?
Northgate, as the run read it, and what two pickers and three made of the same day
2 pickers
Orders waiting 9
Done today 330
3 pickers
Orders waiting 3
Done today 336
The run itself, replayed: 15:40 to 16:00 of the first generated day, a busy stretch of the afternoon, at 20 times speed. Every dot on a rack is an item of an open order, waiting to be picked.
- 2 pickers
- 3 pickers
-
Average order lead time (min)
-
Slowest 10% of orders wait over (min)
-
Longest wait (min)
-
Pickers busy, 09:00–18:00 (%)
-
Walking per order (m)
A third picker cuts the longest order wait from 5.9 to 3.0 minutes. All 450 orders get done either way, and total walking doesn't change. Each picker is busy half the day instead of three-quarters.
Lead time runs from an order arriving to its items being dropped. Each day ran through the simulator engine on its own, in about 11 seconds.
- Floor
- Northgate, our modelled reference DC: 80 × 50 m, 15 rack runs, 130 pick locations, three drop zones
- Orders
- 450 a day, arriving 09:00–18:00, one to three items each
- Pickers
- Two, then three. 1.4 m/s walking, 8 s per pick
- Held equal
- The same orders and the same routing in both runs; mean of three generated days
Your site’s know-how, and your floor plan, read for you
AI Agent
A question, the documents it is answered from, and the answer
The assistant
A chat panel in the dashboard. Ask in words, and it answers from the documents your organisation has given it — procedures, safety rules, equipment manuals, your own notes.
Illustrative exchanges. The answers come from your own documents, so yours will read differently.
- AskedHow do we log a near miss?
- AnswerTell the shift lead before the end of the shift, then fill in the near-miss form: where, when, and the people and vehicles involved. Your safety procedure asks for a photo of the spot if it is safe to take one.
- AskedWhat does the pre-use check on a reach truck cover?
- AnswerForks and mast, chains, tyres, brakes, horn, lights, and the battery charge, plus any damage noted — done and signed before the first lift of the shift.
- AskedWhich aisles are closed to pedestrians on nights?
- AnswerYour site rules make aisles 4 to 7 vehicle-only on the night shift. People cross at the marked crossing by dock door 2.
- Answers from
- The documents you upload to it
- Where
- A chat panel in the dashboard, switched on per organisation
- Runs on
- Open models we host on Google Cloud, in Frankfurt
- Sign-in
- The same account as the dashboard
- What it does not do
- Read live positions, or change anything on the floor
A floor plan, one known distance, and the draft floor read from it
Floor plans, read into a draft
Setting up a floor starts from the drawing you already have. Upload it, or photograph a paper plan, in the Logidot mobile app: the reader finds the building outline, walls, racks and rooms, in metres, and lays them out as suggestions on a draft floor.
- 1 Upload the plan A DXF, PDF or photo of the plan. Photos have no scale: mark one distance you know, such as a dock door, and enter its length.
- 2 The reader finds the floor Outline, walls, racks and rooms are taken from the drawing, each with a confidence score. A guessed scale is flagged.
- 3 You review, then publish Accept or reject each suggestion, or all of them at once. Nothing changes on your live map until you publish.
No plan at all? Tap your building on the map in the app and its outline comes from open map data (Overture Maps), ready to place.
- Reads
- PDF, PNG, JPEG, WebP, DXF; a DWG is exported to DXF or PDF first
- Finds
- Building outline, walls, racks, rooms and areas
- Scale
- From the DXF’s units, or one distance you know
- Where
- The Logidot mobile app, when you set up a floor