Reuse Existing Wi-Fi APs · No Rewiring · AI Fingerprinting

Wi-Fi Indoor Positioning (Reuse Existing APs, Lowest Deployment Cost)

Where a site already has Wi-Fi, Wi-Fi positioning is the lowest-cost option — existing APs are reused and no rewiring is required, delivering data transport and positioning from the same infrastructure. P-Square's proprietary AI Fingerprinting (RF Fingerprinting) algorithm overcomes Wi-Fi signal disturbance in partitioned environments, and measured positioning performance outperforms solutions from major international networking vendors such as Cisco and Ruckus. Deployed at a leading domestic semiconductor manufacturer for emergency response team safety, night-shift safety, duress alerts and restricted-zone warnings.

Architecture 1: Smartphone Positioning (Wi-Fi AP / BLE Beacon)

Smartphone positioning architecture diagram using Wi-Fi AP and BLE beaconsSmartphone positioning architecture: the site deploys Wi-Fi APs or BLE beacons, the handset scans the signals and computes its own position before reporting the coordinates to the positioning server for display on the web platform. On-device computation is a distributed architecture that scales with user numbers and applies self-adaptive correction for differences between handset brands.Smartphone Positioning: Wi-Fi AP and BLE BeaconSite side (no badge issued to users)Wi-Fi APreuse existing APsBLE Beaconbattery, no cablingUser smartphoneposition computed on devicesignal scanPositioning serverreceives computed coordinatesWeb platformwayfinding / navigation / BITwo benefits of on-device computation1. Distributed computationperformance holds as user numbersgrow, excellent scalability2. Handles device heterogeneityself-adaptive correction acrosshandset brands (Huawei, OPPO,Xiaomi, HTC, Samsung, Sony)Note: Wi-Fi positioning is cheapest where Wi-Fi already exists, but iPhone does not support Wi-Fi scanning, Android 10+ needs developer mode and scanning draws more power. BLE beacons offer better cross-platform compatibility.

← Swipe horizontally to view the full diagram →

Figure: smartphone positioning — the position is computed on the handset before being reported, a distributed architecture.

Personnel carry a Wi-Fi Mobile device or Wi-Fi tag and are positioned from the signals of multiple Wi-Fi APs; the scanned signals are sent to the back-end server for computation and display on the web platform. An alternative is for the user's handset to compute its own position before reporting it. This distributed architecture has two benefits: (1) performance holds as user numbers grow, giving excellent scalability; (2) it resolves differences between handset brands (Huawei, OPPO, Xiaomi, HTC, Samsung, Sony and others).

Position computation on the handset is handled by the positioning library we supply, which drops into the customer's existing Android app. The site reuses the Wi-Fi it already has, with no extra hardware to install. It is in service for merchandise location in retail stores and personnel location in factories.

Architecture 2: The AI Fingerprinting Algorithm

AI Fingerprinting algorithm architecture diagramP-Square AI Fingerprinting algorithm architecture: IMU inertial sensing, Wi-Fi and Bluetooth RF signals and structured map data are combined through RSSI pattern recognition, probability-based sensor fusion and decision fusion to output coordinates, floor and partition determination, improving accuracy by over 30% versus conventional trilateration in partitioned environments.AI Fingerprinting (RF Fingerprinting) Algorithm ArchitectureIMU Inertial SensingAccelerometer / GyroscopeMagnetometerRF SignalsWi-Fi / Bluetooth RSSIAoA / ToA / TDoA / UWBStructured Map DataFloor plans / image featuresPartition and door positionsRSSI Pattern Recognition — matched against the surveyed signal fingerprint databaseProbability-based Sensor FusionDecision Fusion — determines room, inside/outside a door, and floorPositioning output: coordinates + floor + partition, with realistic trajectoriesOver 30% higher accuracy than conventional trilateration in partitioned environments | Hardware independent: supports RSSI, AoA, ToA, TDoA and UWB

← Swipe horizontally to view the full diagram →

Figure: AI Fingerprinting architecture — fuses IMU, RF signals and structured map data to output positioning with partition determination.

Conventional positioning degrades as partitions increase and is poorly suited to indoor environments with many obstacles and rooms; trajectories often jump erratically or land in impossible places such as inside a pillar, which makes projects difficult to sign off. P-Square addresses this with three innovations protected by eight patents:

The algorithm is hardware independent and supports RSSI, AoA, ToA, TDoA and UWB, which is why P-Square can partner with many classes of hardware vendor and keep its business model flexible.

Four Limitations You Should Know About (We State Them Openly)

LimitationExplanationMitigation
iPhone does not support Wi-Fi scanningiOS does not expose surrounding Wi-Fi AP scan results to third-party apps, so pure Wi-Fi positioning cannot cover iPhone users.Pair with BLE beacon smartphone positioning where iPhone coverage is required, or use site-side tag positioning.
Android 10 and above requires developer modeAndroid 10 onwards restricts Wi-Fi scan frequency; continuous scanning requires developer mode.Use a Wi-Fi tag or a Bluetooth solution instead, or configure company-owned devices centrally via MDM.
Wi-Fi scanning draws more powerCompared with Bluetooth, Wi-Fi scanning consumes noticeably more power, affecting battery life.For staff wearing a device all day, use Bluetooth tags (about one year of runtime); Wi-Fi positioning suits handheld terminals or mains-powered devices.
Scanning is throttled with the screen off or the app in the backgroundTo save power, both Android and iOS heavily reduce or suspend scanning once the app goes to the background or the screen turns off, so position updates slow down or stop.We have an answer on both platforms: on Android a foreground service with a persistent notification keeps scanning alive; on iOS Core Bluetooth background modes and beacon region monitoring wake the app. Where the site needs continuous, reliable updates, switch to site-side tag positioning, which is unaffected by handset power saving.

Recommended Positioning Server Specification

ItemSpecification
Operating systemWindows 11 (64-bit) or Windows Server 2022 / 2025; Linux also supported (Ubuntu 22.04 LTS or later)
CPUIntel Core i5 12th generation or later, or equivalent Xeon, quad core and above
Memory16 GB or more
Storage512 GB SSD or more, depending on how long history is retained
Supported methodsWi-Fi AP smartphone positioning, BLE beacon smartphone positioning
Beacon requirementBluetooth 4.0 or above
Handset supportAndroid for Wi-Fi positioning; BLE beacon positioning recommended for iOS

Case Study: Leading Domestic Semiconductor Manufacturer

Personnel carry a Wi-Fi Mobile device or Wi-Fi tag and are positioned from multiple Wi-Fi AP signals, which are sent to the back-end server for computation and displayed on the web platform. Benefits realised:

The site also applied Wi-Fi positioning to live headcount during fire drills and went on to take first place in the group-wide inter-fab drill assessment. The solution reuses existing plant Wi-Fi APs with no rewiring, making it the lowest-cost route for an operating facility.

A leading domestic petrochemical group uses the Mobile Beacon system: users carry Android or iOS handsets, fixed BLE beacons are deployed across the site, the handset computes its position from beacon signals and reports it over Wi-Fi or 4G/5G to the back-end server for wayfinding, navigation and business intelligence analysis on the web.

Where It Is Used

Frequently Asked Questions

We already have Wi-Fi in the plant — can we really use it for positioning?
Yes, and it is the lowest-cost approach: existing APs are reused with no rewiring. Four limitations apply: iPhone does not support Wi-Fi scanning, Android 10 and above requires developer mode, Wi-Fi scanning draws more power, and scanning is throttled with the screen off or the app in the background. In practice a mixed Wi-Fi and Bluetooth deployment is common, and P-Square assists with selection during assessment.
What accuracy does Wi-Fi positioning achieve?
Roughly 3–5 metres, suitable for wide-area approximate positioning, material distribution by zone and evacuation headcount. For 1–3 metres use Bluetooth; for 10–30 cm accuracy use UWB.
How does performance compare with Cisco or Ruckus?
Measured positioning performance outperforms solutions from major international networking vendors. The difference lies in the algorithm rather than the hardware — AI Fingerprinting combined with structured map data improves accuracy by over 30% versus conventional trilateration in partitioned environments, and the algorithm is hardware independent so existing APs can be reused.
Why emphasise that the algorithm is hardware independent?
Because it means customers do not have to replace existing network equipment to gain positioning, and it means P-Square can work with many classes of hardware vendor. The algorithm supports RSSI, AoA, ToA, TDoA and UWB, giving high extensibility.
What server specification is required?
Windows 11 64-bit or Windows Server 2022 / 2025, with Linux also supported (Ubuntu 22.04 LTS or later); Intel Core i5 12th generation or equivalent Xeon, quad core; 16 GB of memory or more and a 512 GB SSD or larger. BLE beacons must comply with Bluetooth 4.0 or above.

Turn the Wi-Fi You Already Have into Positioning

Contact the P-Square project team to assess whether your existing APs can support your positioning requirements.

Professional Consultation