What Are Indoor Positioning Systems?

Indoor positioning systems are sensing and computing arrangements that estimate the location of a person, device, or vehicle inside a building, where signals from the Global Positioning System and other satellite constellations are attenuated by roofs and walls to the point of being unusable. They substitute terrestrial signal sources, onboard inertial sensors, or environmental features for satellite ranging, and they report position in a building-referenced coordinate frame such as a floor plan grid, a room label, or a zone identifier. Accuracy targets vary sharply by application, from room-level resolution for asset tracking to centimeter-level for automated forklifts and robot docking.

The engineering problem differs from outdoor navigation in three respects. Indoor radio propagation is dominated by multipath and non-line-of-sight conditions, so time-of-flight and signal strength measurements are biased rather than merely noisy. The environment changes as furniture, inventory, and people move, which degrades any model calibrated in advance. Finally, there is no single global infrastructure, so each deployment depends on whatever anchors, access points, or beacons the building already has or is willing to install.

Ranging and Geometric Estimation

Ranging approaches measure a physical quantity related to distance or bearing between the target and fixed anchors, then solve a geometric problem. Time of arrival and time difference of arrival convert propagation delay to range and resolve position by trilateration or hyperbolic multilateration; angle of arrival uses antenna arrays to intersect bearing lines. Ultra-wideband is the dominant radio choice for high-accuracy work because its nanosecond-scale pulses separate the direct path from reflections, and a survey of ultra-wideband indoor positioning for industrial IoT reports decimeter-level accuracy under favorable anchor geometry. Bluetooth Low Energy direction finding and IEEE 802.11 fine timing measurement bring similar geometric methods to commodity hardware at reduced precision. Dilution of precision, anchor placement, and clock synchronization between anchors set the practical error floor.

Fingerprinting and Radio Maps

Fingerprinting avoids propagation modeling entirely. In an offline survey phase, received signal strength or channel state information is recorded at known reference points to build a radio map; in the online phase, a live measurement vector is matched against that map by nearest neighbor search, probabilistic inference, or a trained classifier. An overview of Wi-Fi fingerprinting for indoor positioning sets out the accuracy dependence on reference point density and the cost of resurveying when access points are moved or replaced. Deep learning has reduced that burden: autoencoder and convolutional architectures such as the CDAE-CNN fingerprinting system denoise sparse signal measurements and generalize across devices with different radio front ends. Crowdsourced map construction and transfer learning address the same calibration cost from the data collection side.

Sensor Fusion and Dead Reckoning

No single indoor modality is reliable across a whole building, so production systems fuse several. Pedestrian dead reckoning integrates accelerometer and gyroscope output from a phone or wearable to propagate position between radio fixes, with step detection and heading estimation bounding the drift. Kalman and particle filters combine that motion model with radio observations and with map constraints, since walls and corridors eliminate physically impossible trajectories. Magnetic field anomaly matching, visible light communication, ultrasonic beacons, and visual simultaneous localization and mapping serve as additional observation sources where the radio environment is hostile.

Applications

Indoor positioning systems have applications in a wide range of settings, including:

• Warehouse and factory asset tracking, including forklift and autonomous mobile robot guidance
• Firefighter and first responder tracking during interior operations
• Hospital equipment location and patient flow analysis
• Retail analytics, wayfinding, and in-store navigation
• Airport, museum, and campus wayfinding for visitors
• Contact tracing, occupancy sensing, and building energy management