RFID, GPS, BLE or UWB: choosing by the question the asset asks

01 / FIELD NOTE

Keep the decision tied to the operating context.

Asset tracking choices usually fail at the first sentence, because identification and location are treated as one question. Identification answers what an asset is and where it was last seen at a defined point; location answers where it is now. A portal read, a satellite fix, a beacon heard by a receiver and a time-of-flight measurement are four different answers to two different questions, and the technology that fits falls out of which question the operation is actually asking, at what scale, and against what budget.

Passive RFID is identification at checkpoints. The tag carries no power of its own; the reader’s field energizes it, so the tag is cheap, small, durable and free of batteries, at the cost that a read happens only near a reader. That makes it a natural fit wherever assets pass defined points in volume — a dock door, a shelf, a conveyor, a cabinet — and it reads whole batches without line of sight. What it does not give is continuous position: between checkpoints the asset is invisible, and the value of the system is that the checkpoints are placed exactly where the operating decisions are made.

BLE is the indoor proximity layer. A battery-powered tag broadcasts a low-energy signal on a duty cycle, receivers within range pick it up, and the system estimates position from signal strength — which is to say a zone, not a coordinate. The tag’s power budget is where this technology wins: a low duty cycle can keep a tag transmitting for years on a small cell, so the total cost is dominated by the receivers and their placement, not by the tags. In return, accuracy is measured in metres, metal and obstructions push the estimate around, and the answer is deliberately coarse: which zone an asset is in, not exactly where within it.

GPS answers where for outdoor motion. A receiver takes a fix from satellites and the module sends the position over a cellular network, so coverage spans any geography where a device can see the sky and hold a connection — and requires no local infrastructure at all, which makes it the natural fit for vehicles, containers and equipment moving across regions. The cost is power: receiving, computing and transmitting a fix drains far more than a beacon ever would, so the device either draws on the vehicle’s supply or runs on a battery that sets its own life. The sky is the requirement: inside a building, under a deck, or in a deep yard, the fix degrades or disappears completely.

UWB is the precision indoor layer. Instead of signal strength, it measures the time a pulse takes to travel between the tag and several fixed anchors, which yields distances and therefore a position accurate to centimetres — in conditions that defeat the other radio methods, because the wide-band pulses resist the reflections and multipath of dense metal environments. The price is infrastructure: a dense, well-placed grid of powered anchors, with tags and anchors both costing more than their indoor alternatives. UWB earns its cost where the process needs the exact position continuously — a fast-moving item on a production line, a machine coordinating against the assets around it — and it is the wrong answer where the process only needs to know whether the asset is present.

The printed label is the honest baseline. A barcode or QR code costs nearly nothing, needs only a scanner or a phone, and is understood everywhere — and it pays for that simplicity in labor, because every read is a line-of-sight scan performed one item at a time. Wear, dirt and damage end a label’s life early, and nothing about a label reports an asset on its own. Where volume is low and movement rare, the label is not a legacy choice to be defended; it is the correct minimum, and the comparison starts from it rather than from the newest radio.

The selection axes are five, and they are physical before they are commercial. Power: does the tag harvest from a reader field, carry a battery, or connect to external power — each sets the tag’s size and life. Cost: per tag, but also per reader, per anchor, per gateway, per mount and per cable run, which for the indoor technologies can dwarf the tag itself. Latency: checkpoint-time, near-time, or real-time, and the tolerance the process can actually live with. Environment: indoor or outdoor, and how much metal, liquid and obstruction the signal has to cross. And scale: how many assets, how often they move, and how many positions one location question has to resolve.

The match follows from the axes. High-volume processing at defined points, where the value is in batch identification, points to passive RFID at portals, shelves and conveyors. Movement across wide outdoor geography points to GPS on the asset, with the power budget to feed it. Indoor zones with many moderate-value assets and a tolerance for coarse position point to BLE. Continuous centimetre position indoors points to UWB, and the process has to be able to spend what its anchors require. None of these matchings is a ranking; each is a physical fit between what the operation asks and what the technology does.

A deployment can answer different questions in different phases of the same journey, and the layered answer is normal. The same identity can be a passive tag at the packing bench and the dock, then a GPS device on the vehicle as it crosses regions; the record is one asset identity, and the capture layers differ by phase. What keeps the hybrid coherent is that every layer writes to the same identity and the same event stream; what breaks it is a second identity created for the second layer. The integration decision is as real as the hardware decision, and it is made at the same time.

Two honest limits close the comparison. The first is that total cost is infrastructure, not tags: a cheap tag served by a demanding grid of anchors and gateways can outspend a costlier tag served by fewer, better-placed readers, so the per-tag figure never stands alone. The second is that no technology combines continuous centimetre position indoors with wide-area outdoor tracking at passive-tag cost; each axis is a budget, and the selection is where the operation chooses to spend it. And where volume is low and movement rare, the cheapest answer may be no continuous tracking at all — a label, a periodic count, and the recognition that the question did not exist.