Definition
Hospital Asset RTLS (Real-Time Location System) achieves meter-level or room-level positioning through UWB, BLE AoA, RFID, Wi-Fi fingerprinting, etc., serving defibrillators, infusion pumps, wheelchairs, and high-value consumables. Location events mostly converge in the hospital LAN and are transmitted to the hospital asset platform via gateways or edge servers over cellular networks, usually without opening a cellular connection for each individual tag.
Technical Features
Must handle multi-floor drift, elevator shaft jumps, and interference; the positioning engine often interacts with HIS and SPD work orders. Transmission links should distinguish between real-time alarms and batch trajectories to control air interface and cloud costs.
Use Cases
Around the emergency department, ICU, operating rooms, and logistics warehouses; implementation requires calibration of base maps, number segments, and departmental cost allocation. SimtoLink can provide manageable and controllable cellular access and card management strategies for off-campus branches, emergency stations, etc.
Key points for cooperation with IoT SIM cards
- Divide data pools by campus/gateway, silent cards and abnormal traffic go into work orders.
- Private network or dedicated APN, TLS, and whitelists; patient-related trajectories and logs are anonymized with minimal retention.
- ICCID is bound to gateways and anchor batches, with local caching and retransmission in weak network scenarios.
- Evaluate roaming and multi-campus replication fees and compliance boundaries in advance.
Aggregate at the Gateway, Do Not Open a SIM per Tag
What a location event means on the way back
Hospital asset location means a tag is seen by an anchor or a base station inside the building, the location engine computes a room or a floor, and a gateway then sends "this device is here now" to the hospital-wide platform. The cellular SIM is usually only on the gateway or the edge server, because the building already has a location network. Opening a wide-area SIM for every infusion-pump tag still needs the in-building radio to locate the tag. The SIM only adds a charge and a batch of silent numbers. Alarms (leaving an agreed zone, or no movement for a long time) should go out live. History tracks can be backfilled in batches.
The SIM number maps to the gateway and the campus. The tag number maps to the asset. Bind them the wrong way and the platform shows campus A's pump at campus B. The join to clinical systems is covered under HIS and EMR, but a location event is not an order. A change of position must not rewrite the patient record by itself.
When a branch site uses cellular
In-building location is already calibrated, and cellular only solves how this building's aggregation point reaches the central platform. A branch, an emergency station, or a warehouse with no private line fits that. If the main hospital plant room already has a private line, cellular should be the backup, not a second primary that fights the first. The allow-list admits only the asset platform. If a track can imply where a patient went, retention and de-identification follow hospital rules, not the default log of the SIM system.
Elevator shafts cause floor jumps. Filter those on the engine side. Do not upload every jump as an urgent work order on the SIM side. Size the plan by the number of gateways, not by the number of tags. Before copying across campuses, confirm roaming, addresses, and account boundaries, so campus A's gateway does not register into campus B's platform.
Floor drift that opens a false recovery ticket
A common failure turns alarms on before the base map is calibrated. A tag in the elevator is computed on another floor, the system opens a find-and-return ticket, and the device is still in the original department. Another has no sequence number on the live stream and the batch track, so a backfilled old position covers the new one and the device walks backward on the map. After the gateway SIM is out of credit, in-building location still shows locally, assets on the central platform all go grey, and dispatch thinks the tags are broken.
Opening a SIM per tag also leaves a large number of identities that never attach, mixed with gateways that are truly offline, and the alarm stops meaning anything. Acceptance should push one device into the elevator and back to the original room: no cross-building recovery ticket, no position time running backward after the network returns, and a SIM count close to the gateway count rather than the tag count.
An anchor losing power and a tag sitting still are two completely different kinds of "not seen." When an anchor loses power, a whole area of devices disappears from the map at once. That should be reported as a gateway or anchor fault, not as a lost-ticket for every infusion pump. A tag that does not move for a long time may be equipment parked in a store. Thresholds should differ by asset type. A defibrillator still for a few hours is worth a question. A wheelchair still overnight may not be. If cellular backhaul packs both events as "position empty," the central platform cannot assign them. Cache backfill should carry the event kind and the anchor id. If gateways in different buildings reuse one set of anchor coordinates, a device teleports to another campus on the map while the SIM looks normal, because both gateways are online. The coordinate calibration file should be versioned with the gateway SIM number. A SIM swap or a gateway swap must not carry the calibration onto the wrong building by itself.
Low tag battery and actually leaving the hospital are not the same event. Low battery should prompt a battery change. Leaving the fence is what opens a recovery. If low battery is drawn as the device vanishing, nurses go looking for a pump that is still at the bedside. Keep the event codes apart, and do not backfill low battery as a departure. After a gateway changes SIM number, old events not yet sent should still carry the old gateway identity and time. The platform inserts them by time. It does not pile them all at the moment of the swap.