When a device sends a signal in a highly reflective environment, it will hardly ever reach the gateway directly. The signal will bounce from one or multiple buildings until it reaches the gateway. This will create a longer path, later fine timestamp and a lower accuracy.
If the device is mounted directly next to a window (with coating) or concrete material, reflection will negatively impact the accuracy of Geolocation.
In general, you do not need a special device. However, instead of just having a connection (one gateway) as with regular devices, messages need to be received by multiple gateways to perform the most effectively. So, to increase quality, special attention should be given to antenna design, signal direction and configuration of the device (concerning ADR, for example).
For now, we see that static performance is most important for our customers, so we are improving both accuracy and stability (success rate) for static device profiles first. We also noticed that the “first time Geolocation accuracy” is relevant for our customers. We are constantly innovating our IoT services and, as we are the first operator in the world to offer nationwide LoRa Geolocation, we are still expecting many improvements of this service in the future.
There is quite some pull from the market for such services. Most of the time, solutions based on LoRa only are not sufficient for indoor Geolocation, but this does not mean that KPN is not able to help you. Please talk to our IoT team; we have quite some experience with devices that support multiple technologies (RF, Bluetooth, GPS) to answer your needs.
The best practice is to enable Adaptive Data Rate (ADR) on your device, so that the network can calculate the best Spreading Factor. In the field tests KPN has performed together with customers, it turned out that using SF12 does not always provide the best results, since using SF12 results in a bigger chance of collisions in the air.