Adaptive Data Rate
The solution supports Adaptive Data Rate (ADR) device management as specified in LoRa spec v1.1. The main goal of the ADR is to optimize the network for maximum capacity ensuring devices always transmit with their best settings possible (highest data rate, lowest power), you can find more ADR information on this page.
Adaptive Data rate is always initiated and set on the device side. ADR should never be used with moving devices. Our solution currently implements the Semtech proposed Algorithm for ADR and has been tested against EU868 region plan. In this algorithm the data rate and transmission power calculation is done as follows:
Signal-to-noise ratio (SNR) is calculated based on the maximum SNR detected over the recent transmissions, the required SNR for the given region (which depends on the spreading factor of the current data rate) and SNR margin (always equal to 5dB). The calculated SNR determines the number of steps that will be executed for the calculation of new data rate and transmission power.
In each step we try to increment the data rate, as long as it's lower than the maximum data rate supported for the region (DR5 in case of EU868). We then increment the TX power index as long as it's lower than the highest TX power index for the region.
To determine the maximum TX power index we use the TX power table which is documented in LoRaWAN Regional Parameters specification. The table defines the mapping of TX power indices (0 - 7 in case of EU868) to EIRP (Equivalent Isotropically Radiated Power) in dB. The highest EIRP is mapped to the lowest TX power index, hence in order to achieve a lower transmission power we increment the TX power index in each step of the algorithm.
Once the remaining steps are equal to 0, the new data rate and TX power index calculated this way are returned.
In the case where the number of steps calculated in the beginning is negative the algorithm decrements the TX power index (as long as it's greater than 0) but it does not try to lower the data rate, because the end-devices implement automatic data rate decay. The algorithm can only actively increase the data rate.