We are using a data logger with a water flow meter and I would like to create an algorithm to detect certain water anomalies using Tago.io and TTN.
The manufacture use their own peivate platform and network and says data is extracted from the message every 4:36 minutes when using their network and platform.
When we extracted the data from our platform we found that the times are random like every 20 seconds. When we asked the manufacturer they say it is TTN behavior. So now we are at a standstill. My developer needs an exact message occurrence to design a proper algorithm.
Is there a fix for this on TTN? Or maybe a different solution?
Define exact - to the second, milli-second, micro-second?
This will never happen with normal LoRaWAN - you may get very very very very close, but never exact - unless exact means a few seconds either way. If you put a GPS to get an accurate timebase on the device, you can surely trigger a send at an exact time but whether the whole infrastructure receives & processes to your requirements is another matter.
@Gianpaul It would also help if you advised what datalogger/meter solution you are using/having problems with - other forumites may have already solved any problems or have working solutions using TTN
Sounds a lot like they are processing the individual messages (more or less the same excessively frequent ones TTN is getting spamm’d by) to extract additional meaning from what is available in an individual one.
It doesn’t really sound like this sensor fits in TTN usage guidelines.
Even apart from that, you’d need an algorithm to distill down the frequent unreliable messages into less frequent reliable summaries. And that algorithm would need to be informed by knowledge of exactly what the node transmits. For example, does it send the same message multiple times for redundancy?