TTN Saarpfalz, Germany

It’s been almost a year since we initiated the TTN community Saarpfalz. Since then, several gateways and some sensors have been installed. But to unleash the community to become official, things like postings in this forum are still missing. I would like to change that with this.

The link to our TTN community page is:

We get a lot of help from the commune Kirkel which supports the idea of an open IoT-network. Together with Armin Jung and the support of the mayor Frank John it was possible to launch some interesting pilot projects:

  • measuring flow level of the Blies
  • soil temperature measurement for winter services
  • co2 measurements in schoolrooms

You can find some more information as well as links to the measured data (grafana dashboard) here:

Currently we are working on the development of a traffic counter for cycle paths.

Join our community. We welcome new members.


1 Like

Sounds very interesting, please share this development with the global community as well if there is any progress to report.

Yes gladly - here are some short information about the traffic counter project:

Armin Jung - who is also the local officer for cycling (Fahrradbeauftragter) here in Kirkel - asked me about a way to count bicycles on bike paths. That was when I told him about IoT and my plans for a local TTN community. I decided to try my luck …

The traffic counter uses radar to detect objects passing. Compared to the classic solution with road tubes, radar has the advantage that the counter is easier to install: Simply screw it to one pole from a traffic sign. Road tubes also always present a certain risk of tripping when not properly installed. The counter uses a prebuild radar doppler module. The signal is amplified with a low noise, rail to rail op-amp and then send to the ADC of an MKR 1300 WAN arduino. The signal is sampled with about 4 kHz to get the necessary information which allow to recognize what kind of road user pass. With a certain error rate because the main indicator is currently the speed. And there are fast pedestrians and slow cyclists … Another challenge are cyclists who pass in groups. All in all, there is about a 5 … 10 % error rate. For the purpose of measuring the magnitude of the traffic volume, this is perfectly fine.

We have followed some approaches with deep learning but are currently using classic methods in the arduino firmware.

The counter sends the values via LoRaWAN to the backend but shows the values also on a little e-ink display. An optional GPS-module allows georeferencing. Power is supplied via two 18650 batteries. That means energy for 2 … 3 days, which is sufficient for temporary measurements (a weekend and a working day). With this power supply the unit is relatively small. Optional you can attach a bigger external battery for extended operating time.

Here are two pictures of the traffic counter and of the pcb:



1 Like