Improve LoRa connection logic

[MeierTobias]

The MKRWAN library has some flaws regarding the timing and handling of the downlink message. Unfortunatly there is a lot of work to do to improve the library so here are some options:

Use a better antenna:

• order antennas
• test if this makes the setup more reliable

Change to MKRWAN_v2 and update library:

• move to v2
• test if everything works and if the issues still remain
• improve the library to solve the join issue
• fix the sendAndReceive example bug

STM dev kit

• order STM dev kit
  test if the ST Murata dev kit faces the same issues or if this is much more reliable
  test the mcci lcmi library and if this works much better consider changing to the MCCI Cartena 4610. (https://github.com/mcci-catena/arduino-lorawan library is a high level abstraction of the https://github.com/mcci-catena/arduino-lmic which provides an adapted LoRaWAN-MAC-in-C library (from IBM))

[guidosch]

Below a screen shot of the July on the Adlisberg Trail with fresh loaded batteries. It first was online a few days then offline and then felt again in the endless sending mode and the offline again. Later no more data but the batteries where drained within just a few weeks by doing something...

[MeierTobias]

Regarding the MKRWAN_v2 library:

• There is a quiet some effort needed to fix some stuf in the library to be able to use it for the BikeCounter. (16h+ invested and still not fully functional.)
• So I changed the code on the clean-up branche back to the v1 library.

Another issue is the signal strenght from my desk to the next gateway. It is very hard to debug the application if I encounter a signal error every second time. To fix this I orderd a cheap Gateway (<100chf) for my office place. (As we discussed last time)

So the current open points are:

• setup the gateway
• test the new version with the MKRWAN_v1 library
• Update the field devices (maybe replace the antenna)
• setup the STM lora dev kit.
• transfer the BikeCounter app to the STM board and test the reliability.