@rmackay9 The flight on March 12th fails with latest version of ardupilot(at this time) where the MQTT features have been include. During this try the motor emit a strange sound.
However before the drone crash, we had successfully done another try in the morning with an older version of arducopter (Copter-3.4-rc1) download from the emild HP.
At home, we have done some tries (no flight, just motor) with the base version of ardupilot( Copter-3.5 without MQTT). The result is that the motors have a strange behaviour and making some strange noise, the same that we noticed during the failure and probably make the drone crashes.
Currentely, locally we moved the commit where you had the AP_Telemetry library on the Copter-3.4 branch. The compile successed and the motor seems to be fine. We are now trying to import the MQTT features on that new local branch based on Copter-3.4.
We made the demonstration of MQTT successful
I used Copter-3.4.
We did two demonstrations.
First, I connected Drone to the Internet with WiFi, got log (latitude, longitude, etc.), and operated (ARM, DisArm, take off, land etc.) Drone with MQTT from a remote place.
Second, connect Drone to the Internet with SORACOM (LTE / 3G), perform simple operation (ARM, DisArm) with MQTT from a remote place, loaded Drone in the car and got position information.
We shot a demonstration video, we will show you it when editing is over.
@rmackay9 We used the SORACOM USB dongle.
And we used a linux flight controller with the Navio2,
we didn't use a companion computer on the drone.
The Solacom band of both upload and download is 2 Mbps at the fastest speed class(s1.fast).
The communication rate is 0.2 yen / MB.
By 2 Mbps bandwidth, I think slow/laggy video is possible.
In this development, we communicate with Drone using the MQTT protocol
With the MQTT protocol, we basically only send text data.
Additional development is necessary for us to send and receive videos over the Internet, but it is possible in the future, because we succeeded in connecting Drone to the Internet.
@tajisoft @norika-takumi @p-goudet
It's the agenda of today's meeting.
(1) Development of Raspberry Pi + Navio 2 configuration
· Refactoring: Pierre
· Adding a debug log: Pierre
· Arrives the buzzer sound informing of the state of ArduPilot
(2) Development of Pixhawk + Companion configuration
· on hold
(3) KDDI's demonstration experiment at SIM
· Schedule of demonstration experiment
· Scheduled to be confirmed on Saturday, April 15
· Additional development of Web UI
· Introduction of framework
· Server side: Express
· Front side: React or Angular etc.
(4) Video streaming of ArduSub
@rmackay9 We wanted to check if the strange motor sound remains on the latest commit. So we rebased the dev branch on the master branch of the ardupilot origin project. The result is that the strange still remains and the disarm command seems to not be transmit correctelly (need long time to stop engine). The testing setup is a navio2 board with esc "LHI simonk 12A” mounted on a rubocop drone. The previous version 3.4 is still fine and the MQTT function seems to go smoothly.
Do you have any idea about where this problem might come from? (and if it might easily be solve?)