Menü
Aktuelles
Von:
Filter

CaSSAndRA

hamselvdk

hamselvdk

Member
Is there a dedicated forum for CaSSAndRA somewhere? is it "only" on git?

Can we start / are we allowed to discuss it here?

After I switched the robot seems to not be going in straight lines - more a wobbly approach to a line :)

+ the docking fuction seems not to be working (ok didnt see this at first: "Current restrictions: If the charging station is outside the perimeter and the mower is supposed to drive from the charging station to the desired point, the task will be aborted with an error." so that explains that ... )
 
Zuletzt bearbeitet:
Just switch back to the sunray app to verify ... straight lines and no wobbling ...
Does CaSSAndRa make too many setpoints? or what is it that give the wobblyness in the forward motion?
 
Hm...that is strange. Cassandra only sets the waypoints and does not "really" control the mower. That still does Sunray. Does Cassandra maybe use a different speed than Sunray?
 
hamselvdk schrieb:
the docking fuction seems not to be working (ok didnt see this at first: "Current restrictions: If the charging station is outside the perimeter and the mower is supposed to drive from the charging station to the desired point, the task will be aborted with an error." so that explains that ...
Zum Vergrößern anklicken....
If you on current version of sunray then there are no restrictions anymore
 
Hmm - it is still not docking (also not with the sunray app if I change and upload from that) ... guess its a sunray mower thing..
 
Question - is there a way to visualize the lawnmower progress in CaSSAndRA for the scheduled tasks? (if I missed it Im sorry for my question up front)
 
hamselvdk schrieb:
Is there a dedicated forum for CaSSAndRA somewhere? is it "only" on git?

Can we start / are we allowed to discuss it here?

After I switched the robot seems to not be going in straight lines - more a wobbly approach to a line :)

+ the docking fuction seems not to be working (ok didnt see this at first: "Current restrictions: If the charging station is outside the perimeter and the mower is supposed to drive from the charging station to the desired point, the task will be aborted with an error." so that explains that ... )
Zum Vergrößern anklicken....
The wobble is probably caused due to a unstable http connection that initialy is caused by cassandra due to connection establishment overhead, eating up firmware runtime every now and then loop. The sunray UI app is much more efficient in that..... And the orig sunray FW on the mower has no protection for that, if the message is not transported completely or with errors in it.... Causing 1 seconds hang with each cassandra connections. In this time while the system hangs on some http transmit data, no sensor is updated. The mower looses track or does funny stuff that is not comprehensive when it comes back to normal readings... resulting in heavy offsets of sensor states and triggering obstacles or just does "snake" lines because the tracker could not track.. it´s miniature firmware freeze. This is what i am familiar with, traced it back and tried to debug it last year. It was a pain until realisation. On my side, it´s fixed. This is also part of why you cannot fluently steer the mower manually with cassandra. This is called blocking an embedded system because of a function that waits too long for something.
 
Zuletzt bearbeitet:
I am not sure how mqtt is collected. It may use the same connection routine, but it's worth a try or someone has more knowledge about that. best is to unblock the httpserver.cpp. if you are familiar with vscode and or compiling, you could try to unblock it. Also you can check if this thesis is right for your situation. In the log you would see warning loop time notifications during tracking.
 
Yes. The serial Robot Driver is a complete other issue. It's Alfred and I hope not owl Plattform. On Alfred, comms can be stuck just because of electro magnetical disturbance mainly from mowmotor. I am not sure if the http server cpp is acting up in Alfred as well or if the data is forwarded a different way, but the 1 sec freeze lookes like that. So, I would not trust this analysis completely but check disabling Cassandra client in a mow session. If robot works much better it's absolutely the http server architecture. That's what I observed last year. Alfred working shit: with Cassandra absolute katastrophe.
 
so I changed to MQTT with a local mosquitto on the PI as gateway - had CODEX help me rewrite sunray/mqtt.cpp and mqtt.h + robot.cpp for service/TTY alignment... and I now have a stable connection (as long as my wifi adaptor isnt loosing connectivity) and Cassandra recieves the correct values and the send start/stop/coordinates are recieved correctly. I wont commit it as it is AI edited ..
 
Zuletzt bearbeitet:
Du musst dich einloggen oder registrieren, um hier zu antworten.
Oben