-
Caileigh Fitzgerald authored
- `comms_slots_` - `nav_slots_` - `active_slots_` - fixed and tested nav slots and comms slots with split slots so the active slots on the modem sim were a superset of the comms and nav slots. The modem sim only pinged in nav slots and only sent packets in comms slots (and in nav slots if comms slots overlap and there was remaining time to send a packet). This was tested using test_slotted_aloha.launch. This should probably get a basic test case and a seperate slotted aloha test file. - updated tdma status messages in subclasses. They are now all flat instead of nested super classes msgs as attrs. So a TdmaScriptedStatus msg has a `msg.we_are_active` from tdma_basic and so does a TdmaAdvancedStatus msg - the datatype for `comms_slots`, `nav_slots` and `active_slots` has changed to int8[] from int8 (since it's actually a list) in .msg files. - fixed bug in config_slot_mask - got `nav_slots` and `comms_slots` working in sim - added error handling and better warnings so users can't run with a slot configuration that would not allow a transmit (can be overriden and base class. It is in tdma_advanced) - added handling for `tdma_advanced` node so the node keeps the old values for `packet_length_seconds` and `guard_time_seconds` in the above case if dynamic reconfig passes a value for one or both that mean their sum is >= `slot_duration_seconds`. The node will print warnings. Turns out TDMA will not transmit in this case ever. So you can't start up this way without TDMA throwing an error (since in base TDMA it does not use dynamic reconfigure). BUT, if a user passed a value for `packet_length_seconds` or `guard_time_seconds`, it now handles the error thrown in tdma and just keeps the prior values and prints an error so the user knowns their values did not stick (it also resets the dynamic config param so rqt will go back to saying the prior value as soon as they hit enter)
Caileigh Fitzgerald authored- `comms_slots_` - `nav_slots_` - `active_slots_` - fixed and tested nav slots and comms slots with split slots so the active slots on the modem sim were a superset of the comms and nav slots. The modem sim only pinged in nav slots and only sent packets in comms slots (and in nav slots if comms slots overlap and there was remaining time to send a packet). This was tested using test_slotted_aloha.launch. This should probably get a basic test case and a seperate slotted aloha test file. - updated tdma status messages in subclasses. They are now all flat instead of nested super classes msgs as attrs. So a TdmaScriptedStatus msg has a `msg.we_are_active` from tdma_basic and so does a TdmaAdvancedStatus msg - the datatype for `comms_slots`, `nav_slots` and `active_slots` has changed to int8[] from int8 (since it's actually a list) in .msg files. - fixed bug in config_slot_mask - got `nav_slots` and `comms_slots` working in sim - added error handling and better warnings so users can't run with a slot configuration that would not allow a transmit (can be overriden and base class. It is in tdma_advanced) - added handling for `tdma_advanced` node so the node keeps the old values for `packet_length_seconds` and `guard_time_seconds` in the above case if dynamic reconfig passes a value for one or both that mean their sum is >= `slot_duration_seconds`. The node will print warnings. Turns out TDMA will not transmit in this case ever. So you can't start up this way without TDMA throwing an error (since in base TDMA it does not use dynamic reconfigure). BUT, if a user passed a value for `packet_length_seconds` or `guard_time_seconds`, it now handles the error thrown in tdma and just keeps the prior values and prints an error so the user knowns their values did not stick (it also resets the dynamic config param so rqt will go back to saying the prior value as soon as they hit enter)
Loading